IRCloggy #git 2011-01-19

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2011-01-19

← calvinstephens left00:00
deitarion How would I retroactively move a bunch of commits into a new branch so, when I push the stuff, the old HEAD on the master branch doesn't change?00:00
→ ki__ joined00:00
nollidj deitarion: have you looked at --amend for git commit?00:01
← eletuchy left00:01
← Cstep4 left00:01
← mloki left00:02
SethRobertson deitarion: You would actually git branch newbranch; git reset --hard <ORIG_SHA> (or HEAD^^^^^)00:02
← cehteh left00:02
← davidc_ left00:02
→ rhl joined00:03
deitarion SethRobertson: That makes sense now that I think about it. Create a new branch, then roll back the master to the published head and let git's internal consistency preservation keep things intuitive.00:03
→ cehteh joined00:04
← smolyn left00:05
← mrtadis left00:05
deitarion nollidj: Not sure how that'd work. As far as I can tell from the manpage, it's just for injecting new changes into a commit back down the line. What I want is an "oops. I destabilized the 'stable dev' branch and now I want to share it, but it's going to take too long re-stabilize it" solution. SethRobertson's looks like it'll do that for me.00:05
noel_v So the waiting longer totally worked, thanks NfNitLoop!00:05
nollidj deitarion: if you look at what it says it's an equivalent for, it is more or less what SethRobertson suggested00:06
← SIGBUS left00:07
SethRobertson nollidj: how is it going to change the branch? It also works only on the last commit00:07
deitarion nollidj: Even if the manpage actually used grammatically valid english, I don't see how a command which includes a commit is what I want.00:07
← cwebber left00:07
← jrmuizel left00:08
← zombor left00:08
deitarion Sentences which aren't even parseable aren't very helpful --> "Prepare the tree object you would want to replace the latest commit as usual (this includes the usual -i/-o and explicit paths), and the commit log editor is seeded with the commit message from the tip of the current branch."00:09
SethRobertson The "would" is a little distracting, but it is a good description00:10
→ tgunr joined00:10
← Dave^| left00:11
→ Mannequin joined00:11
deitarion To someone who knows how to mentally complete the sentence perhaps.00:11
Is that first bit supposed to be "Prepare the tree object *that* you would want to replace the latest commit..."?00:11
← normanrichards left00:12
← psoo left00:12
SethRobertson Well, I deleted the would instead of adding the that, but yes00:12
← EnginA left00:12
tgunr I'm trying to move a svn repo over to git, I had previously done a git init on the local svn working copy so now the folder has both .svn and .git There are no svn branchs so is safe to just rm all the .svn folders and then make a bare repo to put onto my git server?00:12
deitarion Adding a "that" would still clarify that it's not really "Prepare the tree object. You would want to replace the latest commit as usual."00:13
SethRobertson tgunr: Are you trying to save history or just move the most recent checkout?00:13
← sterlo left00:14
← Anubisss left00:14
→ davidc_ joined00:14
nollidj SethRobertson, deitarion: i'm referring to the bit after "it is a rough equivalent for"00:15
→ EricInBNE joined00:15
nollidj not that it's very important now00:15
deitarion And I already knew git reset... It just hadn't occurred to me that git has a strong enough consistency guarantee that a reset after a branch creation would do what I wanted.00:16
nollidj ah, ok00:16
tgunr would be nice to have the history but not necessary00:16
← alip left00:16
→ eletuchy joined00:16
SethRobertson tgunr: Having a .svn and .git in the same directory is a sign (to me) that you are probably losing history. Others can comment on using git-svn to import history, or using one of the many third party tools (I can point you) to do that.00:17
← oracal left00:17
← ajpiano left00:17
← Chillance left00:17
SethRobertson However, if you are just taking the most recent copy, do a svn checkout, remove or modify all svnish files, then do a `git init; git add .; git commit -m "Import of svn 341 into git"` and you are done00:17
→ luxurymode joined00:18
← giallu left00:18
Arrowmaster svn export would be a better idea than a checkout00:18
SethRobertson Depending on whether you want to populate .gitignore, but yes00:18
→ alip joined00:20
tgunr this was a recent git init so it only has a couple of commits in it, but a lot of svn history00:20
ok, will try it and see what happens00:21
→ pr0wler joined00:21
SethRobertson I am only afraid that .svn might have been checked into git. If this did not happen then you are golden.00:21
tgunr no, it was not committed00:23
→ robotarmy joined00:23
← pr0wler left00:24
← johnw left00:25
← scelis left00:25
← DiaPo left00:28
→ k3yvn joined00:29
k3yvn → keyvan[mac]00:29
← benzap left00:30
← aah left00:30
→ bhosie joined00:31
← M1ghtyDOC left00:36
← jense_ left00:36
→ fogus joined00:37
→ jense joined00:37
← fogus left00:37
← jxpx777 left00:38
→ Chaser joined00:38
→ alberto56 joined00:39
← karihre left00:39
← jacobrelkin left00:39
← bhosie left00:39
→ karihre joined00:40
← albel727 left00:42
→ albel727 joined00:43
→ bkosborne joined00:43
bkosborne I pushed a tag to github with a missing file00:44
I'm trying to retag and repush the tag00:44
but I can't seem to get it right00:44
← Agos left00:45
wereHamster bkosborne: create the tag, git push -f origin tag <tag>00:45
→ normanrichards joined00:45
bkosborne i ran "git tag -a -f v1.2 d163fc9"00:45
→ mjf joined00:45
bkosborne then "git push github --tags"00:45
wereHamster however, keep in mind that it's usually not a good idea to rewrite a tag you've already pushed00:45
bkosborne yes ive read that00:46
wereHamster it's better to create a new tag00:46
bkosborne i need the -f in push eh00:46
amerine → amerine|away00:46
← Yuuhi left00:47
→ jrmuizel joined00:48
bkosborne wereHamster: I'm getting "everything up to date"00:48
but the tag is not updated on github00:48
hmm00:48
Counting objects: 1, done.00:48
Writing objects: 100% (1/1), 377 bytes, done.00:48
Total 1 (delta 0), reused 0 (delta 0)00:48
To [email@hidden.address]00:48
419571a..bfaf0f2 v1.2 -> v1.200:48
← keyvan[mac] left00:48
bkosborne that's what I got when i recreated the tag with -f, and then ran "git push github --tags"00:48
Arrowmaster i dont think you can use --tags00:50
wereHamster bkosborne: what about 'git push -f origin tag v1.2' ?00:50
← timepilot left00:50
← ki__ left00:51
bkosborne its still giving me everything up to date00:51
let me try deleting the tag and readding it00:51
wereHamster the tag is there!00:51
bkosborne yes00:51
but not with updated files00:51
wereHamster tag v1.2 is pointing to d163fc, just like you wanted00:52
bkosborne its missing the /js directory00:52
→ warlock_mza joined00:52
bkosborne oh god00:52
wereHamster "git tag -a -f v1.2 d163fc9"00:52
that's what you wanted00:52
bkosborne wait a sec00:52
d163fc9..f6750b4 master -> master00:52
does that mean i updated the master branch from d163fc9 to f6750b00:52
wereHamster yes00:52
bkosborne damn00:52
wereHamster or just look at the master branch on github00:53
how surprising, it shows up as f6750b400:53
bkosborne i read that quickly and thought it was just the beginning and end of the hex -_-00:53
← miloshadzic left00:53
wereHamster ranges are always <from>..<to>00:53
bkosborne see that now00:53
sorry for frustrating ya there lol00:53
all good now... has updated files00:54
thanks again wereHamster00:54
← tewecske left00:54
← mikepack left00:55
← bkosborne left00:55
← noahcampbell left00:57
← EricInBNE left01:01
tgunr Don;t know if this is a git think or a bash thing but when I try to clone a repo using git clone ssh://user@host/git/repo.git I receive the message bash: git-upload-pack: command not found01:01
I can ssh user@host just fine without a password01:01
← normanrichards left01:01
Arrowmaster sounds like git-upload-pack isnt in the path on the remote host01:01
→ EricInBNE joined01:02
← Heimidal left01:02
tgunr hmm, ok let me check01:02
← eletuchy left01:03
← vitor-br left01:03
tgunr ssh user@host echo $PATH shows the correct path: /opt/local/bin and which git-upload-pack shows /opt/local/bin/git-upload-pack01:04
→ QinGW joined01:04
← bmahe left01:05
→ jbrennan joined01:05
→ ForceBlast joined01:05
Eridius tgunr: the PATH you get in an interactive shell is different than in a non-interactive shell01:06
please tell me you actually tried `ssh user@host 'echo $PATH'` (e.g. with the single-quotes)01:06
otherwise, all you did was evaluate your local PATH, pass the string to the remote server, and tell it to echo it back to you01:06
wereHamster tgunr: 'ssh user@host which git-upload-pack'01:06
tgunr thanks Eridius, that was it01:07
Eridius fyi, `type` is a bash built-in that tells you precisely how the shell evaluates the command, wheras `which` does a search through PATH. In other words, `type` is a) faster and b) more accurate. Use it.01:07
tgunr: way to use ssh as an echo server ;)01:07
tgunr been a long day01:08
ayust Is there any way to get 'git archive' to also export the contents of submodules (or some other equivalent operation)?01:08
→ infinity0 joined01:09
→ miloshadzic joined01:09
Eridius ayust: not that I'm aware of. You could script it, by appending each subsequent archive into the same tar file with the submodule path prefixed, but I don't know of any prebuilt solution01:09
→ smolyn joined01:09
jbrennan I've got 2 computers both with clones of my git repo, and a server with a --bare repo that I push and pull from. I've got one branch, master. I pushed my changes to my server, but now when I pull into my master branch on my other machine, I get "error: untracketed working tree file: 'blah'. But git status doesn't show the file as being untracked on my machine. Any ideas how to solve this?01:09
infinity0 hey, what's the best way to use git-filter-branch to keep only certain paths from a repo01:10
→ vitor-br joined01:10
infinity0 "git rm" does the opposite, and it's hard to predict what was in the source tree earlier in the history, i need something like "git keep"01:10
Eridius jbrennan: maybe it's currently ignored, but it was added to master so the `git pull` would overwrite your untracked-and-ignored local file?01:10
jbrennan: try `git ls-files -o`. That should show you untracked files without using gitignore01:10
jbrennan Eridius: Yes it sounds possible. It's one of the new Xcode project file types I forgot to ignore originally on my first machine01:11
Eridius infinity0: you could create a new index file, add only the entries from the current index that you want, and then replace the current index with the new one01:11
← kenichi left01:11
infinity0 Eridius: how do i create a new index file and replace the old one with it? .git/index or something?01:11
hm, how do i edit that file?01:12
← delinquentme left01:12
wereHamster infinity0: git ls-tree, filter it out, nuke the index (rm .git/index) and load only the entries you want (git read-tree)01:12
Eridius jbrennan: you may be interested in using https://gist.github.com/785500 in your ~/.gitignore01:12
← tjholowaychuk left01:12
← JimmyRcom left01:12
→ cylence joined01:13
Eridius infinity0: you can set the GIT_INDEX_FILE env var on commands you want to affect the new index01:13
jbrennan Eridius: Yep, that's essentially what I'm using. But I hadn't updated my ignore file on my laptop before I added all the junk01:13
→ bmahe joined01:13
jbrennan So I think it's in that list of ignored files.. So how can I resolve this?01:13
Eridius infinity0: you could in fact dump the current index-info into a file, filter it to contain just the paths you want, then reset the current index and load up your filtered index-info01:13
← FilipeMaia left01:13
← justin-george left01:13
Eridius jbrennan: delete the file locally?01:13
jbrennan Eridius: OK, I'll give that a whirl01:14
← jrmuizel left01:15
jbrennan gawd Xcode project files really sully a git repo01:16
← rorist left01:16
jbrennan ironic, considering they've built in git support to xcode 401:16
Eridius just set up your ignores appropriately01:16
I set that once in .gitignore and have never had an issue01:16
of course, I also double-check what I'm checking in instead of blindly saying `git add -A && git commit` ;)01:16
← phantomcircuit left01:16
→ phantomcircuit joined01:17
← ForceBlast left01:17
jbrennan Haha, well yes, like I said, it was just an omission on my other machine, and Xcode has added new files between DPs I think01:17
→ ki__ joined01:17
jbrennan and now I've got to clean up it's mess. Oh well.01:17
← vitor-br left01:17
wereHamster git support in xcode4 is really poor01:18
Eridius yeah there are new files01:18
wereHamster any one of the recently released gui tools for mac is miles better01:18
jbrennan wereHamster: Yeah, I'm still using git from bash. I played with it in Xcode but yeah, it's still too crappy01:18
→ clarkfischer joined01:18
← jfkw left01:18
Eridius wereHamster: really? All of the recently-released mac tools rub me the wrong way. Mostly because they all seem to be implementing their display in webkit and have a lot of flashing and slow loading01:19
wereHamster I didn't say they were any good. Just better than what xcode4 has01:19
jbrennan I liked gitbox when it was in beta (but not enough to pay 40$ for it or however much)01:20
Kyso → KyleXY[2]01:20
jbrennan Not that it wasn't worth that much. Just not to me.01:20
Eridius for history, perhaps. Xcode 4's diff viewer is pretty nice though01:20
→ FilipeMaia joined01:20
jbrennan gitx is fairly nice when merging. But mostly I stick with the Terminal01:20
← kukks left01:21
→ quizme_ joined01:22
→ vitor-br joined01:22
← dpino left01:22
← clarkfischer left01:22
← muep left01:23
→ banisterfiend joined01:24
← banjiewen left01:24
banisterfiend is it just me or are git submodules a major headfuck01:24
← urbanmonk left01:24
SethRobertson I think I have suggested gitslave to you a few times due to that, though of course gitslave just inserts into a different body-part.01:25
← FilipeMaia left01:26
banisterfiend do ppl actually use submodules or avoid them?01:26
wereHamster most try to avoid them01:26
Eridius my current project has a nice tree of 16 submodules01:27
banisterfiend well github now has a 'wiki' feature where the wiki for your project is a separate repo01:27
i thought it was appropriate to add the wiki repo as a submodule?01:27
but reading up on submodules it seems like that's going to be a pain01:27
amerine|away → amerine01:28
Eridius unless you really think people who clone your repo need to be able to have really easy access to the wiki, it seems like a bad idea. Especially because you'd need commits on your main repo just to update the wiki link01:28
banisterfiend how would you guys recommend organizing the project wiki and the project? since they're in some sense connected01:28
wereHamster separate branches01:28
or even separate repos01:28
or simply have a directory with the wiki in it01:28
Eridius separate branches requires everyone who downloads one to also download the other. That may not be ideal01:29
banisterfiend can i have a separate repo as a subdirectory in my main repo or not?01:29
SethRobertson Really the only thing you need to do is tag the repo when you tag your release (assuming you work on a release concept)01:29
AAA_awright I wouldn't say submodules are hard, once you have it setup, but unless you are actively extending an existing Git repository (say, cgit with Git) I don't see a reason to use it. If you need a library idk maybe it does make sense to use submodules.01:29
Eridius AAA_awright: we're using submodules basically for libraries01:29
← bmahe left01:29
Eridius got a couple of libraries, they get developed in their own repo, and get pulled into the main one periodicially via submodule update01:29
→ bhosie joined01:29
SethRobertson It doesn't really make any sense to know what the state of the wiki is when you make a specific bugfix commit, unless you are using hte wiki as your ticket tracking system01:30
Eridius if the wiki is your active code documentation, it may make sense to keep it in sync, as it were, with the source. But it seems more useful to keep the wiki as user documentation rather than code documentation01:30
← cylence left01:31
banisterfiend Eridius: yes, so you would recommend separate repos? but can i have my wiki repo as a subdirectory under my main repo or is tha tnot possible? i want to somehow keep them together01:31
Eridius I don't know your specific use case, but it seems the simplest solution is to just keep separate repos01:32
← SirFunk_ left01:32
→ Zvpun joined01:35
→ gretch joined01:36
← miloshadzic left01:37
quizme_ http://www.pastie.org/147587201:37
Zvpun Hey, I am new to git and have used "git cvsimport -d $CVSROOT -C foo -r cvs -k foo" I thought (after reading git help cvsimport" that this creates some branch with the name cvs but there is only master. Futhermore "git config cvsimport.r cvs; git cvsimport" says: "fatal: bad config value for cvsimport.r" I appreciate the help of a person that understands what I am doing wrong.01:37
quizme_ I'm trying to checkout thrift-0.2.0.01:38
→ Mage joined01:38
quizme_ from github01:38
← bhosie left01:39
wereHamster quizme_: you're trying to clone the thrift repo and then checkout the 0.2.0 tag or branch01:39
pheaver hmm, i think the way you're using checkout, it means you're trying to create a local branch to track the remote branch, but it's not a branch, it's a tag01:39
wereHamster quizme_: there is no 'origin/tags/thrift-0.2.0'01:40
pheaver also, tags don't really have a notion of being remote or local. if the tag exists on the remote, the fetching will make it available to you, so "git checkout thrift-0.2.0" should work01:40
quizme_ wereHamster: like git checkout remotes/origin/0.2.x ?01:40
wereHamster there is a 'thrift-0.2.0' tag, but that one lives in refs/tags/thrift-0.2.001:40
quizme_: you could use that, it's a branch though, not a tag01:41
← mjf left01:41
quizme_ how do i checkout the tag?01:41
wereHamster quizme_: please run 'git branch -a' and 'git tag' to see which brnaches and tags are available01:41
quizme_ git checkout refs/tags/thrift-0.2.0 ?01:41
pheaver quizme: yes01:41
wereHamster quizme_: just like you did in the pastie01:41
quizme_ pheaver tyu01:41
wereHamster ok01:41
pheaver quizme: git checkout thrift-0.2.0 would work too, if it's unambiguous01:42
quizme_ thnx01:42
necromancer → necromancer`bzzy01:42
pheaver note that you won't be on a branch when you do that, so if you commit now, they'll be dangling or whatever01:42
quizme_ http://www.pastie.org/147588401:42
git checkout thrift-0.2.001:43
yeah baby01:43
pheaver :p01:43
→ jrmuizel joined01:43
quizme_ thanks guys01:44
you collectively rock01:44
→ bmahe joined01:45
← BiggFREE left01:45
banisterfiend riiight01:47
how do i delete a submodule01:47
← alvaro_o__ left01:48
→ xiangfu joined01:48
→ pheaver_ joined01:49
banisterfiend ??01:49
→ zombor joined01:49
banisterfiend there doesnt seem to be a 'git submodule delete' or 'git submodule rm' command01:49
← pheaver_ left01:49
banisterfiend Eridius ping01:49
← zombor left01:50
→ albel727_2 joined01:50
← albel727 left01:50
← airborn left01:52
SethRobertson My understanding is that it is entirely manual. Manipulate the .gitsubmodules file, delete the subdirectory, and see what git status says.01:52
Renaming is even more painful01:52
← jrmuizel left01:52
SethRobertson Or....if you have not done much else, just rebase those commits away01:52
← pheaver left01:53
← xiangfu left01:53
← ka1ser left01:53
→ morninj joined01:54
← Chaser left01:54
← morninj left01:54
→ normanrichards joined01:55
→ fivetwentysix joined01:56
→ laj joined01:58
← warlock_mza left01:59
Eridius banisterfiend: pong02:00
← vitor-br left02:00
← albel727_2 left02:00
→ albel727_2 joined02:00
Eridius banisterfiend: you can just `git rm` to remove it from the index, and you can edit the .gitmodules file to remove its entry entirely. The dir itself will remain, though, as git will refuse to delete an entire other repo02:00
but of course you can rm -rf it02:00
→ jrmuizel joined02:01
→ justin-george joined02:01
→ frakturfreak_ joined02:01
← Schmallon left02:02
← brandFU left02:03
→ kukks joined02:04
→ xiangfu joined02:05
← frakturfreak left02:05
→ ChrisPitzer joined02:05
→ pgib joined02:06
← Silanus_ left02:06
← sraue left02:07
→ albel727 joined02:07
ChanServ set mode: +v02:07
ChrisPitzer does anyone know of an irc bot that will shout out commit messages given a git repo? (not on github)02:07
→ Silanus joined02:07
← engla left02:08
nfiniteset i see that in the #coffeescript room. might ask there02:08
Eridius ChrisPitzer: how about CIA?02:08
ChrisPitzer not open source.02:08
→ albel727_3 joined02:08
← frakturfreak_ left02:08
← albel727 left02:08
ChrisPitzer Eridius: i need a bot i can plug in to a private repo.02:09
Eridius why does the bot itself need to be open source?02:09
oh, you just want to run your own02:09
ChrisPitzer isn't CIA only for public codebases?02:09
Eridius I wrote one years ago in ruby. It's pretty easy02:09
I don't know if CIA cares about the source project, it just relies on a hook on the repo to send it the info to shout out02:09
womble ChrisPitzer: kgb.02:09
ericholscher ChrisPitzer: http://code.google.com/p/cia-vc/source/browse/ ?02:09
→ frakturfreak_ joined02:09
→ banjiewen joined02:10
← albel727_2 left02:10
Eridius of course CIA also documents the commit messages on its website. I don't know if there's any way to turn off that part02:10
← banjiewen left02:11
← ISF left02:12
→ eletuchy joined02:12
→ ISF joined02:12
qDot_ → qDot|Away02:12
← Guest21165 left02:12
← eletuchy left02:13
← bmahe left02:13
← smolyn left02:13
→ dl9pf joined02:14
← fgro left02:14
← dl9pf_ left02:15
ChrisPitzer yeah - we want to use this for client work... so that wouldn't fly.02:15
← noel_v left02:15
infinity0 how do you ls-tree teh index?02:16
← Zvpun left02:16
→ bmahe joined02:16
ChrisPitzer ericholscher: we could set up our own cia-vc instance... but that seems like a lot for just an irc bot... have you set it up before? is it particularly painless?02:16
womble ChrisPitzer: Just use kgb.02:17
A lot less infrastructure involved.02:17
← robotarmy left02:18
ChrisPitzer womble: i'm going to give it a shot. thanks.02:19
← jaeckel left02:20
→ robotarmy joined02:21
← robotarmy left02:21
← oriba left02:22
→ fivetwentysix_ joined02:22
← ISF left02:22
← alberto56 left02:23
→ sraue joined02:23
← amerine left02:23
→ ISF joined02:24
← fivetwentysix left02:24
fivetwentysix_ → fivetwentysix02:24
→ GodEater_ joined02:25
← ISF left02:27
→ jaeckel joined02:27
← jaeckel left02:27
→ jaeckel joined02:27
← GodEater left02:28
ChrisPitzer womble: i'm not having a great time finding docs on how to set up kgb bot... any suggestions?02:28
← dstorrs left02:29
womble ChrisPitzer: apt-get install kgb-bot kgb-client-git; less /usr/share/doc/kgb-*/README* will probably get you somewhere02:30
← Vampire0 left02:30
→ jfkw joined02:30
← pdelgallego left02:31
→ ISF joined02:31
← nfiniteset left02:33
← ISF left02:33
→ gnufied joined02:34
→ ckrailo joined02:34
→ ISF joined02:34
← ISF left02:35
→ shentino joined02:35
← klunky left02:35
→ webPragmatist joined02:35
← webPragmatist left02:35
→ webPragmatist joined02:35
→ ISF joined02:35
← ISF left02:36
→ klunky joined02:36
← Evious left02:38
→ SirFunk_ joined02:38
← spearce left02:38
← jvv left02:39
→ eletuchy joined02:39
→ ISF joined02:39
infinity0 what's the most efficient way to get a commit's tree's sha1?02:40
→ dfr|mac joined02:41
← leavittx left02:41
Eridius infinity0: probably `git rev-parse $SHA^{tree}`02:42
→ legumbre joined02:42
infinity0 ah thanks02:42
→ Vampire0 joined02:43
→ Silanus_ joined02:44
← dermoth_ left02:45
← Silanus left02:45
→ eletuchy_ joined02:47
← eletuchy left02:47
← ISF left02:48
qDot|Away → qDot_02:48
→ robotarmy joined02:48
Chillax → Chillaway02:50
→ themarmot joined02:50
→ ISF joined02:50
← necromancer`bzzy left02:51
themarmot i have a repo that includes about 30 submodules coming in from different developers, however, one developer has gone rogue and decided to make his master repo under 3.0/master instead of just master ... so updating submodules requires updating each one individually? or is there a way to update them all accordingly(previous command i used was git submodule foreach git pull origin master)02:52
Eridius why would you use that foreach command? What's wrong with `git submodule update --recursive`?02:53
oh wait, you're trying to update the commit so you can commit in the parent?02:53
themarmot yes02:53
Eridius in any case, just use `git pull` instead and let the repo itself define what its upstream branch is namd02:53
← peshalto left02:54
themarmot tried that and i get error: pathspec 'git' did not match any file(s) known to git.02:54
n/m, i left out foreach in the command02:54
← ISF left02:56
→ smolyn joined02:57
ChrisPitzer womble: thanks - i'm getting it worked out.02:58
→ dermoth_ joined03:00
→ CamC joined03:00
← beatak left03:01
banisterfiend how do i delete pages from my github wiki? (sorry bug #github is dead)03:03
but*03:03
← bulbous left03:03
← Elfe left03:04
→ Elfe_ joined03:04
→ jvv joined03:04
Eridius banisterfiend: I don't know offhand, but I can guess - check out the wiki itself as a git repo, edit it by hand?03:05
ayust banisterfiend: go to the Git Access header, clone the wiki's repository, and git rm?03:05
(in general, the url is [email@hidden.address]03:06
infinity0 is there a tool that linearises git history?03:06
Eridius infinity0: rebase?03:06
infinity0 ie. given 2 branches, go through the commits in turn and rebase all the commits in chronological order?03:06
Eridius trying to intermingle the commits from two branches together has a rather high risk of introducing spurious conflicts03:07
ayust that... would have very bad implications for merges03:07
Eridius with a regular rebase, the only conflicts you should get are ones where the merge itself had a conflict (I think)03:07
banisterfiend ayust: is taht the only way? wtf?03:07
ayust: what about through the web interface?03:07
infinity0 Eridius: i've ran filter-branch on the repo to restrict the commits only to a small subdir03:08
we're trying to factor that out into a separate library03:08
ayust banisterfiend: github's wiki isn't all that advanced, maybe make them a ticket to add one03:08
infinity0 but --prune-empty didn't get rid of the empty merge commits03:08
Eridius what's an "empty merge commit"?03:09
either it has 2+ parents, at which point you can't remove it, or it has 1 parent, at which point it's not a merge03:09
infinity0 a commit with >1 parents with a tree equal to the most recent common ancestor of the parents03:10
← smolyn left03:10
ayust you could run another filter-branch with the --commit-filter option03:11
and define a filter that skips such commits03:11
infinity0 right, so i need to write that manually, ok03:12
← newbiz left03:12
→ newbiz joined03:12
infinity0 is there a git utility for getting the common ancestor of a series of commits?03:12
Eridius merge-base?03:12
← QinGW left03:13
infinity0 ah, thanks03:13
ayust ^03:13
pgib you are all so smart03:13
← newbiz left03:13
← gnufied left03:13
→ newbiz_ joined03:13
→ Schmallon joined03:15
→ killerchicken_ joined03:15
← jbrennan left03:16
→ envi joined03:16
→ peshalto joined03:16
→ QinGW joined03:17
ChanServ set mode: +v03:17
← newbiz_ left03:17
→ newbiz joined03:17
CamC hi, using TortoiseGit for the first time, tut I am following says "right click folder, select Git clone, but that option is not there, any ideas?03:18
← newbiz left03:19
← killerchicken left03:19
→ newbiz joined03:19
→ mahata joined03:19
← newbiz left03:20
→ newbiz joined03:20
← ph^ left03:20
← newbiz left03:21
← ki__ left03:22
dcope exit03:23
ayust wrong buffer, dcope ;)03:23
dcope ayust :-D03:24
← dcope left03:24
→ FilipeMaia joined03:24
← Amorphous left03:25
→ ph^ joined03:25
infinity0 --commit-filter isn't passing any args to my script, any ideas?03:29
i have echo >&2 ARGS "$@" and it just prints out "ARGS"03:29
← hagabaka left03:31
→ intripoon joined03:32
→ pheaver joined03:33
infinity0 oh, you need to specify <command> as "command $@"03:34
that's confusing as hell :|03:34
← intripoon_ left03:35
→ hagabaka joined03:35
← hagabaka left03:35
→ hagabaka joined03:35
Eridius infinity0: it's confusing that you have to pass the command as a single argument?03:36
if you didn't, how would filter-branch know where your command ended?03:37
→ frooh joined03:37
infinity0 no, it's confusing that you have to pass the argument as "command $@"03:37
scp1 Is it possible to use git with hg like git svn?03:37
← helv12 left03:37
frooh how can I do git show refs/tags/foo and *only* get tag information? The idea being automated querying of the authorname etc03:37
infinity0 with the $@ as part of the single argument03:37
also, this is undocumented03:38
Eridius infinity0: the command is executed with `sh -c "$command" git-commit-filter $ARGS` in essence03:38
infinity0 so it's not really a command, it's a shell expression03:38
Eridius which means your command gets the arguments in the variable $@, just as if it were a shell function03:38
infinity0: the command you said you were using included "$@". Why did you type that if you weren't really using it?03:38
→ roop joined03:38
infinity0 what do you mean03:38
i'm writing a script to use for commit-filter03:39
it needs to take arguments03:39
Eridius you said you typed: echo >&2 ARGS "$@"03:39
that, as a command, looks fine. It has the "$@" and everything.03:39
now you're saying you didn't type the $@03:39
infinity0 that's inside my script03:39
Eridius oooh, but your script was just referenced by name. I gotcha03:39
← jaql left03:40
→ Amorphous joined03:40
→ noahcampbell joined03:41
→ alberto56 joined03:41
← ckrailo left03:42
← ph^ left03:43
← peshalto left03:43
← SimonP86_ left03:44
← alberto56 left03:45
→ ki__ joined03:46
→ gnufied joined03:47
frooh breaks out perl03:48
ianmcorvidae|alt → ianmcorvidae03:49
← eletuchy_ left03:49
→ eletuchy joined03:50
offby1 I think the tradition thing is to break out Emacs.03:51
At least, "The Social Network" says so03:51
← eletuchy left03:51
← chrisshattuck left03:52
→ eletuchy joined03:52
→ smolyn joined03:52
→ timj__ joined03:52
← verto left03:53
← pheaver left03:54
→ tty1 joined03:55
← QinGW left03:55
→ chrisshattuck joined03:55
→ towski joined03:55
→ QinGW joined03:56
← timj_ left03:56
infinity0 lol03:58
← quizme_ left03:58
infinity0 did it mention it was free software and any of that stuff?03:58
← qDot_ left03:59
offby1 SURE it did. Even said "GNU/Linux" instead of "Linux".03:59
offby1 whistles innocently03:59
infinity0 shame, i would seriously pay £10 for a cinema ticket just to hear a mainstream film use the phrase "GNU/Linux"04:00
← LongBeach left04:01
← shentino left04:01
→ glphvgacs_ joined04:02
← gnufied left04:03
womble More importantly, did they pronounce "Linux" correctly?04:04
← kukks left04:04
← roop left04:04
← madewokherd left04:05
← glphvgacs_ left04:05
→ glphvgacs_ joined04:05
← noahcampbell left04:06
glphvgacs_ is there something built into git that would let me pull only directories on the kernel that are relevant to my .config file?04:06
→ madewokherd joined04:06
← booginga left04:07
glphvgacs_ a bare initial pull from kernel.org is something around 500mb, i imagine i would be able to pull 1/3 of this, if git looked into .config file and then decided what is relevant to fetch04:08
← madewokherd left04:08
→ madewokherd joined04:09
← HopsNBarley left04:11
offby1 glphvgacs_: I think it's possible to pull a subset of _history_, but not of the _tree_04:13
and I don't even know how to do the former.04:13
→ jhoepken joined04:13
→ cylence joined04:13
→ pheaver joined04:14
glphvgacs_ offby1: what i'm looking for would be a script or something, made costume to work with kconfing04:14
← gregglind left04:14
glphvgacs_ so script would look into .config, figure out (based on the enabled CONFIG_*) what to pull and take it from there04:15
→ jony joined04:15
Zao glphvgacs_: I really hope there's nothing polluting the Git codebase that's specific to the layout of the Linux Kernel and its config files.04:16
glphvgacs_ so instead of pulling then configuring the kernel, one would configure the kernel (ok pull whatever is needed for config first) and then pullonly relevant directories and subdirectories04:16
Zao: why not?04:17
offby1 glphvgacs_: I would be surprised if that were possible.04:17
I think git only thinks in terms of complete trees.04:17
Zao glphvgacs_: Because Git is a general purpose VCS, and such functionality most definitely belongs in other-party scripts.04:17
glphvgacs_ wouldn't that take some of the loads off the server and also speed up the whole thing?04:17
Zao Which you mentioned in the join/quit spam above, which I saw after my statement.04:17
offby1 server load and speed don't seem to be big problems in git-land04:17
← webPragmatist left04:18
← mdel left04:19
← Rockj left04:19
glphvgacs_ offby1: well, i just sat through 15 minutes of pulling from linus just to get his warning that rsync was doing something slowly and so i had to do it again (for another 15min), to pull 300mb of code. while what i need is just 1/3 of that04:20
offby1 *shrug*04:20
glphvgacs_ i understand that git is not kernel specific and is mean to...04:21
→ Rockj joined04:21
tango_ glphvgacs_: needing just 1/3rd of the linux kernel source has very little to do with git, I'm afraid04:21
glphvgacs_ but i was wondering if someone here would be aware of a script or a module for git that would do what i have in mind04:21
tango_ (although 500mb of initial pull ... I was not aware of the current size of the repo)04:22
glphvgacs_ tango_: i'm getting stuff for hardwares i don't have, every rc cycle which is other freaking week04:25
tango_ glphvgacs_: if you don't care about the history but only care about the snapshot you could clone with depth 104:26
glphvgacs_: that would be about as cheap as getting the tarball04:26
→ roo joined04:26
tango_ glphvgacs_: but even with a tarball you'd get all hardware stuff04:26
glphvgacs_: anyway, I doubt what you propose could be done, unless the kernel repo itself was split into submodules04:27
← bmahe left04:27
glphvgacs_ tango_: isn't it now?04:28
tango_ no04:28
glphvgacs_ how would one go about this?04:28
tango_ glphvgacs_: you wouldn't04:29
those creating and managing the upstream repo would04:29
glphvgacs_ it would sound like something i should ask people at kernelhq04:29
tango_ no, it's something you should ask linus04:29
and my bet is the answer is going to be "no we won't"04:29
← dreiss left04:30
glphvgacs_ i actaully have had this idea for a while and linus just responded to someone on mailin list, explaining how to get a shortlog of the "directories and subdirectories" of one's interest, using git.04:32
→ cwebber joined04:32
glphvgacs_ then i thought if it is possible to track what's changed and what's not then there should be a way to tag-then-track to see a new -rc doesn't even pretain to my machine and so on04:33
→ phantomcircuit_ joined04:34
glphvgacs_ ok, 38-rc1 is begging a reboot, take care you all04:34
← jhoepken left04:34
← glphvgacs_ left04:34
← Risar` left04:35
← phantomcircuit_ left04:35
← phantomcircuit left04:35
→ Risar joined04:36
ChrisPitzer has anyone here set up kgb-bot before?04:37
the man pages aren't quite getting me there...04:37
→ tstclair joined04:37
→ phantomcircuit joined04:38
← ceej left04:40
← cwebber left04:41
← chrisshattuck left04:43
← sivy left04:44
← masonkramer left04:47
→ pratz joined04:48
→ devanjan joined04:48
← Error404NotFound left04:48
→ cwebber joined04:49
← thiago_ny left04:51
← jfkw left04:56
→ jacobrel_ joined05:00
→ gnufied joined05:02
→ chrissbx joined05:02
→ ajpiano joined05:02
chrissbx Hello. I've got a working directory which contains a symlink for one directory, and I want Git to treat the symlink just like the target directory (follow it).05:03
I cannot find the right config option for this. Anyone could tell me?05:04
→ psankar joined05:06
← psankar left05:06
→ psankar joined05:06
← jony left05:06
→ Kharza joined05:09
← jrmuizel left05:09
→ amerine joined05:10
← tty1 left05:10
→ tty1 joined05:11
← jbd left05:12
← cylence left05:12
← homa_ran1 left05:15
→ jrmuizel joined05:15
← montylounge left05:15
← jrmuizel left05:15
→ gregglind joined05:15
← Eridius left05:16
→ webPragmatist joined05:17
← webPragmatist left05:17
→ webPragmatist joined05:17
webPragmatist hey all … how can i serve git with nginx over http? does it support webdav?05:17
← webPragmatist left05:17
→ homa_rano joined05:18
→ speedxcore joined05:18
→ Adaptee joined05:22
→ dissidentss joined05:22
offby1 here's a pointless challenge: how do I find the oldest commit in my repo that has more than one "child"?05:23
AAA_awright The oldest branching commit? hm05:25
offby1 yeah05:25
I'm not even sure I can find the oldest _merge_ commit.05:25
I'd'a thought https://gist.github.com/785734 would do it, but I suspect it's missing some05:26
AAA_awright offby1: Find the oldest common ancestor between all the branches05:27
offby1 hmm05:27
clever indeed.05:27
AAA_awright Or rather, the newest common ancestor05:27
oldestCommonAncestor(branch1, oldestCommonAncestor(branch2, oldestCommonAncestor(branch3, branch4)))05:27
I forget the command but there is one05:28
http://www.kernel.org/pub/software/scm/git/docs/git-merge-base.html05:28
Ah yeah just supply all your branches05:28
offby1 well ... https://gist.github.com/785737 yields a very recent commit05:28
AAA_awright I think05:28
offby1 interesting idea though; I'm probably just doing something wrong05:29
AAA_awright Given three commits A, B and C, git merge-base A B C will compute the merge base between A and a hypothetical commit M05:29
...which is a merge between B and C. For example, with this topology05:29
offby1: You should only supply 2 at a time then05:29
So, iterate?05:29
offby1 I guess05:29
← tty1 left05:30
AAA_awright offby1: Actually try --octopus05:30
offby1 huh05:31
also yields a very recent commit05:31
er, to which command should I pass the "--octopus" option?05:31
AAA_awright hmm maybe no switch does it05:31
git merge-base05:31
You're looking at the man page right?05:31
offby1 ya05:32
fr0sty-away log --raw in date order. save a dictionary of parents, find the first duplication.05:32
← madewokherd left05:32
offby1 fr0sty-away: hmm05:32
offby1 ponders05:32
→ HopsNBarley joined05:32
← HopsNBarley left05:32
→ HopsNBarley joined05:32
offby1 pity there's no way to do that in bash05:33
→ aah joined05:33
fr0sty-away unless I'm inventing options that don't exist, that just might work. maybe --reverse05:33
offby1 oh, there's a --reverse all right, Jerry.05:33
AAA_awright "git merge-base finds best common ancestor(s) between _two_ commits to use in a three-way merge" emphasis mine05:33
fr0sty-away perl is your friend, and nearly as universal.05:33
offby1 AAA_awright: yeah05:33
fr0sty-away: actually perl is my mortal enemy, but never mind :)05:33
AAA_awright How many branches are we talking?05:33
offby1 sitaram seems to like it OK05:33
→ Error404NotFound joined05:33
fr0sty-away though some really creative use of sort |uniq would work along with -nX to git log05:33
AAA_awright Try it by hand just to be sure it works05:34
offby1 AAA_awright: actually I'd be happy to just scan _one_ branch.05:34
→ maleknet joined05:34
← ptl left05:34
offby1 fr0sty-away: your ideas intrigue me. *mumble* newsletter.05:34
fr0sty-away happy to help :-)05:34
offby1 I like Python, but running subprocesses is ... tedious.05:35
→ fr0sty-away_ joined05:35
← fr0sty-away left05:38
AAA_awright offby1: It should be the same, the only reason processing one branch would vary from all of them is if you have a branch that shares no common commits, which is rare (but done, like for documentation sometimes)05:39
What's it called05:39
→ zivester joined05:39
AAA_awright I forget05:39
→ Utkarsh__ joined05:39
← gnufied left05:39
AAA_awright offby1: If you know the branch shares any commits it should have the same result05:39
*another branch shares05:40
→ ptl joined05:40
killerchicken_ charon: thanks again for your ml suggestion, I see it sparked quite a discussion :)05:41
→ carutsu_ joined05:42
→ jony joined05:42
← Utkarsh_ left05:43
← carutsu left05:43
→ ThomasLocke joined05:43
AAA_awright offby1: C="master"; for b in $(git for-each-ref | awk '$2 == "commit" {print $3}' | xargs); do C=$(git merge-base $C $b); done; echo $C;05:44
→ roop joined05:44
AAA_awright That seems to work for me05:45
→ mdel joined05:45
← dissidentss left05:46
→ gnufied joined05:46
Unknown[NF] → Unknown[OFF]05:49
→ muep joined05:49
offby1 https://gist.github.com/785756 seems to work05:50
AAA_awright: I'll try yours now05:51
why "| xargs" with no arguments to xargs?05:52
← muep left05:52
offby1 AAA_awright: nevertheless, it seems to have worked! Thanks.05:53
AAA_awright offby1: To de-listify it05:53
offby1 oh, interesting05:53
→ webPragmatist joined05:53
← webPragmatist left05:53
→ webPragmatist joined05:53
AAA_awright maybe there's a -0 option but I'm too lazy to check05:54
offby1 aw reet!05:54
← ormaaj left05:54
webPragmatist sorry05:55
← speedxcore left05:55
webPragmatist anyway… hey guys where can i find out how to setup git and nginx for http clone and push?05:55
AAA_awright webPragmatist: The Nginx docs?05:55
webPragmatist eh?05:55
→ ormaaj joined05:56
offby1 *shrug*05:56
→ speedxcore joined05:57
webPragmatist AAA_awright: suggestions ? does git support webdav? should i fastcgi through git-core or somethin?05:57
→ muep joined05:57
AAA_awright webPragmatist: Yes Git supports WebDAV, what would you use fastcgi for?05:57
webPragmatist rather the proxy pass? i dunno… just trying to get a direction really05:58
AAA_awright webPragmatist: Did you try looking at the documentation to see what WebDAV support there is for NginX?05:59
← yhager left05:59
AAA_awright http://wiki.nginx.org/HttpDavModule looks relevant05:59
webPragmatist: Also, why not SSH?05:59
→ doener joined05:59
webPragmatist there's a module but i'd rather not use webdav if theres other options?05:59
is that the only option for write?05:59
over http05:59
offby1 write works over https06:00
AAA_awright Same thing inside an SSL wrapper06:00
webPragmatist: Do you know what you're trying to do?06:01
webPragmatist AAA_awright: probably not06:01
right now i think i am just going to use git over ssh06:02
← towski left06:02
→ yhager joined06:02
← jacobrel_ left06:02
AAA_awright That's the recommended option especially if you already have an SSH account06:02
← Exposure left06:03
← psankar left06:04
← Tommy[D] left06:07
← robotarmy left06:08
← yhager left06:08
→ Tommy[D] joined06:09
banisterfiend offby1: what advantages do u think git has to bzr06:09
nevyn it's a 2nd system of the distributed version control problem.06:11
→ Exposure joined06:11
← mahata left06:12
→ johnw joined06:13
← TheMoonMaster left06:15
← cwebber left06:17
← OOPMan left06:19
→ napster_123 joined06:21
→ ianmcorvidae|alt joined06:22
→ offby1` joined06:23
← offby1` left06:23
→ offby1` joined06:23
→ fr0sty-away joined06:23
→ Tuomas_ joined06:24
← offby1 left06:24
offby1` → offby106:24
← werdan7 left06:25
→ SvenDowideit_ joined06:25
← muep left06:25
← GodEater_ left06:25
← adamm left06:25
← bauruine left06:25
← Ratler left06:25
← Gunni left06:25
← hegge left06:25
← HopsNBarley left06:25
← QinGW left06:25
← Elfe_ left06:25
← themarmot left06:25
← MacGyverNL left06:25
← kpreid left06:25
← NickyP left06:25
← comcor- left06:25
← Juri_Agarin left06:25
← cbz left06:25
← stepnem left06:25
← Tronic left06:25
← frooh left06:25
← multi_io left06:25
← lirakis left06:25
← ecksun left06:25
← j0ran left06:25
← metze left06:25
← Risar left06:25
← envi left06:25
← dfr|mac left06:25
← jaeckel left06:25
← alip left06:25
← theoros left06:25
← ymasory_ left06:25
← rootard left06:25
← m0 left06:25
← funfun left06:25
← womble left06:25
← mcjb left06:25
← halfline left06:25
← dean[w] left06:25
← tangledhelix left06:25
← blalor_afk left06:25
← mrtazz left06:25
← CamC left06:25
← alnewkirk left06:25
← an0maly left06:25
← ayust left06:25
← jdav_gone left06:25
← squentin left06:25
← qknight left06:25
← RandomSort left06:25
← tro left06:25
← Khisanth left06:25
← ormaaj left06:25
← Utkarsh__ left06:25
← aah left06:25
← Adaptee left06:25
← FilipeMaia left06:25
← jvv left06:25
← albel727_3 left06:25
← angelsl left06:25
← Bombe left06:25
← jksz left06:25
← Swimming_bird left06:25
← saurik_ left06:25
← Arrowmaster left06:25
← orafu left06:25
← kandinski left06:25
← m_3 left06:25
← avar left06:25
← Derander left06:25
← Purdy left06:25
← divVerent left06:25
← robinsmidsrod left06:25
← elly left06:25
← robinbowes left06:25
← sven^ left06:25
← johnw left06:25
← smolyn left06:25
← frakturfreak_ left06:25
← justin-george left06:25
← EricInBNE left06:25
← rhl left06:25
← marc_kdab left06:25
← bobhope left06:25
← Guest48246 left06:25
← [exa] left06:25
← zorzar left06:25
← jjuran left06:25
← quadz left06:25
← kelvie left06:25
← markatto left06:25
← resixian left06:25
← blathijs_ left06:25
← tmokros left06:25
← ivan left06:25
← spb left06:25
← esc left06:25
← rudi_s left06:25
← phishy left06:25
← sejo left06:25
← lotyrin left06:25
← mjv left06:25
← pasky left06:25
← mdel left06:25
← fr0sty-away_ left06:25
← timj__ left06:25
← banisterfiend left06:25
← Mannequin left06:25
← fr| left06:25
← Soliton left06:25
← drizzd left06:25
← SvenDowideit left06:25
← gebi left06:25
← bran left06:25
← quizme left06:25
← nevyn left06:25
← ricky left06:25
← bgerber left06:25
← ianmcorvidae left06:25
← RSchulz left06:25
← piti left06:25
← homa_rano left06:25
← sunn` left06:25
← agruman left06:25
← Unknown[OFF] left06:25
← orospakr left06:25
← tlonim left06:25
← jds left06:25
← fr0sty left06:25
← ilteris left06:25
← Teo` left06:25
← dmlloyd left06:25
← sycofly left06:25
← teuf left06:25
← drry left06:25
← Tuomas left06:25
→ homa_rano joined06:25
→ Ingmar joined06:25
← Ingmar left06:25
→ Ingmar joined06:25
→ RSchulz joined06:26
→ ricky joined06:26
← MikhX left06:26
→ Soliton joined06:27
→ Unknown[OFF] joined06:27
→ werdan7 joined06:27
→ boombatower joined06:28
← kelvie_ left06:28
→ orospakr joined06:28
→ kelvie_ joined06:29
→ quizme joined06:29
→ bran joined06:30
→ ilteris joined06:31
→ jds joined06:31
→ nevyn joined06:32
→ HG` joined06:32
→ agruman joined06:32
→ dmlloyd joined06:33
← phantomcircuit left06:34
← zivester left06:34
→ bgerber joined06:34
→ dreiss joined06:36
← Exposure left06:39
→ Exposure joined06:42
→ jhoepken joined06:44
→ mastro joined06:45
webPragmatist hey all06:46
← hobodave left06:46
webPragmatist i'm trying to take symfony 1 and make a git repo out of their svn06:46
→ _ollie joined06:47
webPragmatist i'm wondering how i can handle svn externals intelligently06:47
← werdan7 left06:47
→ ivan`` joined06:48
offby1 not sure you can :-|06:51
← pheaver left06:53
← Yuffster left06:54
→ beneg225 joined06:54
webPragmatist offby1: should i just store the svn inside of git?06:55
i mean this is code i rarely use06:55
or rather change06:55
the solutions is to go through each external and convert it to it's git equivalent06:55
ivan`` → ivan06:56
← legumbre left06:56
← hyperair left06:57
webPragmatist http://yergler.net/blog/2009/07/21/git-svn-and-svnexternals/06:57
spiffy06:57
→ hyperair joined06:58
→ avinashhm joined06:59
→ Bass2 joined07:01
← Bass10 left07:02
← avinashhm left07:05
webPragmatist offby1: how useful is this i wonder though07:05
it's not like you can do gsc update07:06
← Naikrovek left07:06
→ krosenvold_ joined07:07
webPragmatist rather07:08
you lose the ability to rebase07:08
← _ollie left07:09
→ mdel joined07:09
→ Rabbitbunny joined07:10
→ cbreak_work joined07:10
→ Naikrovek joined07:12
WebDragon → WebDragon|away07:12
→ TrineLocke joined07:13
← quizme left07:13
→ blaenk joined07:13
← eletuchy left07:15
→ aah joined07:15
← webPragmatist left07:18
← roo left07:21
← LouisJB left07:25
→ Zhonghua joined07:25
← blaenk left07:25
← beneg225 left07:27
← Bass2 left07:28
← Rickardo1 left07:30
→ SIGBUS joined07:36
→ sonnym joined07:37
→ Matthewds joined07:39
← Matthewds left07:39
← dreiss left07:42
← gnufied left07:42
→ saschpe joined07:44
→ gnufied joined07:44
→ Theravadan joined07:44
← coldboot left07:44
→ eletuchy joined07:46
← klunky left07:47
→ steph021 joined07:48
← Mage left07:50
← eletuchy left07:53
← mdel left07:53
→ avinashhm joined07:53
→ skoop joined07:54
→ hobodave joined07:54
← vdv left07:54
→ Titosemi joined07:54
→ coldboot joined07:55
← roop left07:55
→ drl joined07:55
→ rapidmax joined07:57
→ _ollie joined07:58
→ Badiss joined08:00
← cannonball left08:01
← Zhonghua left08:01
→ raichoo joined08:02
← dvaske left08:02
← SvenDowideit_ left08:03
→ zheng joined08:04
→ serard joined08:05
→ mdel joined08:05
← sftp left08:08
→ sftp joined08:08
→ OOPMan joined08:08
→ roop joined08:09
← acid left08:10
← andrewvos left08:11
→ andrewvos joined08:11
→ workmad3 joined08:12
→ Veejay joined08:13
→ thierryp joined08:14
→ clarkfischer joined08:16
← Plaidrab left08:17
← Schmallon left08:23
→ Stem joined08:23
→ dr_win joined08:23
← Stem left08:25
← drl left08:25
→ BiggFREE joined08:26
→ redtricycle joined08:27
→ mahata joined08:27
→ mikepack joined08:27
→ teweWork joined08:27
→ SvenDowideit joined08:27
→ frenzz joined08:27
→ EnginAy joined08:27
→ ph^ joined08:27
→ ludde joined08:27
→ bauruine joined08:27
→ apalto joined08:27
→ WAACAYX joined08:27
→ opalka joined08:27
→ tewecske joined08:27
→ matjas joined08:27
→ dc5ala joined08:27
→ yoh joined08:27
→ pasky joined08:27
→ mjv joined08:27
→ lotyrin joined08:27
→ sejo joined08:27
→ phishy joined08:27
→ rudi_s joined08:27
→ esc joined08:27
→ spb joined08:27
→ blathijs_ joined08:27
→ resixian joined08:27
→ markatto joined08:27
→ tmokros joined08:27
→ kelvie joined08:27
→ quadz joined08:27
→ jjuran joined08:27
→ zorzar joined08:27
→ Guest48246 joined08:27
→ [exa] joined08:27
→ bobhope joined08:27
→ EricInBNE joined08:27
→ frakturfreak_ joined08:27
→ johnw joined08:27
→ LongBeach joined08:27
→ gebi_ joined08:27
→ sycofly joined08:27
→ Khisanth joined08:27
→ tro joined08:27
→ RandomSort joined08:27
→ qknight joined08:27
→ squentin joined08:27
→ jdav_gone joined08:27
→ ayust joined08:27
→ an0maly joined08:27
→ alnewkirk joined08:27
→ Mannequin joined08:27
→ drry joined08:27
→ hegge joined08:27
→ Gunni joined08:27
→ Ratler joined08:27
→ GodEater_ joined08:27
→ muep joined08:27
→ fr0sty joined08:27
→ ch077179 joined08:27
→ drizzd joined08:27
→ sambarino joined08:27
→ ormaaj joined08:27
→ Utkarsh__ joined08:27
→ HopsNBarley joined08:27
→ Adaptee joined08:27
→ Risar joined08:27
→ QinGW joined08:27
→ timj__ joined08:27
→ frooh joined08:27
→ FilipeMaia joined08:27
→ envi joined08:27
→ jvv joined08:27
→ Elfe_ joined08:27
→ themarmot joined08:27
→ jaeckel joined08:27
→ albel727_3 joined08:27
→ alip joined08:27
→ angelsl joined08:27
→ MacGyverNL joined08:27
→ kpreid joined08:27
→ Bombe joined08:27
→ theoros joined08:27
→ Purdy joined08:27
→ NickyP joined08:27
→ jksz joined08:27
→ Swimming_bird joined08:27
→ ymasory_ joined08:27
→ fr| joined08:27
→ saurik_ joined08:27
→ Arrowmaster joined08:27
→ lirakis joined08:27
→ multi_io joined08:27
→ Tronic joined08:27
→ stepnem joined08:27
→ cbz joined08:27
→ Juri_Agarin joined08:27
→ comcor- joined08:27
→ ecksun joined08:27
→ rootard joined08:27
→ m0 joined08:27
→ funfun joined08:27
→ orafu joined08:27
→ womble joined08:27
→ kandinski joined08:27
→ m_3 joined08:27
→ mcjb joined08:27
→ halfline joined08:27
→ dean[w] joined08:27
→ j0ran joined08:27
→ divVerent joined08:27
→ robinsmidsrod joined08:27
→ avar joined08:27
→ Derander joined08:27
→ metze joined08:27
→ mrtazz joined08:27
→ blalor_afk joined08:27
→ tangledhelix joined08:27
→ piti joined08:27
→ sunn` joined08:27
→ tlonim joined08:27
→ Teo` joined08:27
→ elly joined08:27
→ robinbowes joined08:27
→ sven^ joined08:27
→ teuf joined08:27
→ Leonidas joined08:27
→ selckin joined08:27
→ tomaw joined08:27
zelazny.freenode.net set mode: +v08:27
dean[w] At the moment we're doing "git fetch origin" then "git merge --no-ff some-branch"08:27
← sejo left08:28
→ sejo joined08:28
← OOPMan left08:28
Mannequin → Guest2661908:28
→ leavittx joined08:29
→ resmo joined08:29
→ OOPMan1 joined08:29
→ OOPMan joined08:29
→ drl joined08:29
← _ollie left08:29
→ bburhans_ joined08:29
← bburhans left08:30
← amerine left08:30
← OOPMan left08:30
← srcerer left08:31
← coldboot left08:31
→ Zoresvit joined08:32
dean[w] urgh netsplit lost my question08:32
If i check out a remote branch, locally. How can I ensure that I have the latest changes of it. e.g. git checkout test-local-branch origin/some-branch - if the origin makes a change to that branch, how do i make sure my test-local-branch has those changes also? At the moment we're doing "git fetch origin" then "git merge --no-ff some-branch"08:33
→ darkredandyellow joined08:35
← ch077179 left08:36
→ ch077179 joined08:36
← Rabbitbunny left08:36
→ Rabbitbunny joined08:37
→ psoo joined08:37
→ coldboot joined08:38
→ fgro joined08:38
← Utkarsh__ left08:39
← avinashhm left08:40
← boombatower left08:40
← codetonowhere left08:40
→ avinashhm joined08:40
→ danlucraft joined08:43
← drl left08:43
← hobodave left08:43
→ hobodave joined08:44
← EnginAy left08:44
→ EnginAy joined08:45
← doener left08:49
cbreak_work dean[w]: you should not check out remote branches, you should always make a local tracking branch for them08:49
→ DiaPo joined08:49
← deitarion left08:50
cbreak_work if you do git checkout -t remotename/branchname, you get a local branchname that has the remote's branch as upstream08:50
→ doener joined08:51
cbreak_work then you can just do a git pull while on that branch to pull it08:51
→ PigeonFriend joined08:51
dean[w] right, thanks :)08:51
→ nevyn^1 joined08:52
← avinashhm left08:52
cbreak_work pull is like fetch and merge08:52
jaeckel cbreak_work: what's the difference between checking out a remote branch and explicitely creating a local branch that tracks the remote branch? doesn't git do this automatically?08:52
cbreak_work sometimes08:52
sometimes you just get a detached head08:53
jaeckel if you do a git checkout "exact remote branch name"08:53
← serard left08:53
cbreak_work sometimes08:53
jaeckel ok08:53
cbreak_work it depends on your git version08:53
→ avinashhm joined08:53
jaeckel right08:53
thx08:53
→ banisterfiend joined08:54
→ koltroll joined08:54
→ chrischambers joined08:55
→ _ollie joined08:55
→ serard joined08:55
← theoros left08:57
← DiaPo left08:58
→ unon joined09:01
← laj left09:04
← SIGBUS left09:06
→ thermal joined09:09
silex → Silex09:10
→ LouisJB joined09:10
→ mziulu joined09:11
→ rhl joined09:11
→ drl joined09:11
→ DiaPo joined09:12
← apalto left09:13
→ everyman joined09:17
ianmcorvidae|alt → ianmcorvidae09:20
← aah left09:21
← drl left09:21
← everyman left09:21
← banisterfiend left09:22
→ petrux joined09:25
→ lundtor joined09:25
← gretch left09:26
lundtor Question: how to checkout the latest version before a given date?09:26
→ Cro_Crx joined09:26
→ refund joined09:27
refund How can i get the date of the last commit on a branch?09:27
lundtor refund: you can switch to the branch and do a git log, but i guess its too manual for you?09:28
dean[w] git log -n1 | grep <something clever goes here>09:28
refund thanks09:28
cbreak_work no need to switch branches09:29
just use git log with a format09:29
← fgro left09:30
← Cro_Crx left09:31
→ ExtraSpice joined09:31
← jjore left09:31
cbreak_work git log -n1 --format=%cd branchname09:31
→ tatsuya joined09:32
refund cbreak_work, "fatal: unrecognized argument: --format=%cd" :o09:33
→ engla joined09:33
← napster_123 left09:33
← engla left09:33
→ engla joined09:33
← mikepack left09:33
lundtor none knows how to checkout a version before a given date?09:34
→ jjore joined09:34
→ gbjk joined09:35
→ gretch joined09:35
→ drl joined09:35
ribng hi all, how to get the list of commits on particular file ?09:36
gbjk Hi. I want to remove a commit, so that a file revrts to a previous stage. I thought I wanted to do git rebase -i SHA1..SHA1~3 and then just remove it.09:36
→ ProLoser joined09:37
gbjk I get "Needed a single revision. Invalid base"09:37
ProLoser git gui broke for me on msysgit09:37
cbreak_work refund: get a non-ancient git09:37
ProLoser it thinks there are no changes in any files i view09:37
even though git diff shows the changes just find09:37
*fine09:37
cbreak_work lundtor: want to use the reflog?09:37
git checkout HEAD@{10.days.ago} for example09:38
lundtor ribng: git log <path>09:38
cbreak_work that will not consider the date of the commit09:38
gbjk: read what it says09:38
lundtor cbreak_work: background is that i have an android checkout, where all git repositories are on master branch.. well i updated all 200+ git repositories a few times and now all hell is loose, since everythings broken...09:39
but i know when it was working the last time (the exact date).. thus i want to reset every git repositoru to the version it was on that date09:40
gbjk cbreak_work: Yep. So I switch to just git rebase -i HEAD and I get "noop" in the editor and "Rebase d4a7392..d4a7392 onto d4a7392"09:40
lundtor i knwo i screwed up :-)09:40
gbjk cbreak_work: Which lead me to try to specify a range.09:40
← MetaCosm_ left09:41
← engla left09:41
lundtor cbreak_work: error: pathspec 'HEAD@{100.days.ago}' did not match any file(s) known to git.09:42
→ MetaCosm joined09:42
→ moonk joined09:42
→ kipibenkipod joined09:42
→ dvaske joined09:44
→ banisterfiend joined09:45
→ Cro_Crx joined09:45
ribng lundtor, thx.09:45
ProLoser how can i download the latest tag: http://repo.or.cz/w/msysgit.git/09:46
i don't understand this git repo tool09:47
← danlucraft left09:47
← tango_ left09:47
wereHamster ProLoser: git clone git://repo.or.cz/msysgit.git; cd msysgit; git checkout Git-1.7.3.2-preview2010102509:48
→ tango_ joined09:48
→ newbiz joined09:49
← orafu left09:50
→ orafu joined09:50
ribng hi all, i have something liek this #git show track/fnt.c , also track is a symbolic link. But git show is not resolving track ?09:52
any way to do that ?09:52
← avinashhm left09:52
← gregglind left09:53
→ avinashhm joined09:53
Rabbitbunny → Rabbitbu2ny09:54
→ ningia joined09:55
ningia hi all09:55
ProLoser wereHamster: thanks09:56
ningia is there a way to see the list of repository?09:57
wereHamster ningia: where09:57
Rabbitbu2ny → Rabbitbu3ny09:58
→ carloratm joined09:58
→ codebeaker joined09:58
carloratm hi all09:58
n00b question here09:58
codebeaker Hi all, I've asked some colleagues, and done some reading - but I can't figure this out, can anyone tell me why the master->production merge made by recursive, not by fast-forward?09:58
https://gist.github.com/f16c1aef61d3ee4b9bdf09:58
Rabbitbu3ny → Rabbitbu2ny09:58
→ ayanes joined09:59
← ayanes left09:59
→ ayanes joined09:59
← carloratm left09:59
wereHamster codebeaker: because before the merge the merge-base of master and production was not master09:59
ningia wereHamster: I've pushed a commit on my server, and reset my local to a previous commit... now I want to return to the server version, but git pull say me that dunno wich repo pull...09:59
codebeaker wereHamster: I'm going to have to read that a few times :)09:59
wereHamster codebeaker: sorry, the merge base was not production10:00
codebeaker wereHamster: is that user error on my part, or a bad historical branch maneuver, that I need to fix10:00
wereHamster codebeaker: that is, there were commits in master which were not in production and commits in production which were not in master.10:00
thre is nothign to 'fix'10:00
it just happens10:00
codebeaker wereHamster: it makes a real mess of my history10:00
ProLoser anyone know where i can get the LATEST msysgit installer?10:01
← mapreduc` left10:01
mlq jup10:01
ProLoser getting the latest tag doesn't provide me with something i can install10:01
→ mapreduce joined10:01
wereHamster you can rebase if you wish so.. however be aware of the consequences10:01
ProLoser and the latest download doesn't have some fixes to bugs i have10:01
codebeaker wereHamster: is there any major risks ?10:01
wereHamster ningia: git pull origin master10:01
mapreduce Hi. Does #git have some mode that prevents changing your nick while unauthenticated?10:01
← b_jonas left10:02
ningia wereHamster: fatal: 'orgin' does not appear to be a git repository10:02
wereHamster codebeaker: man git-rebase, google.com/index?q=git+rebase+...10:02
jast codebeaker: the 'git-rebase' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html [automatic message]10:02
→ danlucraft joined10:02
wereHamster ningia: origin10:02
mapreduce Or was mapreduc` banned in some way I can't fathom?10:02
ningia wereHamster: shame on me... -.-"10:03
→ mabrand joined10:03
Rabbitbu2ny → Rabbitbu3ny10:04
codebeaker thanks, wereHamster - I know about rebase, but master/production should be the same10:04
wereHamster codebeaker: if they are the same, then there is nothign to merge10:04
codebeaker master only ever gets +1 commit ahead10:04
wereHamster codebeaker: gitk --all10:05
ProLoser any recommendations on what i should do if i want to install a later version of msysgit than provided on the downloads page?10:05
Rabbitbu3ny → Rabbitbu2ny10:06
← QinGW left10:07
→ borwar joined10:08
→ gregglind joined10:08
← borwar left10:09
noobdearia → necrodearia10:09
→ borwar joined10:09
→ bluenovember joined10:10
← bauruine left10:10
← sraue left10:10
→ giallu joined10:10
← thierryp left10:11
→ spaceonline joined10:11
Rabbitbu2ny → Rabbitbu3ny10:12
→ dpino joined10:13
→ c_nick joined10:13
Rabbitbu3ny → Rabbitbu2ny10:14
← ProLoser left10:15
→ curtana joined10:15
→ drizzd_ joined10:15
Rabbitbu2ny → Rabbitbu3ny10:15
← c_nick left10:15
→ guardian joined10:17
→ alus joined10:18
← FilipeMaia left10:18
alus is there a way for me to peak at the git history without pulling?10:18
I want to know what changes await me upstream10:18
← zheng left10:18
patrikf fetch10:19
← carutsu_ left10:19
← drizzd left10:19
← fivetwentysix left10:20
Rabbitbu3ny → Rabbitbu2ny10:21
wereHamster alus: pull = fetch+merge10:21
codebeaker wereHamster: thanks for the tip, but it's a mess :) - I figured if the diff between the two HEADs is clean, then a merge with +1 commit (you see my log) on one branch would fast-forward onto the other10:21
→ fivetwentysix joined10:21
← Theravadan left10:21
alus wereHamster: after "git fetch", I see nothing in "git log"10:21
codebeaker fortunately I think I can blow away production, next time it is (== master), and then I can fork it10:21
alus nothing additional, I mean10:21
wereHamster codebeaker: actually, I don't see your log10:21
codebeaker: pastebin a screenshot of gitk --all10:22
← fivetwentysix left10:22
← hyperair left10:22
→ pdelgallego joined10:22
wereHamster alus: git log ..@{u}10:22
alus: of course you don't see anythig new, because log still shows you the same local branch10:23
you need to tell git to display the commits you just fetched10:23
→ sraue joined10:23
alus I have not idea what "..@{u}" means10:24
wereHamster alus: man git-rev-parse10:24
jast alus: the 'git-rev-parse' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rev-parse.html [automatic message]10:24
alus is it an emoticon for a snail?10:24
Weust`afk → Weust`10:24
alus I don't know how "git-rev-parse" is applicable here10:24
wereHamster alus: it explains what @{u} means ... ?!?10:25
→ hyperair joined10:25
codebeaker wereHamster: you know an image paste bin that doesn't suck ?10:26
← Rabbitbu2ny left10:26
wereHamster codebeaker: imageshack, bayimg, skitch, ... just pick the next best one you know10:26
← HG` left10:27
timj__ → timj10:27
codebeaker http://img5.imageshack.us/i/screenshot20110119at112.png/ - thanks wereHamster10:27
← _ollie left10:27
→ airborn joined10:27
alus whoever made that revision/ref syntax needs to be stabbed with regular expressions until science discovers a cure for their disease10:28
wereHamster codebeaker: nd you want to do what?10:28
codebeaker wereHamster: have master merge FForward into production10:29
wereHamster alus: propose a better syntax, the mailing list is in the topic10:29
codebeaker: merge production into master, then master into production, that'll be ff10:29
codebeaker ah, really ?10:29
thanks10:29
yeah, I tried a rebase, but that didn't really work out (and I know exactly why!)10:30
→ wtfness joined10:30
← hyperair left10:30
→ err_ok joined10:30
alus wereHamster: the syntax I propose is "english"10:31
but I would settle for "words"10:31
wereHamster alus: ok. But I'm not willing to write prosa on the commandline10:31
→ phoque joined10:31
phoque hey everybody10:31
→ tvw joined10:31
phoque I have a problem where my local clone of a repo doesn't hold the same code as one of the remotes10:32
← kanru left10:32
← foocraft left10:32
→ HG` joined10:32
phoque Git seems to think that they match (pulling says everything up to date) but they don't10:32
→ aeiou joined10:32
← gretch left10:32
→ Yuuhi joined10:33
aeiou i was working in branch "speedtest" and had commited a new change, but had a dirty working tree - so i did git stash. then i checked out master and did git merge speedtest10:33
phoque specifically, I've forked https://github.com/symphonycms/symphony-2/tree/integration to https://github.com/nils-werner/symphony-210:33
aeiou but it seems to have added and merged my stashed changes? :/10:33
patrikf phoque: which branches do not match?10:34
phoque integration10:34
← alus left10:34
← mlq left10:35
phoque oh hangon10:35
patrikf phoque: do a git remote update, then git log integration...origin/integration; does it output anything?10:35
phoque no10:35
weird10:36
I'm having a look at the file I thought didn't match10:36
but now they do10:36
(it's not that my file caught up but that the newer file wasn't changed at all)10:36
→ artefon joined10:37
← johnw left10:37
phoque patrikf: sorry for stealing your time10:37
:-)10:38
← killerchicken_ left10:38
→ kreemer joined10:39
← roger_padactor left10:39
patrikf never mind, i knew it was user error anyway ;-)10:39
← Pacman_Home left10:41
← soidexe left10:41
→ codespectator joined10:41
→ soidexe joined10:42
→ Pacman_Home joined10:42
→ ereslibre joined10:43
← ereslibre left10:43
→ ereslibre joined10:43
phoque patrikf: oh of course... :-D10:43
→ Titosemi1 joined10:45
← devanjan left10:46
← Titosemi left10:47
→ leo2007 joined10:48
← leo2007 left10:50
→ tomboh joined10:50
→ Pacman_Home` joined10:50
← tvw left10:50
← drl left10:50
→ tvw joined10:51
← Pacman_Home left10:51
← avinashhm left10:52
→ marc_kdab joined10:52
→ avinashhm joined10:53
← banisterfiend left10:54
← ereslibre left10:55
← gebi_ left11:00
→ vdv joined11:01
← nevyn^1 left11:02
← guardian left11:05
← Zoresvit left11:05
→ kanru joined11:05
← vdv left11:06
→ Zoresvit joined11:07
→ banisterfiend joined11:07
→ ereslibre joined11:08
→ darkling joined11:09
→ vitor-br joined11:12
→ devanjan joined11:13
← spaceonline left11:16
→ napster_123 joined11:18
← agateau_ left11:20
→ agateau joined11:21
← vitor-br left11:21
→ Ryback_ joined11:22
← pdelgallego left11:22
→ xiong joined11:23
→ leo2007 joined11:23
← normanrichards left11:24
← mahata left11:24
→ Pacman_Home joined11:25
← Pacman_Home` left11:27
→ jfkw joined11:29
← Pacman_Home left11:30
→ Pacman_Home joined11:30
→ zorzar_ joined11:30
→ bentob0x joined11:32
← zorzar left11:33
→ vdv joined11:38
→ thiago_ny joined11:39
← ExtraSpice left11:39
→ pdelgallego joined11:40
← thiago_ny left11:43
→ thiago_ny joined11:43
→ miloshadzic joined11:45
← mdel left11:46
← codebeaker left11:47
← devanjan left11:49
darkling If I'm developing a patch series over a period of time, and I rebase interactively to clean up the history (i.e. merge incremental bug fixes into major patches), how do I get git to update the date on the main patches to the date I rebased?11:49
→ g0bl1n joined11:49
darkling Alternatively, how do I get it to use the current date when I send patches to the relevant mailing list?11:50
phoque does it really matterß11:50
:-)11:50
darkling Well, yes, because at the moment, I'm sending mails dated several days (or even months) ago.11:50
phoque ah ok11:50
→ Pacman_Home` joined11:50
← Pacman_Home left11:51
← avinashhm left11:52
darkling Is there a good write-up of development workflows using git? How do other people develop patch series?11:52
phoque can't you just export the patch11:53
and send the mail yourself?11:53
→ avinashhm joined11:53
phoque (I'm assuming there's a function in git that sends mails, dated to the time and date of the commit)11:53
→ gebi joined11:53
darkling I'm using git send-email for it at the moment.11:53
phoque darkling: I haven't used anything besides Github (genius!) and repos I had write access to...11:54
← pratz left11:54
darkling Doing it manually would be a right pain in the backside for anything more than a few patches.11:55
phoque correct11:55
:-)11:55
darkling People are clearly using git in the way that I am,11:55
but the effects that they're getting are different to the effects that I'm getting,11:55
so I thought I'd ask in here and try to work out what the difference in usage it.11:55
phoque do you have a repository that can be accessed?11:56
you could ask people to pull from there11:56
isntead of sending patches11:56
darkling I'm working on the Linux kernel (on btrfs),11:56
phoque by "can be accessed" I mean "read-only from anywhere in the word"11:56
darkling so patches to mailing lists are expected.11:56
phoque ah ok11:56
← leo2007 left11:57
→ pratz joined11:57
← pratz left11:58
← unon left11:58
→ unon joined11:58
→ hyperair joined11:59
divVerent is there a good alternative to git-send-email that preserves commit IDs?12:04
→ nevyn^1 joined12:04
divVerent git-bundle is ugly, it makes binaries12:04
wereHamster you can't preserve commit ids with send-email (format-patch). push/fetch or bundle is your only option12:05
divVerent well, I had read that git-fast-export possibly does that12:05
but didn't work out in practice, maybe I was using it wrong12:05
wereHamster but that's for a completely different use case12:05
divVerent git fast-export HEAD^1..HEAD12:05
makes a HUGE file12:05
containing files the commit has not even touched12:05
wereHamster as it should12:05
divVerent You can use it as a human-readable bundle replacement (see git-12:06
bundle(1)), or as a kind of an interactive git filter-branch.12:06
I had assumed it would be as useful as git-bundle :P12:06
from this manpage item12:06
wereHamster what's wrong with bundle?12:07
divVerent but maybe I just missed an option to fast-export to allow excluding objects?12:07
→ winterweaver joined12:07
divVerent it is binary12:07
I would prefer something that has human readable diffs12:07
AND preserves commit IDs12:07
(as long as the diffs are not touched, of course)12:07
→ bauruine joined12:07
wereHamster send a diff and tell the receiver to fetch from <url> if he wants the commit12:07
divVerent sure, this is a workaround for now... but I prefer a real solution12:08
why cannot format-patch preserve that info?12:08
wereHamster format-patch + attached bundle12:08
divVerent assuming linear history in the changeset12:08
wereHamster because it doesnt' preserve the committer12:08
divVerent that should be a simple addition...12:08
→ marcob joined12:08
divVerent so that can't be the only problem12:08
→ leo2007 joined12:09
wereHamster sure, pass the committer identity+timestamp along with the patch, and make sure the receiver knows what to do with it12:09
← opalka left12:09
divVerent I just wonder why this hasn't been done yet12:09
there must be something that makes it not work that easily12:10
not just "nobody has coded these options for format-patch and am yet"12:10
wereHamster because you shuldn't (and can't, rightfully so) force your committer identity onto other people when they apply your patches12:10
divVerent but I am the committer...?12:10
→ DiGitalX joined12:10
wereHamster git distinguishes between committer and author12:10
no. you are the author12:11
divVerent but I did the commit to my local repo12:11
and if they pulled, I would stay the committer12:11
wereHamster and whoever commits your pathc is the committer12:11
yes12:11
divVerent I just don't see the reason for distinction12:11
← SirFunk_ left12:11
divVerent why does it matter whether they get my change via git protocol, or via email?12:11
it's just another protocol12:11
← lundtor left12:11
avar beacuse with one you get commits, with another you get text that you make commits from12:11
hence making you the "committer"12:11
→ opalka joined12:12
divVerent basically... asking a different way12:12
what I really want is send a bundle12:12
but also have an easy way for reviewing12:12
is there a way to automatically view the diff a bundle would cause to a ref?12:12
without actually importing it12:12
wereHamster no12:12
divVerent that could be a nice mutt handler for bundle files12:12
of course it'd need to know where the repo is12:12
avar just import it, it's a reversible operation.12:13
divVerent avar: I mean as mailcap viewer12:13
so when someone attaches a bundle, I see the diff from within my mail client12:13
→ devanjan joined12:13
avar Well that's between you and your mail client, I'm sure you could make a plugin for it.12:13
divVerent because, that would be what I really want12:14
avar i.e. unpack the bundle, apply it in /tmp, diff it.12:14
divVerent and what format-patch/send-email is an approximation to12:14
avar sounds like something that's easy to write.12:14
→ goshawk joined12:14
divVerent needs a full extra copy of the git repo, though12:14
avar sure12:15
divVerent still... why is there no REAL text encapsulation for commit information?12:15
avar you're not going to be able to view a custom binary blob in a mail viewer without support for that custom binary crap.12:15
cbreak_work encapsulation?12:15
divVerent format-patch basically lacks two things:12:15
- committer info12:15
- revision graph12:15
cbreak_work there's real encapsulation for commit info12:15
→ thierryp joined12:15
divVerent it only supports linear/linearized history12:15
cbreak_work it's in the commit object12:15
wereHamster revision graph? Why woyuld it need it?12:16
avar Because people haven't made it yet, because it hasn't been solving their problems. You could easily make one if you need it.12:16
wereHamster you mean you want to be able to mail a merge commit?12:16
divVerent will probably have to take quite a while to make a real "bundle viewer"...12:16
Silex just embed the whole git repo with each of your emails and you're good12:16
divVerent that ideally uses e.g. a gitweb remote reference to "resolve" missing objects12:16
or a git:// URL12:16
and shows the bundle as a diff of what gets applied to each ref, and additional commit messages12:16
I just wonder why it has to be that complex12:17
when format-patch is already a nice approximation to it12:17
→ Rickardo1 joined12:17
divVerent I could probably write a quick shell script replacing format-patch12:17
cbreak_work format patch does not transfer any history12:17
it doesn't transfer commits either12:17
divVerent essentially executing "git show" on each commit in the difference set12:17
and another tool that applies that info12:17
format-patch doies transfer commits12:18
cbreak_work no12:18
divVerent but without their relation12:18
instead, as linearized history12:18
basically, it's like a rebase operation12:18
cbreak_work a commit is a snapshot12:18
format-patch does not contain the snapshot12:18
divVerent sure12:18
cbreak_work it only contains a diff12:18
divVerent and I do not want such a snapshot12:18
cbreak_work it's worthless12:18
divVerent I want "diff to a known state"12:18
cbreak_work without something to apply it to12:18
divVerent of course, so?12:18
← jds left12:18
cbreak_work if you want to transfer a commit, you should use git archive12:19
divVerent I want to transfer "enough information to recreate the commit, etc. objects as is, human readable"12:19
cbreak_work that way you at least get the whole commit tree12:19
divVerent and by human readable, I also mean that it is easy to see what has changed12:19
so fast-export does not qualify, as it contains so much info it is not human readable any more12:20
(mainly: because it contains full files, no diffs, and even unrelated files)12:20
cbreak_work what does format-patch do with images?12:20
divVerent let's take these out12:20
for images review is quite unimportant anyway12:20
cbreak_work can't really do anything human readable with non-text info12:20
divVerent exactly12:20
cbreak_work not even with diffs on things like XML12:20
divVerent not that important12:20
for source I want to see a diff12:20
cbreak_work it's important that the information is not lost12:20
divVerent of course12:21
but format-patch handles THAT already well12:21
← aeiou left12:21
cbreak_work how?12:21
divVerent by doing a "binary diff"12:21
of coruse not human readable12:21
but human can skip it at least12:21
and read the diff of the next source file12:21
cbreak_work so it somehow ascii encodes the diff12:21
divVerent and git-am can apply it12:21
yes12:21
in fact... I wonder if one can do an export like I want it, without needing a special tool on the receiver end12:22
cbreak_work does it encode ancestor relationships?12:22
divVerent by using shell syntax to create all objects12:22
no, these are not encoded12:22
it basically rebases, then makes patches for the rebased history12:22
one of the two things I dislike about it12:22
cbreak_work does it encode committer and author info (name, date, email)?12:22
divVerent lacks committer12:22
→ donri joined12:22
cbreak_work does it contain any hashes, like tree hashes, for verification/error checking?12:22
divVerent yes12:23
donri Can I "lift out" paths from a repo, creating a new repo?12:23
Retaining the history for those paths12:23
cbreak_work donri: man git-filter-branch12:23
jast donri: the 'git-filter-branch' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-filter-branch.html [automatic message]12:23
cbreak_work just delete everything else12:23
donri Thanks12:23
cbreak_work (use --index-filter or --subtree-filter)12:23
wereHamster donri: google 'git subtree github'12:23
it's a user friendly wrapper around filtr-branch12:24
donri To clarify, by paths I mean file paths12:24
cbreak_work I assumed you meant the files themselves12:24
sorry12:24
git filter-branch won't help you get the paths12:24
if you just want the paths, you'd probably have to use something like git ls-files12:25
donri I want "mv repo/foo/* foorepo/" but with git history :)12:25
← phoque left12:25
cbreak_work --subdirectory-filter12:25
donri I have a project contained in another project's repo, and want to move it to its own12:25
selckin just google split git repo12:25
donri Why didn't I think of that search term, thanks12:26
cbreak_work it's self explaining, more or less12:26
if you read the man page :)12:26
btw, do that in a clone of your repo12:26
→ hagebake joined12:27
← hagabaka left12:27
→ errordeveloper joined12:27
← davidc_ left12:28
errordeveloper hey, does anyone know if there is a way checkout all forks from github ??12:28
to be able to test all of them quite quckly ..12:28
wereHamster I'd use: git branch foorepo `git subtree split foo/`; and then fetch foorepo into a separate repo12:28
nevyn you can add them as individual remotes12:28
errordeveloper yeah .. but is there a record of all forks in the original repo ?12:29
← goshawk left12:29
wereHamster errordeveloper: maybe you can get that through the api12:30
-> develop.github.com12:30
errordeveloper wereHamster: yeah, that basically what I was asking ..12:31
wereHamster so basically you want us do your job?12:31
→ stamina joined12:31
← gebi left12:35
→ martin_ joined12:37
→ Agos joined12:38
cbreak_work you mean branch when you say fork?12:38
donri filter-branch left lots of "empty" commits12:41
cbreak_work yeah12:42
that's because you skipped the part of the man page12:42
→ felipe joined12:42
→ goshawk joined12:42
cbreak_work that contained the --prune-empty flag12:42
→ Dave^| joined12:43
donri True :) thanks12:43
→ brandFU joined12:43
divVerent cbreak_work: to the old topic: with a quickly hacked shell script, I got an output file like this: http://rm.endoftheinternet.org/~xonotic/temp/exported.txt12:45
executing this with the shell on the right repository recreates the commit hashes12:45
but it doesn't support merges yet12:45
but you see the general idea12:45
schweizer how can i see which changes are staged and will be committed?12:46
divVerent git diff --cached12:46
schweizer thx12:46
divVerent cbreak_work: obviously this format has some flaws, especially that the commit message might contain the marker12:47
← err_ok left12:48
schweizer how can i merge only one commit from a branch into master?12:49
divVerent the fun part is, this would be ALMOST doable with merely git-log with the right args12:49
only problem is the terminating line12:49
schweizer: a) you can't, b) but you can cherry-pick it (causing it to have a new commit hash, though)12:49
← koltroll left12:49
schweizer like "git cherry-pick HASH" in master working dir?12:52
divVerent yes12:53
→ tty1 joined12:56
← diago left12:57
← donri left12:57
← clarkfischer left12:58
← luxurymode left12:58
divVerent anyone happen to know a command to "apply" a merge as displayed by git-log?12:58
with the two-source diff thingies12:58
(+-, etc. stuff)12:58
→ teolicy joined12:59
← winterweaver left12:59
→ roger_padactor joined13:00
← unon left13:01
→ unon joined13:01
← koo5 left13:01
teolicy Hi. I'm writing a script that needs a checkout of the working directory representing a particular commit, do stuff and then bring the working directory to the exact same state it was before the street ran. I think the safest way (and the one which represents the semantics I want) is: 'cd /tmp ; git clone <path_to_repo> <tempname> ; cd <tempname> ; git checkout <desired_ref> ; <do stuff> ; rm -fr /tmp/<tempname>'.13:02
Any way to achieve this quickly and cleanly?13:02
wereHamster teolicy: git checkout <commit>; do stuff; git checkout -13:02
teolicy (emphasis on quickly, possibly even at the expense of cleanly, we'll have to discuss this)13:02
wereHamster: What if I had changes in my repo before the first checkout? They'll be carried over.13:02
I need to stash them.13:02
And what if I have untracked files?13:03
wereHamster sure, stash them13:03
git won't touch untracked files13:03
→ codebeaker joined13:03
teolicy I know, but I want my checkout to be clean. Very clean.13:03
wereHamster then do what you came up with13:03
teolicy I don't even want ignored files (build products).13:03
wereHamster: But that's slow... I was hoping there's something like hardlinking most of the existing repo and checking out only the changes, or doing something like two working directories using the same repo, or something like that.13:04
wereHamster don't clone into tmp, clone inside the same filesystem13:04
(I'm assumign your /tmp is on tmpfs or somethign like that)13:04
teolicy Cloning on the same filesystem will hardlink by default?13:05
wereHamster it'll use hardlinks for the stuff inside .git13:05
← Error404NotFound left13:05
wereHamster and if you are doing your dance often, you might want to keep the temporary clone instead of rm-rf'ing it13:06
← KyleXY[2] left13:06
→ Kyso joined13:06
← Kyso left13:06
→ Kyso joined13:07
← Kyso left13:07
→ mahata joined13:07
ribasushi just sanity checking: if I am running a bisect on master, the bisect will not dive into merged branches right?13:07
wereHamster sure it will13:07
→ MetaCosm_ joined13:08
→ Kyso joined13:08
← MetaCosm_ left13:09
→ koo5 joined13:09
→ keystr0k joined13:09
→ d0k joined13:09
← thermal left13:13
← dekroning left13:14
teolicy wereHamster: My nefarious plan is to: cd <repo_path> ; cd .git ; [ ! -d _teolicy_ ] && git clone -shared . _teolicy_ ; cd _teolicy_ ; git reset --hard <desired_ref> ; git clean -fdx ; <do stuff>13:16
wereHamster: I know it's very unorthodox, but it seems to me like it has precisely the semantics I want and the least garbage lying about.13:17
(you could easily argue that garbage inside .git is quite horrific, and I'd understand, but still, I think this command has the least change of going wrong than any alternative I can think of)13:17
ribasushi wereHamster: how can I prevent this then?13:18
→ v0n joined13:19
ribasushi wereHamster: say I have commits A and B, and between them there was a branch-off and a merge for a feature implementation which has commits with failing tests13:19
← errordeveloper left13:19
← Zao left13:19
wereHamster why do you want to skip those commits?13:19
ribasushi wereHamster: how do I tell bisect to stay in the "shortest path" between A and B13:19
wereHamster: because if I am looking what caused a regression these dirty commits will not help any13:19
I will have to be stopping to skip all the time13:20
← thierryp left13:20
→ anon33_ joined13:20
→ tizbac joined13:21
→ jaql joined13:21
anon33_ On Github: Does adding collaborators to a project mean that they can push to the master branch at any time - ie no pull requests?13:21
← Zoresvit left13:22
→ gebi joined13:22
← gebi left13:22
→ gebi joined13:22
→ _pingu joined13:23
← thiago_ny left13:24
→ Chillance joined13:25
divVerent I now got my weird shell script tool able to even export merge commits... but they show as diff to the first parent as I know no tool to apply a git diff --cc13:25
← Tabmow left13:26
→ Zoresvit joined13:29
→ funsheep joined13:29
→ Tabmow joined13:32
→ normanrichards joined13:32
divVerent cbreak_work: in case it interests you - my current export-git-history shell script: http://paste.pocoo.org/show/323223/13:33
it writes a shell script that imports the same history back in13:33
← alip left13:33
funsheep Hi, i'd like to move from SVN to git. I searched on the net and found several how-tos. But every howTo assumed that there was only one project per svn repository. Now we have a lot of projects, categorised in various directories. Also every project does have its own trunk, branches and tags directory (e.g. /coreProjects/utils/trunk/)13:34
Before i dive deeper into git, can i use the same structure with git? Couldn't find anyone clearly saying "yes" or "no".13:34
divVerent you SHOULD use separate repository for this when using git13:34
← _pingu left13:35
→ alip joined13:35
→ cwebber joined13:35
← leo2007 left13:35
wereHamster funsheep: would you be so kind and point us to a how-to which tells you to put multiple projects into a single git repo?13:36
← gebi left13:36
← serard left13:36
→ gusnan joined13:37
← goshawk left13:37
→ henrikbjorn joined13:37
divVerent wereHamster: BTW, you may find this tool interesting too... basically, it MIGHT become a new alternative to git-format-patch that preserves enough metadata to even generate the same commit IDs, but is still "mostly human readable"13:38
only thing it lacks at the moment, is good merge support13:38
→ serard joined13:38
divVerent merges currently contain the diff to their first parent, instead of a "merge diff"13:38
→ thiago_ny joined13:39
← thiago_ny left13:39
→ thiago_ny joined13:39
wereHamster I don't see why I'd need it13:40
it's.. uhm.. 'interesting' for certain though13:40
divVerent sure13:42
I'd rather have something git-internal become that13:42
← Rickardo1 left13:42
divVerent e.g. a text version of "git bundle"13:42
as "shell archives" have one big flaw: what if someone is being malicious...13:42
quite easy to hide harmful stuff in them13:42
← SvenDowideit left13:43
FauxFaux If only you could trivially verify a git bundle by inspecting HEAD or refs/*.13:43
wereHamster just like someone could hide stuff in his repo before he sends you a pull request..13:43
divVerent sure, but a pull is quite easy to verify13:43
you pull, then git diff @{u}..13:43
wereHamster I mean.. you don't blindly apply stuff without reviewing it, no matter which source it comes from13:44
I just don't see what that has to do with bundles13:44
divVerent bundles may change multiple refs13:44
→ khmarbaise joined13:44
wereHamster uhm, so.. ? fetch the bundle into FETCH_HEAD and not refs/...13:44
funsheep @wereHamster: There is none, but there is also no howTo that explains my situation. Something like: "So you used one repository for several projects, then you should do this…. etc."13:45
divVerent wereHamster: ah, I see, yes, can be used that way too13:45
then they are sort of safe, yes13:45
→ pratz joined13:46
wereHamster if I tell you to 'git fetch <bundle> refs/heads/foo:refs/heads/master' or 'git fetch git://server/repo.git refs/heads/foo:refs/heads/master', where is the difference?13:46
← artefon left13:46
→ koltroll joined13:46
divVerent then the only problem of bundles is that they are binary13:46
wereHamster would you do the former but not the later?13:46
divVerent e.g. you cannot verify them right when reading the email with them13:46
wereHamster so is the git protocol13:46
divVerent sure13:46
← koltroll left13:46
divVerent format-patch has the advantage that it can be reviewed directly13:46
and basically, what I would like is a bundle with that ability :P13:47
codebeaker is there an option for format-patch which compresses it to ONE patch ?13:47
( I reviewed the --help, but didn't spot anything )13:47
wereHamster codebeaker: --stdout > the-big.patch13:47
divVerent of course, there is another idea to "fix" bundles for email verifying...13:47
codebeaker ach so :) makes sense, I was afraid about patch headers, etc13:47
divVerent make a bundle, attach it...13:47
as body text, add a FULL patch of the ref to be updated13:48
and when applying, have a tool that verifies the text part13:48
wereHamster that's what I suggested an hour or so ago..13:48
divVerent yes, but also needs integration13:48
→ _pingu joined13:48
→ rgr_ joined13:48
divVerent as otherwise I cannot know whether what I reviewed is actually in the bundle13:48
wereHamster as does everything else13:48
divVerent currently, we solve this problem a very weird and stupid way13:49
we have the users push to username/* branches in our big repo13:49
and when they want us to merge, they use a merge request system13:49
that sends an email that, among other stuff, contains a combined diff of all the changes, and git commands to verify and merge13:49
but... this basically forces a centralization13:50
wereHamster why?13:50
where?13:50
divVerent without having safety that the email is from MY merge request system13:50
I have no guarantee that the patch I reviewed is what I merge13:50
so we can't e.g. generate these emails from a script the user runs13:50
they have to go through a website for this13:51
wereHamster what if a user pushes, sends the merge request, you review it, in the meantime the user pushes some other code, you pull from his brnach and there you have the bad code in your repo13:51
divVerent that is handled :P13:51
← Adaptee left13:51
wereHamster how?13:51
divVerent wait, pulling up one of these requests13:51
anon33_ Is there a reason that 'git submodule foreach git commit -m "Initial Commit"' gives me an error? 'pathspec 'commit' did not match any files known to git13:51
divVerent wereHamster: the "How" is answered in a query, not publishing the URL here13:55
← teolicy left13:56
← alip left13:57
← martin_ left13:57
→ arhyttin joined13:57
← devanjan left13:57
← koo5 left13:58
← ciaranm left13:59
→ gebi joined14:00
→ generale joined14:00
→ ciaranm joined14:00
← arhyttin left14:01
← OOPMan1 left14:01
→ alip joined14:01
← roop left14:02
→ keystr0k_ joined14:03
← generale left14:03
→ generalecluster joined14:04
← normanrichards left14:04
ChanServ set mode: +v14:05
generalecluster hello to everyone14:05
← keystr0k left14:06
keystr0k_ → keystr0k14:06
generalecluster anyone here?14:06
→ cp joined14:06
→ zaxis joined14:07
cp → Guest6009414:07
← Guest60094 left14:07
zaxis whats the best way to deploy to test/prod using git14:07
→ teolicy joined14:09
← teolicy left14:09
← keystr0k left14:09
→ keystr0k joined14:09
← _pingu left14:09
← brandFU left14:09
→ _pingu joined14:10
→ brandFU joined14:10
← zaxis left14:11
← generalecluster left14:11
→ SvenDowideit joined14:12
→ jxpx777 joined14:14
→ sivy joined14:14
← pratz left14:14
→ alberto56 joined14:17
→ normanrichards joined14:17
← pgib left14:19
→ normanrichards_ joined14:19
→ fivetwentysix joined14:20
← saurik_ left14:21
← _pingu left14:21
→ saurik joined14:21
→ Adaptee joined14:21
← bauruine left14:21
→ romeo404_ joined14:22
← normanrichards left14:23
normanrichards_ → normanrichards14:23
→ Maletor joined14:24
→ posciak joined14:25
← brandFU left14:25
offby1 wounds his knee14:27
← maleknet left14:27
← kanru left14:29
→ kylehayes joined14:29
→ ISF joined14:31
→ smuf joined14:31
← marc_kdab left14:31
→ kanru joined14:31
← ISF left14:31
→ marc_kdab joined14:31
← opalka left14:31
→ ISF joined14:32
← ISF left14:32
← borwar left14:32
→ ISF joined14:33
→ Pewpewarrows joined14:33
Chillaway → Chillax14:34
← ruskie left14:35
→ blee joined14:38
→ flaguy48 joined14:38
← ajpiano left14:39
→ apalto joined14:39
blee Anyone here use Assembla's git services?14:39
offby1 watches the tumbleweeds skitter down Main St14:39
← BiggFREE left14:40
→ jbd joined14:40
← ph^ left14:40
→ LiamH joined14:41
blee lol :-(14:43
← ISF left14:43
blee im having an issue, where when I try to clone my assembla repo, its completely ignoring my ssh key14:43
→ ISF joined14:43
→ scelis joined14:44
← DiGitalX left14:44
blee if i try a simple ssh it looks like its looking for is_rsa.14:45
and not .pub14:45
http://pastebin.com/bVT01W3c14:45
workmad3 blee: the key is split over 2 files14:45
blee: id_rsa is the private key, id_rsa.pub is the public part14:45
xiangfu Hi, I run "git rebase OTHER_BRANCH" --> "git push -f" update the remote git.14:45
in COMPUTER_A14:46
workmad3 blee: if you removed the id_rsa file, you removed the key14:46
xiangfu then in COMPUTER_B, I run "git pull"14:46
← gusnan left14:46
blee workmad3: they are both there14:46
→ zivester joined14:46
blee i guess thats a period at the end, not ssh looking for a file with no extension14:46
xiangfu it's always can not success, and give me : Your branch and 'origin/master' have diverged,14:46
how can I do a force update on my local branch on COMPUTER_B . thanks14:47
← Yuuhi left14:47
workmad3 blee: probably ;)14:47
← Zoresvit left14:47
workmad3 blee: do you know if your public key has been given key-based access on the remote machine?14:48
offby1 xiangfu: can you paste a transcript?14:48
← anon33_ left14:48
blee workmad3: It has been added14:48
← Badiss left14:48
← rhl left14:49
→ Bass10 joined14:49
→ Badiss joined14:49
workmad3 blee: and it's definitely looking for id_rsa, and the file is located in ~/.ssh ?14:49
← alberto56 left14:49
bluenovember xiangfu, COMPUTER_B has made a commit to its working tree that COMPUTER_A does not have. You can either take in the branch from B by fetching from origin on COMPUTER A ("git fetch origin master", "git rebase origin master") then force pushing, or ignore what B has done; COMPUTER B git reset --hard origin/master14:49
blee workmad3: yes, and I use the exact same keys to ssh into my own servers14:49
xiangfu bluenovember: there is no commit on COMPUTER_B, COMPUTER_B only have read access.14:50
blee workmad3: i might be onto something, just a moment :)14:51
bluenovember xiangfu, there is, even if it hasn't pushed it14:51
that's what "diverged" means14:51
xiangfu offby1: here is some info: http://pastebin.com/biiiZ97T14:51
workmad3 blee: heh, I only just looked at your pastie14:51
blee: it's actually not complaining that it can't find the file... it's saying that the file doesn't contain an RSA1 key14:51
blee yeah i am just realizing that14:51
xiangfu bluenovember: oh. I mean not commit anything in COMPUTER_B, COMPUTER_B is a buildhost, just run git pull always.14:52
workmad3 serves me right for not reading your error message ;)14:52
one of the first things I always tell other people to do :/14:52
→ jrphelps joined14:52
← jxpx777 left14:52
blee workmad3: im comparing a successful ssh against the assembla one now14:53
→ Miika-- joined14:53
Miika-- I have cloned a project with submodules from github... Does git pull update submodules also?14:54
xiangfu bluenovember: Hi, for the buildhost, you know always run 'git pull', never commit, never push, I can run command like this:14:55
1. git fetch -a14:55
2. git reset --hard origin/master14:55
3. git pull14:55
→ Cipher-0 joined14:55
bluenovember that looks like a good plan14:55
xiangfu bluenovember: is it better then just "git pull"14:55
bluenovember xiangfu, perhaps B has pulled a commit that later got removed (via a forced push after a rebase or something)14:55
blee workmad3: lol wow, looking closer at the log14:55
← jrphelps left14:55
blee workmad3: debug1: Remote protocol version 1.99, remote software version OpenSSH_2.3.0_Mikrotik_v2.914:56
→ jxpx777 joined14:56
bluenovember xiangfu, yes, as it avoids the situation where an upstream rebase has occured14:56
blee my router is like intercepting the ssh request14:56
and this is actually and SSH to our router14:56
workmad3 blee: ouch!14:56
blee wtf!14:56
hahaha14:56
→ pr0wler joined14:56
← coldboot left14:56
blee workmad3: I was so confused because I was on my laptop last night using git like no ones business14:56
cbreak_work xiangfu: git pull after a git reset --hard is redundant14:56
← napster_123 left14:56
cbreak_work (unless you pull from somewhere else)14:56
workmad3 blee: heh14:56
← Miika-- left14:57
workmad3 blee: so from your perspective, it was working fine at work, you came home and it suddenly stopped working14:57
blee: commercial routers are crap14:57
blee workmad3: i had a suspicious of it being my router, but it seemed so unlikely14:57
xiangfu bluenovember: ok. thanks. I have another question: I must sync our master branch to 'upstream master', sometimes. then I must run "git rebase", how you people manage this ? just give me some advice.14:57
cbreak_work: thanks.14:57
workmad3 blee: I once had MSN (and only MSN) stop working because my router didn't like windows XP14:57
→ tyteen4a03 joined14:57
workmad3 or vista... it may have been vista actually14:58
blee workmad3: well if it was vista...14:58
cbreak_work no one likes vista14:58
xiangfu I know 'git merge' may easy, but I also want make our commit clear, easy generate patch, send patches to upstream.14:58
workmad3 and to fix it, I had to upgrade to the firmware release just before the latest one for my router14:58
blee legit fix14:58
workmad3 blee: it's the fact it was broken again in the latest firmware :P14:59
point is... nothing surprises me when it comes to routers nowadays14:59
cbreak_work I use Tomato and DDWRT with some success14:59
xiangfu the 'git rebase' can not avoid? right?14:59
offby1 xiangfu: from your paste, it looks like git has refused to pull because you haven't told it your name and email address. Try fixing those problems.15:00
xiangfu offby1: I have tried that, same problem. and I think it's just a warning not ERROR :)15:00
→ andrewvo1 joined15:00
offby1 xiangfu: well, it says "fatal: empty ident <xiangfu@fidelio15:01
.(none)> not allowed"15:01
blee I cannot think of 1 logical reason for for my router stealing my ssh connection15:01
s15:01
offby1 that sounds like an error to me.15:01
→ DiGitalX joined15:01
→ t0rc joined15:01
← avinashhm left15:01
← zivester left15:02
cbreak_work blee: try to ssh to some random location15:02
does it intercept it?15:02
like ssh whitehouse.gov15:02
blee lol15:02
i can SSH to IP's it seems15:02
one se15:02
c15:02
→ al joined15:03
→ smolyn joined15:03
→ dfr|mac joined15:03
blee cbreak_work: no when I try to connect to one of my servers using an aname record it connects fine.15:03
al can somebody tell me what i'm missing when using a bare repo with --git-dir/--work-tree? http://pastebin.com/w5kk6Qvy15:03
i always get "fatal: This operation must be run in a work tree"15:04
→ SirFunk_ joined15:04
thiago_ny al: a bare repo has no work tree15:04
blee cbreak_work: could be a port 22 thing15:04
al thiago_ny, that's why i'm specifying --git-dir resp. --work-tree15:05
offby1 that's some port, that port 2215:05
thiago_ny al: remove the "core.bare" configuration then15:05
cbreak_work al: you know you have to specify both?15:06
al cbreak_work, no, i thought they were mutually exclusive15:06
→ Yuuhi joined15:07
→ aah joined15:07
→ cptG joined15:08
al cbreak_work, that works, thanks15:08
← jfkw left15:08
al funny thing is, on a different box it's even working without15:08
→ cannonball joined15:10
← thiago_ny left15:10
→ TheMoonMaster joined15:11
blee does git use port 22?15:12
→ EnginA joined15:12
→ thiago_ny joined15:12
→ ruskie joined15:14
← EnginA left15:14
← EricInBNE left15:14
← angelsl left15:14
→ Error404NotFound joined15:14
← rmblr left15:15
← sambarino left15:16
← EnginAy left15:16
Silex blee: ssh does15:16
→ roop joined15:16
Silex git over ssh thus does15:16
→ Ramblurr joined15:16
← Ramblurr left15:16
→ Ramblurr joined15:16
blee okay15:16
← gnufied left15:16
→ OOPMan joined15:17
blee ahhhh yes, i know exactly whats wrong now15:17
rofl15:18
→ Hutch[S] joined15:19
→ alvaro_o joined15:20
← agile left15:20
← TheMoonMaster left15:20
→ MattDiPasquale joined15:20
→ priidu joined15:21
← Tommy[D] left15:21
← thiago_ny left15:21
← OOPMan left15:21
al it's really weird. i know these bare repo scripts used to work and still do on certain boxes15:22
but for this box i had to add a few --git-dir=$PWD15:22
→ thiago_ny joined15:23
← thiago_ny left15:23
→ thiago_ny joined15:23
→ bhosie joined15:24
→ bhosie_ joined15:25
← DrNick left15:25
← ereslibre left15:25
→ Henrik_ joined15:25
blee workmad3: cbreak_work: Yeah someone wrote a retarded rule into the firewall15:25
grabbed port 22 and redirected onto itself15:25
workmad3 blee: was it you? :P15:25
blee no15:26
workmad3 heh :)15:26
→ DrNick joined15:26
→ werdan7 joined15:26
← skoop left15:26
blee but now15:26
im laughing at a new guy who started yesterday15:27
who was trying to ssh to his box at home15:27
and it wasnt taking his password15:27
because he was using port 2215:27
I dont use 22 on any of my boxes for ssh, so never had the issue15:27
Cipher-0 Do you have a forty column monitor?15:27
blee lol15:27
← bhosie left15:28
← kreemer left15:28
bhosie_ → bhosie15:28
← henrikbjorn left15:28
← xiangfu left15:29
← Henrik_ left15:29
→ napster_123 joined15:29
→ montylounge joined15:30
← jxpx777 left15:30
→ Tommy[D] joined15:30
← alvaro_o left15:31
← al left15:32
← roop left15:32
← bhosie left15:33
← khmarbaise left15:34
→ ajpiano joined15:35
→ bhosie joined15:36
→ OOPMan joined15:37
← g0bl1n left15:37
→ booginga joined15:38
→ mw joined15:39
← m_3 left15:40
→ m_3 joined15:41
← workmad3 left15:41
→ jfkw joined15:42
← m_3 left15:44
← scp1 left15:44
← pr0wler left15:45
→ Yuffster joined15:45
← Cro_Crx left15:46
→ vano joined15:46
→ pr0wler joined15:46
← hobodave left15:46
← rgr_ left15:47
→ m_3 joined15:48
→ rgr_ joined15:48
→ sergiomc joined15:49
→ scp1 joined15:49
vano I have this code base with modules/components which compile different executables, right now I have separate git repos for all the modules and I just cherry pick them up for the projects. I also use tags to identify the versions for the exectuables. Thing is managing them gets to get a bit complicated manually, any tips on any existing scripts which use git to say do do something on several cherry picked repos at once? Or any other tips15:49
what is the best way to tackle this? Thanks15:49
← jony left15:50
← rgr_ left15:50
→ Zoresvit joined15:50
→ newbiz_ joined15:50
→ rgr_ joined15:51
offby1 dunno what you're talking about, unfortunately.15:52
"I just cherry pick them up for the projects" -- that means nothing to me15:52
← shorg left15:52
→ rayners joined15:53
vano for project a (executable a) I use moduleA, moduleB, but for project b (exectuable B)15:53
)15:53
offby1 sorry, dunno what you mean by "moduleA"15:53
vano bexecutable b) I use moduleB, moduleC ...15:53
→ ExtraSpice joined15:53
offby1 are you talking about git submodules?15:53
vano moduleA = git repo15:53
no.. just separate repos.15:53
offby1 so you've got a collection of separate git repositories.15:54
→ pgib joined15:54
vano correct.15:54
→ parth-vader joined15:54
← newbiz left15:54
offby1 and you need to combine their "outputs" (by which I mean, exectuables built from them)15:54
vano and each projects uses any one of them or a mix.15:54
offby1 I have no particular advice, except to note that the Android project works something like this --15:54
they have a whole crapload of separate repositories, and a big script called "repo" that somehow (like The Dude's rug) ties it all together15:55
might be worth spending 15 minutes reading about the problem that their "repo" script solves, and how it solves it15:55
→ mkramer joined15:55
← mkramer left15:55
vano well, no everything is ok. The problem is managing the whole thing manually. I use tags to pinpoint a specific version within each repo / module... and it gets a bit complicated to manually track which repo goes with which project and the tags themselves etc. So before I code something for this, wanted to know whether there is anything out there already or some other options.15:56
→ beatak joined15:56
→ Layke joined15:56
vano Can you point me to the project ? Name / url ?15:56
or you mean the generic Android sdk ?15:56
← Titosemi1 left15:56
← ymasory_ left15:57
→ Schmallon joined15:57
← newbiz_ left15:57
→ yhager joined15:57
offby1 not the SDK, I don't think; that's relatively small.15:57
hold on15:57
vano k15:57
Hutch[S] → Hutch[S]|away15:58
Layke I have really messed up with something, I don't know what. I made some changes yesterday while I was working away from my usual workspace, and I think I forgot to push those changes. Now I can't push them at all. I can't even pull to abandon those changes.15:58
→ jds joined15:58
← rgr_ left15:58
← raichoo left15:58
vano offby1: I have to go now for a while, I'll read up when you write something. Thanks.15:58
Layke I don't know where to start explaining the problem. I think git is also now stuck in the middle of a merge15:58
and I don't know what to do to fix?15:59
→ mike22 joined15:59
offby1 vano: nuts, I can't find what I thought I remembered15:59
→ rgr_ joined15:59
offby1 Layke: can you paste a transcript?15:59
PerlJam Layke: If you use git from the command line, you should look into using the git-completion.bash script to show you when you're in a git repo and what state it's in.15:59
offby1 also "gitk --all" is handy15:59
→ TheMoonM_ joined16:00
ChanServ set mode: +v16:00
Layke What am I looking for?16:00
mike22 hi, is there a way to get a list files that were added to the repo after a revision?16:00
i know about git diff --names-only but that shows all changes16:01
← m_3 left16:01
offby1 Layke: anything funny-looking :)16:01
I don't really know.16:02
← parth-vader left16:02
→ parth-vader joined16:02
Layke offby1: I don't know much about git really. All I ever do it git add , git commit, and git push/pull16:02
No idea how it all works16:02
offby1 mike22: I would think "git log --stat revision.."16:02
← parth-vader left16:02
← banisterfiend left16:02
← cptG left16:02
offby1 Layke: well, your best bet is to paste a transcript of the session where you "can't push"16:02
and then tell us how what you saw differed from what you expected to see16:03
→ seme joined16:04
← ptl left16:04
→ ymasory_ joined16:04
← marc_kdab left16:04
→ thierryp joined16:04
← mike22 left16:05
→ marc_kdab joined16:05
Layke offby1: http://pastebin.com/bjZKW8s016:05
seme hi guys... I have a project that I've created a git repo for on my local machine.... I want to have this stored on a central server since I access the project from several workstations... in the past I've just copied the directory to the server using tar/scp and then just did a git clone over ssh... is that the right way to do it?16:06
PerlJam seme: it's a way16:06
→ banisterfiend joined16:07
seme Is there a better way? I am open to change ;)16:07
PerlJam Layke: ah, you're already using git-completion.bash16:07
→ necromancer joined16:07
→ _psychic_ joined16:07
Layke PearlJam: I really don't know what that means.16:08
← saschpe left16:08
← bhosie left16:08
PerlJam Layke: your prompt tells you that you're on the master branch and that you're in the middle of a merge16:08
Layke I didn't tell it to do a merge. It went into a merge automatically16:09
WHen I did a pull a while ago16:09
PerlJam well, that's what pull does :)16:09
Layke and it detected that I was out of sync.16:09
So what can I do?16:09
PerlJam first it fetches remote refs, then it merges16:09
Layke: what does "git status" say?16:09
offby1 Layke: if you don't have any valuable uncommitted work, "git reset --hard" will get you out of the "in the middle of a merge" state.16:09
seme Layke: perhaps you just wanted to do a git fetch which will just grab the changes but not merge them16:10
← darkling left16:10
Layke I have about an hours worth of work16:10
offby1 ok, then don't do it :)16:10
Layke I've probably spent that long on this problem16:10
offby1 well, that's not what I meant; I meant changes to files that haven't been committed, and that you'd hate to lose.16:10
→ lvianna joined16:10
← lvianna left16:10
→ m_3 joined16:11
→ lvianna joined16:11
offby1 I _think_ "git reset --keep" is safe.16:11
hagebake → hagabaka16:11
Layke This: http://pastebin.com/0SARPMtk16:11
offby1 at least, so I've been told.16:11
→ Scorpion joined16:11
← giallu left16:11
← apalto left16:11
Layke That was to PearlJam @16:11
offby1 "git stash" will save those changes16:11
→ bduncan_ joined16:11
Layke Okay, I have done a git stash.16:12
← bduncan_ left16:12
Layke I am now out of MERGING.16:12
← bduncan left16:12
Layke Now how do I force what I have to be the new HEAD on the remote master?16:12
← unon left16:13
→ ashleyw joined16:13
offby1 well, "git push -f" _might_ do it16:13
but I am not sure that's a good idea16:13
since that will more or less lose data on the remote16:13
ashleyw Hey, does anyone have any idea what the issue is if I get "/usr/libexec/git-core/git-sh-setup: fork: Resource temporarily unavailable" when pulling?16:13
Layke mm, I'm not sure why there would be data on the remote16:13
Which I can't afford to lose16:14
offby1 Layke: usually, you would _merge_ what's currently on master into what you have, so that what you push includes all the original stuff.16:14
Layke Because the latest version SHOULD be this version I have locally16:14
offby1 the simplest explanation is: someone else has pushed something to master, which you now should integrate16:14
Layke offby1: But the stuff I have locally is the most up to date. So wouldn't the merge overwrite what I have done16:14
offby1 Layke: look at "gitk --all". That will show you what's what.16:14
Layke That was why it says that #Your branch is ahead of 'origin/master' by 1 commit.16:14
offby1 merges don't overwrite; instead they ... merge :)16:15
ashleyw I've dumped the local repository and cloned it from the server, and the issue still happens. I'm stumped, I can't even find anything via Google?16:16
← jauler_ left16:16
Layke offby1: I don't understand what the gitk console shows. This is a SS http://chrisacky.com/git.jpg16:17
As you can see, I dont take commit messages very seriously16:17
:P16:17
offby1 that console looks perfectly healthy.16:17
I would think "git pull" now would work smoothly; try it16:17
← cbreak_work left16:17
← thiago_ny left16:18
Layke And doing that I will overwrite the local amendments?\16:18
offby1 nope16:18
your data is safe16:18
offby1 crosses fingers16:18
Layke didn't work. I will show you the log16:18
→ thiago_ny joined16:18
← ayanes left16:19
offby1 :-(16:19
Layke http://pastebin.com/iQKNmXba16:19
← ISF left16:19
offby1 actually that _did_ work16:19
← darkredandyellow left16:19
offby1 you just have to fix the conflict16:19
← Tommy[D] left16:19
offby1 I wonder if this is what happened last time.16:19
Layke I am still in (master|MERGING)16:20
now too16:20
offby1 sure16:20
just edit that file, look for the <<<<<<<<<<<<< and >>>>>>>>>>>>>>>>16:20
and fix the code as needed16:20
it _should_ be obvious how to fix it.16:20
unless you've never used a revision control system before, in which case, may God have mercy on your soul.16:20
Just sayin'.16:20
once you've edited the code so that it looks right and passes your tests (you do have tests, don't you?!), do "git add" on the file, then 'git commit'16:21
Layke Here is an example.. http://pastebin.com/mM5Zk9Z216:22
Thats what the file has.. there are about 10 of those..16:22
→ Tommy[D] joined16:22
Hutch[S]|away → Hutch[S]16:22
Layke So I just edit the file so that it looks like proper PHP code.16:22
I don't leave in the <<<<<<<<<< HEAD junk do I?16:22
offby1 nope16:23
that ain't proper PHP code!16:23
10 is rather a lot, unfortunately.16:23
Do you at least understand why this happened?16:23
Layke No I don't.16:23
offby1 oh.16:23
well, you and ... someone else ... have edited the same bits of that file, in different ways, at more or less the same time.16:24
Thus you and that other person, in effect, disagree on how that file should look.16:24
Layke So on two different workspaces,16:24
we edited the file16:24
then we both did a commit?16:24
offby1 You, or perhaps the two of you working together, need to figure out what the file should look like; git cannot guess.16:24
yup, you've got it.16:24
→ joel_falcou joined16:24
Layke offby1: I'm the only developer working on this file.16:24
So it is all my doing.16:24
→ ChUrP joined16:25
offby1 well, then, you in some other "branch" or "repo".16:25
Since both the people involved are you, it ought to be easy to figure out how to "resolve" the conflict.16:25
→ bhosie joined16:25
Layke What I do is.. I code at home16:25
offby1 sure16:25
Layke Then I push my changes to my remote16:25
offby1 sure16:25
Layke Then, I go to the place where the code runs, (as an intranet site)16:25
Then I did a pull16:25
offby1 sure sure sure16:26
Layke Then I saw that something was wrong, so I changed the file yesterday16:26
offby1 yeah yeah whatever.16:26
Layke At the local where the site runs as intranet.16:26
offby1 Just edit the file to look how you want it to look.16:26
then 'git add' it, and then "git commit".16:26
Layke Yup cheers. Thanks for your help.16:26
Should all be okay then16:26
→ tjholowaychuk joined16:27
offby1 should be16:27
Scorpion question, is it possible to put a reference to another git repo to get pulled down when I do a pull from my repo?16:27
offby1 conflicts are a fact of life; better get used to 'em16:27
Scorpion: more or less, yes16:27
that's called a "submodule"16:27
Scorpion gotcha, wasn't sure what to google16:28
thanks!16:28
offby1 unfortunately you have to issue a couple of extra commands to get it pulled down, but it's mostly easy16:28
Scorpion perfect, that will make life much easier :)16:28
Layke offby1: I made the change.. but I am still in MERGING16:29
So I can't a a commit, because it says.. : cannot do a partial commit during a merge.16:30
← jefferai left16:30
Layke So then I did a git stash.. and did a full circle of the problem. That didn't do anything16:30
So now I am back where I was. I did a new pull again16:30
and then fixed my file. And I need to some how leave (master|MERGING)16:30
← andrewvo1 left16:31
offby1 Layke: transcript, please16:31
the way to leave the master|MERGING state is to 1) fix the conflicted file; 2) run 'git add' on that file; 3) run "git commit".16:31
→ jefferai joined16:31
Layke http://pastebin.com/mM5Zk9Z216:31
→ goshawk joined16:31
offby1 that's not a transcript16:32
that's just a conflicted file.16:32
← MattDiPasquale left16:32
← necrodearia left16:32
offby1 obviously you need to choose one of those two lines.16:32
Layke Yes I did16:33
I deleted the stuff around it tho?16:33
The <<<<<<< HEAD16:33
and the long numbers >>>>>>>>>>>16:33
→ zombor joined16:33
wereHamster you need to delete that too16:33
basically make the file look how you want it to16:33
Layke Yeah, I deleted everything to make my file run perfectly16:34
wereHamster then add the file (git add) and then 'git commit' (without any arguments)16:34
offby1 that advice sounds familiar ... where have I heard it before ...16:35
Layke That worked. Things it is all good.16:35
git commit .16:35
didn't work16:35
git commit worked16:35
offby1 of course16:35
refund When I merged i got a "CONFLICT (content)...", how do I open an editor to resolve this?16:35
offby1 *sigh*16:35
refund: I don't want to be mean, but if you don't know how to use a text editor, what the heck are you doing using git?16:35
→ FilipeMaia joined16:35
refund offby1, Of course I do, i just want to get a view of the diff16:36
← posciak left16:36
offby1 so ... edit the conflicted file. What's the mystery?16:36
← ChUrP left16:36
Layke Thanks for help offby1.16:36
offby1 Layke: yay16:36
refund offby1, What is the cause of the conflict+16:36
?*16:36
→ rctay joined16:36
offby1 refund: the cause is: you, and someone else edited the same file, in different branches, at more or less the same time.16:36
the someone else might actually have been you, in a different repo.16:37
Layke just went through exactly this.16:37
→ madewokherd joined16:37
wereHamster refund: http://en.wikipedia.org/wiki/Conflict_(revision_control)16:37
→ kpfleming joined16:37
→ MattDiPasquale joined16:37
refund offby1, obviously, but i want to see what someone else have done, and what i have done16:37
kpfleming this may be a FAQ, and if so please point me to something that describes it... but if i have a branch B that contains a bunch of commits in various files, and i want to create a new branch C that contains only selected hunks from some of those commits, are there git tools (beyond git add -p after applying patches to the working tree) that could help?16:38
→ jxpx777 joined16:38
offby1 refund: ah, you probably want to run 'git mergetool'16:38
it will launch a three-way-diff thingy, if you're lucky16:38
refund offby1, thanks16:39
→ giallu joined16:39
← scelis left16:39
stew does it really depend on luck? :)16:39
→ scelis joined16:39
← smolyn left16:40
→ ilteris___ joined16:40
Hutch[S] → Hutch[S]|away16:41
← smuf left16:41
← teweWork left16:41
→ carutsu_ joined16:41
→ morganick joined16:42
→ pheaver joined16:42
← pheaver left16:42
→ ChUrP joined16:42
→ teweWork joined16:43
← dfr|mac left16:43
→ ashleyw_ joined16:43
← ashleyw left16:44
ashleyw_ → ashleyw16:44
← mastro left16:44
← Schmallon left16:45
← Layke left16:45
→ Layke joined16:45
← FilipeMaia left16:46
→ amondracorp joined16:47
refund So i did git mergetool and it launched vimdiff. I've read a little on how to navigate across the files, but how do I pick one file to use/save? :o16:47
funsheep Hi, me and a coworker just started to test gerrit as an alternative to svn. We both use Eclipse and EGit. We cloned a repository from gerrit to our local computers, added those to eclipse. Then we made changes and when we try to push those changes back to gerrit, we get "You are not commiter work@…". "work" is my local login name, but i never configured that. How can i change this commiter name? To what do i have to change that? I'm lost. Tha16:47
→ noahcampbell joined16:47
→ robotarmy joined16:48
← robotarmy left16:48
→ robotarmy joined16:48
→ rorist joined16:48
← goshawk left16:49
← kpreid left16:49
refund Do I need to edit one of the views and just save?16:49
selckin probably a good idea to learn git a bit before starting with gerrit16:49
→ ashleyw_ joined16:49
→ ben225 joined16:50
← serard left16:51
← ashleyw left16:51
ashleyw_ → ashleyw16:51
Scorpion offby1: that did the trick perfectly, thanks much!16:51
← Scorpion left16:52
refund So my vimdiff changes failed completely, how can i revert? :(16:53
→ nadavoid joined16:53
refund I have lots of untracked files not, they end with .BASE..., .LOCAL... and .REMOTE...16:54
now*16:54
wereHamster delete them16:54
refund but the merge turned out badly16:54
the file did not end up as i expected16:54
curtana git reset --merge16:55
if you want to go back to where yo uwere before you tried to merge16:55
refund thanks16:55
exactly16:55
curtana i think that is a new-ish feature. if you have an old git then you might not have it16:55
← nevyn^1 left16:55
refund is there any shortcut command that will remove all untracked files?16:56
← v0n left16:57
→ carpet_the_walls joined16:57
wereHamster refund: man git-clean16:57
jast refund: the 'git-clean' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clean.html [automatic message]16:57
refund thanks16:57
← marc_kdab left16:57
→ Fullmoon joined16:57
→ marc_kdab joined16:58
← robotarmy left16:58
carpet_the_walls if i create a branch for some experimental work, and it doesn't work out, and i delete the branch, is its history preserved? do the files still take up space in the repository? thanks16:58
wereHamster carpet_the_walls: for a short while (30 or so days)16:58
← fivetwentysix left16:58
wereHamster however you shouldn't worry about it16:59
offby1 *nod16:59
→ mastro joined16:59
carpet_the_walls wereHamster: oh, didn't realise git did anything that was time dependant16:59
← SvenDowideit left16:59
→ ashleyw_ joined16:59
→ saschpe joined16:59
carpet_the_walls wereHamster: thanks16:59
← noahcampbell left17:00
→ SvenDowideit joined17:00
wereHamster carpet_the_walls: man git-gc17:00
jast carpet_the_walls: the 'git-gc' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-gc.html [automatic message]17:00
← codebeaker left17:01
← ashleyw left17:01
ashleyw_ → ashleyw17:01
← ki__ left17:01
→ eletuchy joined17:02
Silex → silex17:03
→ nevyn^1 joined17:03
← funsheep left17:03
→ ceej joined17:03
→ consumerism joined17:04
consumerism i have the following alias, i'm wondering if there is a shorter/better way to recursively initialize and update all submodules:17:04
alias gsm='git submodule init && git submodule update && git submodule foreach --recursive "git reset --hard && git submodule init && git submodule update"'17:04
← teweWork left17:04
→ FilipeMaia joined17:05
← envi left17:05
→ elmob joined17:06
← marc_kdab left17:07
← kipibenkipod left17:08
→ kipibenkipod joined17:08
frakturfreak_ → frakturfreak17:09
← mase_wk left17:09
← dc5ala left17:09
← MattDiPasquale left17:09
← Badiss left17:10
→ mase_wk joined17:10
← mase_wk left17:10
→ level09 joined17:10
→ mkramer joined17:12
→ hobodave joined17:12
← kipibenkipod left17:13
→ ChUrP_ joined17:13
→ kipibenkipod joined17:13
← frakturfreak left17:14
→ pigdude joined17:14
→ sshc joined17:14
← ChUrP left17:15
→ ajpi joined17:15
← pr0wler left17:15
→ bkosborne joined17:15
← SirFunk_ left17:15
yhager is there a version of post-receive-email that sends email on push of tracking branches too?17:16
← ajpiano left17:16
→ MattDiPasquale joined17:16
← Mpenz left17:16
← nevyn^1 left17:16
← rayners left17:17
← pigdude left17:17
→ spearce joined17:18
→ Andy19781 joined17:18
→ v0n joined17:18
← kpfleming left17:18
← airborn left17:18
→ bauruine joined17:18
→ Mpenz joined17:19
FauxFaux glares at git clone -b ignoring the ERROR of the branch not existing.17:19
← rapidmax left17:21
→ smolyn joined17:22
yhager post-receeive-email bails out with *** Push-update of tracking branch, $refname, *** - no email generated.17:22
← v0n left17:22
← Andy19781 left17:23
← hyperair left17:23
yhager (or at least that's what I think is the reason I am not getting any emails)17:23
→ hyperair joined17:24
→ henrikbjorn joined17:24
← henrikbjorn left17:24
← rgr_ left17:25
→ henrikbjorn joined17:25
→ ptl joined17:25
→ frakturfreak joined17:26
← danlucraft left17:27
→ workmad3 joined17:27
→ rgr_ joined17:27
yhager in git-svn, I am pushing to a git bare after 'git svn fetch', and would like to get emails when new refs are pushed. is this possible?17:27
→ Buglouse joined17:27
← rgr_ left17:28
← xiong left17:28
→ steph021_h joined17:28
← steph021 left17:28
→ rgr_ joined17:29
→ abetaha joined17:29
→ transplant joined17:29
← eletuchy left17:31
→ _ollie joined17:32
→ spaceonline joined17:33
← charon left17:33
→ shentino joined17:33
← t0rc left17:34
→ anon33_ joined17:34
→ charon joined17:35
← charon left17:35
→ charon joined17:35
← LongBeach left17:35
→ Schmallon joined17:35
→ gplocke joined17:35
← banisterfiend left17:37
← abetaha left17:37
← henrikbjorn left17:37
← codespectator left17:37
→ towski joined17:37
→ fr0sty-away_ joined17:37
→ acid joined17:38
← thierryp left17:38
← fr0sty-away left17:38
← FilipeMaia left17:39
← refund left17:39
← bentob0x left17:40
→ toddr joined17:40
→ QshelTier joined17:40
→ Andy19781 joined17:40
← QshelTier left17:40
→ QshelTier joined17:40
← Schmallon left17:40
→ ISF joined17:40
→ abetaha joined17:40
→ kenichi joined17:40
toddr is there a way with git log to determine when a file's permissions were changed?17:40
← _ollie left17:40
→ dfr|mac joined17:40
toddr permissions seems to be a poor stepchild in git status and git log output17:40
→ amerine joined17:41
→ guardian joined17:41
gplocke I'm struggling with trying to figure out how best to manage the repo (or repos) for this templating project we're working on. We've basically got 4 HTML designs that we're turning into Drupal, Wordpress, etc. themes.17:41
→ alexmchale joined17:41
gplocke We want to manage them in git, but we don't know if it's better to have a repo for each design and then just have branches for Drupal, Wordpress, etc. or what. Any suggestions or advice?17:41
← rctay left17:42
← bhosie left17:42
bburhans_ → bburhans17:42
← priidu left17:43
← QshelTier left17:43
→ QshelTier joined17:43
→ shruggar joined17:43
← harleypig left17:43
← QshelTier left17:43
→ QshelTier joined17:43
← ashleyw left17:43
→ bentob0x joined17:44
← jxpx777 left17:45
→ bdiu joined17:46
→ Schmallon joined17:46
← alnewkirk left17:46
→ jrochkind joined17:47
← TheMoonM_ left17:47
jrochkind hey, random question maybe someone knows.17:47
With a command line commit message, -m rather than using an editor, is there any way to put newlines in the commit message? Just sticking a "\n" in there doesn't seem to work.17:47
→ hagebake joined17:47
← hagebake left17:47
→ hagebake joined17:47
→ skoop joined17:48
← Bombe left17:48
jrochkind hey, look at that, in my shell anyway, the answer is--- just hit return! the shell wont' execute the command if you're in the middle of a quoted string, it'll wait for the end quote. Okay then.17:48
gplocke Cool. What shell are you in?17:48
→ v0n joined17:49
gplocke Hey, mine does too! On OS X.17:49
I had no idea17:49
→ marc_kdab joined17:49
← carpet_the_walls left17:50
← resmo left17:51
← hagabaka left17:51
← ChUrP_ left17:52
QshelTier → Bombe17:53
← v0n left17:53
charon jrochkind: then again you might just want to use the editor instead :)17:53
→ ChUrP joined17:53
← rgr_ left17:53
← _ruben_ left17:54
scelis consumerism: git submodule update --init --recursive17:54
→ rgr_ joined17:54
→ pheaver joined17:54
← rgr_ left17:55
← mziulu left17:55
← level09 left17:55
→ priidu joined17:55
→ dr_win_ joined17:56
← LouisJB left17:57
← eno__ left17:57
← dr_win left17:57
→ FilipeMaia joined17:58
← DiGitalX left17:58
← FilipeMaia left17:58
→ __name__ joined17:58
← __name__ left17:58
→ __name__ joined17:58
→ eno joined17:59
→ rgr_ joined17:59
← eli left17:59
→ FilipeMaia joined18:00
→ robotarmy joined18:01
← shruggar left18:02
→ SimonP86 joined18:02
← FilipeMaia left18:04
→ eletuchy joined18:04
→ bhosie joined18:06
→ sraue_ joined18:06
← frooh left18:08
← sraue left18:09
← bkosborne left18:09
→ justin-george joined18:10
← toddr left18:10
← DiaPo left18:10
→ v0n joined18:11
→ DiGitalX joined18:11
→ FilipeMaia joined18:12
→ alberto56 joined18:12
← FilipeMaia left18:12
← kipibenkipod left18:12
→ remur_030_v2 joined18:12
← rgr_ left18:13
yhager gplocke: why not just a single repo and a directory for each cms?18:14
gplocke we've got 4 different designs, so we figured it'd make the most sense to have a single design per repo and then branches for each CMS18:14
← sraue_ left18:15
→ sraue joined18:15
← transplant left18:15
gplocke we want to ensure it's simple for people to contribute to and doesn't require a whole lot of extra manual work18:15
amerine → amerine|away18:16
→ funsheep joined18:16
← remur_030_v2 left18:16
← v0n left18:16
→ remur_030 joined18:16
→ rgr_ joined18:16
→ v0n joined18:17
← acid left18:17
remur_030 Hi sitaram, gitolite removes 'git-daemon-export-ok' from the repositories when it updates them, is there a way to avoid that?18:17
→ benishor joined18:17
← moonk left18:18
→ TheFox21 joined18:18
→ dreiss joined18:19
→ codebeaker joined18:19
→ DiaPo joined18:19
sitaram remur_030: yes, grant access to "daemon" user for that repo18:19
← codebeaker left18:19
← jrochkind left18:20
sitaram remur_030: if you want to do *all* repos, "repo @all <newline> R = daemon" ought to do18:20
← DiaPo left18:20
yhager gplocke: up to you. git will support whatever scheme you'll think of18:20
→ irc2samus joined18:20
irc2samus hi guys, I forgot how to get the absolute path to the root of the repo anyone remembers?18:20
gplocke :) yeah, I know it will, but I was hoping for some recommendations or best practices18:21
remur_030 sitaram: aaah nice, I remember that was on the gitweb part, maybe add that to the http-backend doc too?18:21
→ bkosborne joined18:21
→ DiaPo joined18:21
← v0n left18:22
sitaram remur_030: I did not understand... the http-backend doc is about how to use http instead of ssh for authentication. Why would that contain stuff that is common to both types?18:22
→ dunpeal joined18:23
remur_030 sitaram: I am Karsten from the mails, I wanted to avoid setting SetEnv GIT_HTTP_EXPORT_ALL and select single repos18:24
← dunpeal left18:24
remur_030 so for dual use it might be interesting to mark a few repos ssh only18:24
or rather mark a few repos http accessible18:25
→ LouisJB joined18:25
irc2samus ahh git rev-parse --show-toplevel that's it :)18:25
amerine|away → amerine18:25
← amerine left18:26
ajpi → ajpiano18:26
← amondracorp left18:26
→ normanrichards_ joined18:26
Weust` → Weust`afk18:26
← normanrichards left18:27
normanrichards_ → normanrichards18:27
→ M1ghtyDOC joined18:27
sitaram remur_030: thats not how it works. Before I explain, do you know the difference between authentication and authorisation?18:29
otherwise I cant explain18:29
→ v0n joined18:29
remur_030 yes, it's 2 steps, authentication first (do we know the user, and if yes is he who he claims) and second authorization (can he perform the requested action)18:29
second authorisation!18:30
scelis sitaram: heya. wonderful job with gitolite. it's been super-easy to setup on our two servers. just curious, though, what does "pu" stand for? :)18:30
→ sako joined18:30
← irc2samus left18:30
sitaram remur_030: yes, and the http versus ssh thing stops at authentication. Authorisation is a separate and orthogonal activity -- we know what "user" it is, but we don't (or shouldn't) care how he came in18:31
sako is it possible that a developer commits and pushes something out and it somehow disappears18:31
sitaram scelis: proposed updates (same convention as in the main "git" project, as Junio uses it)18:31
sako when he later pulls18:31
sorry for the weird question, that is what a developer here is claiming :/18:31
→ dstorrs joined18:31
remur_030 sako: sounds more like he checked out another branch and forgot18:32
scelis sitaram: ah, thanks. i figured it was something common.18:32
remur_030 or someone else upstream changed history18:32
or someone else changed upstream history18:32
→ peshalto joined18:32
sako how in the world can someone change upstream history?18:33
→ dunpeal joined18:33
scelis sako, remur_030: it still shouldn't become lost even with changed history. youd get a conflict on the pull18:33
remur_030 sako: the same as you would do on your local copy, git reset18:33
scelis: I guess that is what the developer claims18:33
ChanServ set mode: +v18:33
dunpeal Hey. When I do `git push origin foo`, does that mean "push my local HEAD to remote 'origin' as branch foo"?18:33
zivester_ if i do a `git checkout -b newbranch origin/master` will it automatically fetch from the remote repo any updates that I may not have in my local master ?18:33
← matjas left18:34
dunpeal zivester_: no.18:34
scelis dunpeal: no18:34
← v0n left18:34
dunpeal zivester_: the only operation that fetches updates from a remote is `git fetch` (and derived operations like `pull` that call fetch as part of their operation)18:34
remur_030 sitaram: couldn't one use "git-daemon-export-ok" for authorisation as well? as in: you came in the wrong way buddy?18:35
dunpeal scelis: so what does it mean, exactly?18:35
zivester_ so i need to `git fetch origin master` first... ok18:35
scelis dunpeal: look in 'man git-push'. towards the end. under examples. see "git push origin master"18:35
jast dunpeal: the 'git-push' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-push.html [automatic message]18:35
← tvw left18:35
scelis dunpeal: thats a pretty good description for what exactly that command does18:35
dunpeal scelis: thanks.18:35
remur_030 because maybe http authorization and ssh authorization both yield the same username, but I don't want him to use http18:35
← hagebake left18:35
← morganick left18:35
dunpeal zivester_: you should be able to just do `git fetch` actually.18:36
← petrux left18:36
← _Unode left18:36
dunpeal in most cases there's no reason to specify `origin master`18:36
→ ph^ joined18:36
zivester_ git fetch will download all updates from all branches of all repos, yes ?18:36
→ oracal joined18:36
dunpeal zivester_: the default refspec does, yes.18:37
sitaram remur_030: then dont give him a http user id I guess... your solving the problem at the wrong level. It's not clean18:37
dunpeal you can provide or set (in .git/config, via git-config) a refspec that only regards certain refs, but the default one fetches all heads.18:37
→ odus joined18:38
sitaram remur_030: besides, g-d-e-o has a defined meaning in git. I'm not going to subvert it :)18:38
← gebi left18:38
remur_030 sitaram: we use ldap, I can't just take away his http user access, it's not only for git18:38
→ theoros joined18:38
← theoros left18:38
→ theoros joined18:38
→ koo5 joined18:39
→ normanrichards_ joined18:39
→ Zephyrus joined18:39
sitaram remur_030: let me get this right. You have a set of users who have equal ability to authenticate via http and via ssh. You want to run gitolite on both http and ssh. And then you want to distinguish between those accesses and make a further access control distinction.18:39
← ben225 left18:39
→ ISF_ joined18:39
→ engla joined18:39
→ zenazn joined18:40
→ hagebake joined18:40
← hagebake left18:40
→ hagebake joined18:40
← engla left18:40
→ engla joined18:40
sitaram remur_030: do you want read access control or is it sufficient to block writes?18:40
remur_030 sitaram: right, I want to flag special repos where I want people to use the more secure ssh path and not their http access18:40
→ matjas joined18:40
remur_030 but on others I don't care18:41
sitaram and you want to do this "flagging" by using g-d-e-o? again... for read also or only for write?18:41
remur_030 any access over http should be blocked18:41
→ ashleyw joined18:41
remur_030 to this specific repository that is18:41
sitaram remur_030: for reads I cant help but if you want to block right for such repos you just add a update.secondary hook that says "if g-d-e-o file exists and one of the HTTP_ env vars exist, exit 1"18:42
s/right/write18:42
← Andy19781 left18:43
← odus left18:43
← normanrichards left18:43
normanrichards_ → normanrichards18:43
sako remur_030: so it looks like people def tried to reset/revert changes18:43
← ChUrP left18:43
→ ChUrP joined18:44
← ChUrP left18:44
sako remur_030: do you know whats the best way to tell what happened?18:44
→ henrikbjorn joined18:44
← ashleyw left18:44
sako I guess, best way to tell if history was erased/18:44
cbreak you want to know wtf happened? Check the reflog :)18:45
→ v0n joined18:45
remur_030 sitaram: couldn't I just mark the repos where I don't mind http access with daemon?18:45
cbreak .htaccess? :)18:45
remur_030 sako: something like gitolite would limit reverts/resets to authorized people18:45
← jfkw left18:46
→ ChUrP joined18:46
cbreak it is impossible to distinguish a revert from a normal commit18:46
remur_030 also you could just disallow reset and it's buddies on your git host18:46
→ codetonowhere joined18:46
← psoo left18:46
← madewokherd left18:46
remur_030 but in case someone reset the branch to a certain commit now you can't see who did it I think18:46
← napster_123 left18:46
← robotarmy left18:47
→ keystr0k_ joined18:47
← ningia left18:47
← keystr0k_ left18:47
→ ashleyw joined18:48
→ TheMoonMaster joined18:48
→ madewokherd joined18:48
→ robotarmy joined18:48
← keystr0k left18:49
→ vals_ joined18:50
← v0n left18:50
← hagebake left18:51
← tango_ left18:51
← ChUrP left18:51
← ChrisPitzer left18:52
→ FilipeMaia joined18:52
→ SandCube joined18:52
→ codespectator joined18:52
← necromancer left18:53
SandCube how to work with mysql databases with git? Is there any other tool best for database version control?18:53
→ _ollie joined18:53
cbreak SandCube: don't18:53
← _ollie left18:55
vals_ → tango_18:55
→ hagebake joined18:55
remur_030 sitaram: as an example I don't want anybody beeing able to access gitolite-admin over http, so I wouldn't put 'R daemon' into his list, couldn't this be understood as a second level of authorization?18:55
← bluenovember left18:55
← PigeonFriend left18:56
← rgr left18:56
remur_030 the user authenticated successfully over http -> has the right to perform action on gitolite-admin -> doesn't have the right to do it over http18:56
→ ChUrP joined18:57
sitaram remur_030: here's what I'm willing to do: http://pastebin.com/JM2V084618:57
remur_030: and that only because it's generic enough to be useful for a lot more convoluted schemes. I honestly don't like what you suggest being in core gitolite18:58
→ Evious joined18:58
sitaram remur_030: but you will find that works fine; I'll work on the code tomorrow (actually it's be a few lines; documentation always takes more time)18:59
→ v0n joined18:59
→ alarkin joined19:01
remur_030 sitaram: so I would check in that script if the specified repo is using http access or not and return 0 / !0 accordingly right?19:02
← ashleyw left19:02
sako so here is the issue, I found out where the failed "revert" took place in gitk --all19:02
it says Reverts commit #########19:02
← DiGitalX left19:02
sako and it looks like it took the repository way back to that commit #? is that what revert is supposed to do?19:02
cbreak no19:03
tango_ no19:03
sako so anything in between is not showing up in the current repo.. i want to go 1 commit before this revert19:03
cbreak revert undoes exactly one commit19:03
remur_030 I guess this is totally fine, but I don't see why the process bailing out in the git-http-backend would be so bad? because gitolite wasn't the last line of defense?19:03
→ goshawk joined19:03
← tatsuya left19:03
cbreak as I said above, some time ago19:03
you can not distinguish a revert from a normal commit19:03
because a revert IS a normal commit19:03
sitaram remur_030: bailing out in git-http-backend? That's not my program, how do I make it do that?19:03
cbreak it does not modify history, it just undoes the changes a commit introduced19:04
← v0n left19:04
remur_030 sitaram: git-http-backend bails out if there is no g-d-e-o19:04
sako cbreak: so the only way to go back to a previous commit is to git reset?19:05
← Error404NotFound left19:05
cbreak define "go back"19:05
undo history?19:05
reset's the best way to do that19:05
sako theres a chunk of changes in between HEAD and the commit in question that was reverted that does not show up in the code19:05
sitaram remur_030: git-http-backend is not my program; not part of gitolite. It's part of git proper19:05
cbreak if it was reverted, there's a revert commit19:06
← miloshadzic left19:06
cbreak just look at the log19:06
sako cbreak: ya i see the revert commit19:06
remur_030 sitaram: I am aware of that, so it only fails after you authorized the access and handed the further action to git-http-backend19:06
sako problem is there is 2 revert commits19:06
→ keystr0k joined19:06
remur_030 s/after you/after gitolite19:06
sitaram remur_030: best send me email, because I am not able to understand what you're getting at. Must be also partly because it's kinda late and I'm sleepy19:07
sako cbreak: let me take a pic of my gitk all19:07
remur_030 sitaram: will do, thanks19:07
sako cbreak: https://skitch.com/svarozian/rkf8c/gitk-golfchannel19:08
cbreak: if you have a moment, can you please help me look at that?19:08
→ ashleyw joined19:08
← dfr|mac left19:09
← aah left19:09
cbreak what a mess19:09
→ linusoleander joined19:09
sako you can see there are 2 Revert "config corrections", and if you look to where config corrections is down below there under Fixes for ie6, in between the revert and config corrections all that stuff is not showing up in the code19:09
cbreak so, where's the problem?19:10
sako cbreak: tell me about it, i get called in after the shit hits the fan :(19:10
cbreak just look at the diff19:10
and if that doesn't help, look at the files of that commit19:10
→ dfr|mac joined19:10
sako for which one?19:10
cbreak both19:11
→ enherit joined19:11
cbreak git show hashoffirst19:11
git show hashofsecond19:11
← dunpeal left19:12
← robotarmy left19:13
← Bombe left19:13
← frakturfreak left19:13
← Yuffster left19:13
← ilteris___ left19:13
← Maletor left19:13
← ciaranm left19:13
← sycofly left19:13
← bobhope left19:13
← redtricycle left19:13
← Guest48246 left19:13
← [exa] left19:13
← jjuran left19:13
← quadz left19:13
← kelvie left19:13
← markatto left19:13
← resixian left19:13
← blathijs_ left19:13
← tmokros left19:13
← spb left19:13
← esc left19:13
← rudi_s left19:13
← phishy left19:13
← lotyrin left19:13
← mjv left19:13
← pasky left19:13
← yoh left19:13
← seme left19:13
← alarkin left19:13
← oracal left19:13
← marc_kdab left19:13
← ptl left19:13
← SandCube left19:13
← Guest26619 left19:13
← an0maly left19:13
← ayust left19:13
← jdav_gone left19:13
← squentin left19:13
← qknight left19:13
← RandomSort left19:13
← tro left19:13
← Khisanth left19:13
← bauruine left19:13
← sergiomc left19:13
← ruskie left19:13
← kanru left19:13
← Chillance left19:13
← Kyso left19:13
← mabrand left19:13
← muep left19:13
← GodEater_ left19:13
← Ratler left19:13
← Gunni left19:13
← hegge left19:13
← goshawk left19:13
← TheMoonMaster left19:13
← alberto56 left19:13
← justin-george left19:13
← ajpiano left19:13
← tjholowaychuk left19:13
← thiago_ny left19:13
← tizbac left19:13
← Agos left19:13
← Pacman_Home` left19:13
← soidexe left19:13
← MetaCosm left19:13
← sejo left19:13
← HopsNBarley left19:13
← Elfe_ left19:13
← themarmot left19:13
← MacGyverNL left19:13
← NickyP left19:13
← comcor- left19:13
← Juri_Agarin left19:13
← cbz left19:13
← stepnem left19:13
← Tronic left19:13
← multi_io left19:13
← lirakis left19:13
← ecksun left19:13
← j0ran left19:13
← metze left19:13
← alexmchale left19:13
← guardian left19:13
← carutsu_ left19:13
← yhager left19:13
← mw left19:13
← saurik left19:13
← alip left19:13
← dvaske left19:13
← ludde left19:13
← drry left19:13
← Risar left19:13
← jaeckel left19:13
← rootard left19:13
← m0 left19:13
← funfun left19:13
← womble left19:13
← blee left19:13
← mcjb left19:13
← halfline left19:13
← dean[w] left19:13
← tangledhelix left19:13
← blalor_afk left19:13
← mrtazz left19:13
← normanrichards left19:13
← benishor left19:13
← Mpenz left19:13
← zombor left19:13
← ExtraSpice left19:13
← booginga left19:13
← jaql left19:13
← bburhans left19:13
← frenzz left19:13
← ormaaj left19:13
← jvv left19:13
← albel727_3 left19:13
← jksz left19:13
← Swimming_bird left19:13
← Arrowmaster left19:13
← kandinski left19:13
← avar left19:13
← Derander left19:13
← Purdy left19:13
← divVerent left19:13
← robinsmidsrod left19:13
← elly left19:13
← robinbowes left19:13
← sven^ left19:13
← gbjk left19:13
← Leonidas left19:13
← ISF_ left19:13
← TheFox21 left19:13
← SimonP86 left19:13
← abetaha left19:13
← anon33_ left19:13
← mkramer left19:13
← saschpe left19:13
← romeo404_ left19:13
← felipe left19:13
← Ryback_ left19:13
← drizzd_ left19:13
← doener left19:13
← leavittx left19:13
← tewecske left19:13
← fr0sty left19:13
← timj left19:13
← fr| left19:13
← piti left19:13
← sunn` left19:13
← tlonim left19:13
← Teo` left19:13
← teuf left19:13
← selckin left19:13
← tomaw left19:13
← dfr|mac left19:13
← Evious left19:13
← skoop left19:13
← bentob0x left19:13
← spearce left19:13
← MattDiPasquale left19:13
← hobodave left19:13
← jds left19:13
← scp1 left19:13
← OOPMan left19:13
← Ramblurr left19:13
← Cipher-0 left19:13
← tomboh left19:13
← kelvie_ left19:13
← Unknown[OFF] left19:13
← Ingmar left19:13
← ThomasLocke left19:13
← Vampire0 left19:13
← HuntingBears left19:13
← AlexanderS left19:13
← joelpet left19:13
← kenyon left19:13
← plopix left19:13
← trose left19:13
← Kobaz left19:13
← comawhite left19:13
← NfNitLoop left19:13
← Lgb left19:13
← dwmw2_gone left19:13
← elmargol left19:13
← _rane left19:13
← ndim left19:13
← SRabbelier left19:13
← walf443 left19:13
← Sigma left19:13
← f00li5h left19:13
← mullet_ left19:13
← Thumper_ left19:13
← telmich left19:13
← Tianon left19:13
← crash\ left19:13
← mmattice left19:13
← TheUni left19:13
← codejunky left19:13
← tzaeru left19:13
← SethRobertson left19:13
← madduck left19:13
← lamont left19:13
← sigmonsay left19:13
← yrlnry left19:13
← quentusrex left19:13
← joschi left19:13
← Firass-z0r left19:13
← sitaram left19:13
← hackeron left19:13
← logankoester left19:13
← stuffcorpse left19:13
← bjornbjorn left19:13
← osfameron left19:13
← Odd_Bloke left19:13
← hgb left19:13
← Stx left19:13
← dec left19:13
← gabiruh left19:13
← Paleo left19:13
← tokkee left19:13
← enherit left19:13
← ph^ left19:13
← DiaPo left19:13
← pheaver left19:13
← Schmallon left19:13
← fr0sty-away_ left19:13
← charon left19:13
← mastro left19:13
← joel_falcou left19:13
← _psychic_ left19:13
← dpino left19:13
← gregglind left19:13
← tstclair left19:13
← athit left19:13
← jones_ left19:13
← paulproteus_ left19:13
← PerlJam left19:13
← Will| left19:13
← s0ber left19:13
← eean left19:13
← larsrh left19:13
← stillLotR left19:13
← JEEB left19:13
← przemoc left19:13
← joevano left19:13
← ericholscher left19:13
← jpieper left19:13
← jessfishenden left19:13
← Nat_RH_ left19:13
← dsop left19:13
← d1b left19:13
← dan2 left19:13
← ft left19:13
← sutto left19:13
← UukGoblin left19:13
← injekt left19:13
← jpr5 left19:13
← blast_hardcheese left19:13
← kokx left19:13
← Rhonda left19:13
← dennda left19:13
← MadCoder left19:13
← nothingmuch left19:13
← Stravy left19:13
← crab left19:13
← Rapp left19:13
← ACfromTX left19:13
← mnemoc left19:13
← ptrb left19:13
← cworth left19:13
← foom left19:13
← fr0sty_ left19:13
← mett` left19:13
← steph021_h left19:13
sako cbreak: is it easier to somehow get to 1 commit before Revert "config corrections"?19:13
cbreak: should i be doing a revert for that or reset?19:13
← Dave^| left19:13
remur_030 sako: you guys should really think about switching to a real git approach, this looks like everybody is commiting from his localbranch without rebasing it first19:14
→ Dave^|| joined19:14
→ JM joined19:14
← Layke left19:14
→ steph021 joined19:14
← FilipeMaia left19:14
→ Layke joined19:14
→ Evious joined19:15
→ makeworkeasy joined19:15
sako remur_030: i agree. You mean getting everyone to have their own branch?19:16
→ wonder95 joined19:16
sako and merging master into the branch and then pushing to master right?19:16
clusterfuck19:16
← Bass10 left19:16
← SvenDowideit left19:16
→ mdpatrick joined19:17
→ kelvie_ joined19:17
← shentino left19:17
← mdpatrick__ left19:17
← frogonwheels left19:17
← madewokherd left19:17
wonder95 I have a branch in which I added a new directory, and when I switched to another branch, I got a message 'unable to unlink site/to/directory: Operation not permitted'. What do I need to do to fix it?19:17
it's on my local machine, so I don't think it's a permissions issue19:17
and I don't want that directory in my current branch19:18
→ shentino joined19:18
→ SvenDowideit joined19:18
→ frogonwheels joined19:18
→ Error404NotFound joined19:18
→ mloki joined19:18
→ enherit joined19:18
→ dfr|mac joined19:18
→ goshawk joined19:18
→ alarkin joined19:18
→ SandCube joined19:18
→ robotarmy joined19:18
→ TheMoonMaster joined19:18
→ ISF_ joined19:18
→ oracal joined19:18
→ ph^ joined19:18
→ DiaPo joined19:18
→ TheFox21 joined19:18
→ benishor joined19:18
→ alberto56 joined19:18
→ justin-george joined19:18
→ SimonP86 joined19:18
→ pheaver joined19:18
→ marc_kdab joined19:18
→ skoop joined19:18
→ Schmallon joined19:18
→ bentob0x joined19:18
→ Bombe joined19:18
→ alexmchale joined19:18
→ guardian joined19:18
→ abetaha joined19:18
→ fr0sty-away_ joined19:18
→ charon joined19:18
→ anon33_ joined19:18
→ frakturfreak joined19:18
→ ptl joined19:18
→ Mpenz joined19:18
→ bauruine joined19:18
→ spearce joined19:18
→ MattDiPasquale joined19:18
→ ajpiano joined19:18
→ hobodave joined19:18
→ mkramer joined19:18
→ mastro joined19:18
→ carutsu_ joined19:18
→ AAADMWY joined19:18
→ zombor joined19:18
→ tjholowaychuk joined19:18
→ joel_falcou joined19:18
→ thiago_ny joined19:18
→ _psychic_ joined19:18
→ seme joined19:18
→ jds joined19:18
→ yhager joined19:18
→ ExtraSpice joined19:18
→ scp1 joined19:18
→ sergiomc joined19:18
→ Yuffster joined19:18
→ mw joined19:18
→ booginga joined19:18
→ OOPMan joined19:18
→ Ramblurr joined19:18
→ ruskie joined19:18
→ Cipher-0 joined19:18
→ blee joined19:18
→ kanru joined19:18
→ Maletor joined19:18
→ romeo404_ joined19:18
→ saurik joined19:18
→ alip joined19:18
→ ciaranm joined19:18
→ Chillance joined19:18
→ jaql joined19:18
→ tizbac joined19:18
→ Kyso joined19:18
→ felipe joined19:18
→ Agos joined19:18
→ Pacman_Home` joined19:18
→ Ryback_ joined19:18
→ tomboh joined19:18
→ soidexe joined19:18
→ drizzd_ joined19:18
→ dpino joined19:18
→ gregglind joined19:18
→ mabrand joined19:18
→ dvaske joined19:18
→ MetaCosm joined19:18
→ gbjk joined19:18
→ doener joined19:18
→ bburhans joined19:18
→ leavittx joined19:18
→ sejo joined19:18
→ tomaw joined19:18
→ selckin joined19:18
→ Leonidas joined19:18
→ teuf joined19:18
→ sven^ joined19:18
→ robinbowes joined19:18
→ elly joined19:18
→ Teo` joined19:18
→ tlonim joined19:18
→ sunn` joined19:18
→ piti joined19:18
→ tangledhelix joined19:18
→ blalor_afk joined19:18
→ mrtazz joined19:18
→ metze joined19:18
→ Derander joined19:18
→ avar joined19:18
→ robinsmidsrod joined19:18
→ divVerent joined19:18
→ j0ran joined19:18
→ dean[w] joined19:18
→ halfline joined19:18
→ mcjb joined19:18
→ kandinski joined19:18
→ womble joined19:18
→ funfun joined19:18
→ m0 joined19:18
→ rootard joined19:18
→ ecksun joined19:18
→ comcor- joined19:18
→ Juri_Agarin joined19:18
→ cbz joined19:18
→ stepnem joined19:18
→ Tronic joined19:18
→ multi_io joined19:18
→ lirakis joined19:18
→ Arrowmaster joined19:18
→ fr| joined19:18
→ Swimming_bird joined19:18
→ jksz joined19:18
→ NickyP joined19:18
→ Purdy joined19:18
→ MacGyverNL joined19:18
→ albel727_3 joined19:18
→ jaeckel joined19:18
→ themarmot joined19:18
→ Elfe_ joined19:18
→ jvv joined19:18
→ timj joined19:18
→ Risar joined19:18
zelazny.freenode.net set mode: +v19:18
→ HopsNBarley joined19:18
→ ormaaj joined19:18
→ fr0sty joined19:18
→ muep joined19:18
→ GodEater_ joined19:18
→ Ratler joined19:18
→ Gunni joined19:18
→ hegge joined19:18
→ drry joined19:18
→ Guest26619 joined19:18
→ an0maly joined19:18
→ ayust joined19:18
→ jdav_gone joined19:18
→ squentin joined19:18
→ qknight joined19:18
→ RandomSort joined19:18
→ tro joined19:18
→ Khisanth joined19:18
→ sycofly joined19:18
→ bobhope joined19:18
→ [exa] joined19:18
→ Guest48246 joined19:18
→ jjuran joined19:18
→ quadz joined19:18
→ kelvie joined19:18
→ tmokros joined19:18
→ markatto joined19:18
→ resixian joined19:18
→ blathijs_ joined19:18
→ spb joined19:18
→ esc joined19:18
→ rudi_s joined19:18
→ phishy joined19:18
→ lotyrin joined19:18
→ mjv joined19:18
→ pasky joined19:18
→ yoh joined19:18
→ tewecske joined19:18
→ WAACAYX joined19:18
→ ludde joined19:18
→ frenzz joined19:18
→ redtricycle joined19:18
→ Unknown[OFF] joined19:18
→ Ingmar joined19:18
→ ThomasLocke joined19:18
→ tstclair joined19:18
→ Vampire0 joined19:18
→ athit joined19:18
→ HuntingBears joined19:18
→ AlexanderS joined19:18
→ joelpet joined19:18
→ jones_ joined19:18
→ kenyon joined19:18
→ plopix joined19:18
→ comawhite joined19:18
→ trose joined19:18
→ Kobaz joined19:18
→ elmargol joined19:18
→ NfNitLoop joined19:18
→ Lgb joined19:18
→ dwmw2_gone joined19:18
→ paulproteus_ joined19:18
→ _rane joined19:18
→ ndim joined19:18
→ PerlJam joined19:18
→ Will| joined19:18
→ ericholscher joined19:18
→ SRabbelier joined19:18
→ s0ber joined19:18
→ eean joined19:18
→ walf443 joined19:18
→ larsrh joined19:18
→ stillLotR joined19:18
→ JEEB joined19:18
→ Sigma joined19:18
→ f00li5h joined19:18
→ dan2 joined19:18
→ mullet_ joined19:18
→ przemoc joined19:18
→ joevano joined19:18
→ Thumper_ joined19:18
→ quentusrex joined19:18
→ foom joined19:18
→ telmich joined19:18
→ Tianon joined19:18
→ Rapp joined19:18
→ crash\ joined19:18
→ mmattice joined19:18
→ TheUni joined19:18
→ jpieper joined19:18
→ codejunky joined19:18
→ tzaeru joined19:18
→ tokkee joined19:18
→ Paleo joined19:18
→ gabiruh joined19:18
→ dec joined19:18
→ Stx joined19:18
→ osfameron joined19:18
→ bjornbjorn joined19:18
→ Odd_Bloke joined19:18
→ stuffcorpse joined19:18
→ logankoester joined19:18
→ sitaram joined19:18
→ hgb joined19:18
→ joschi joined19:18
→ yrlnry joined19:18
→ sigmonsay joined19:18
→ Firass-z0r joined19:18
→ lamont joined19:18
→ madduck joined19:18
→ hackeron joined19:18
→ SethRobertson joined19:18
→ jessfishenden joined19:18
→ mett` joined19:18
→ fr0sty_ joined19:18
→ mnemoc joined19:18
→ cworth joined19:18
→ crab joined19:18
→ Stravy joined19:18
→ nothingmuch joined19:18
→ MadCoder joined19:18
→ Rhonda joined19:18
→ kokx joined19:18
→ blast_hardcheese joined19:18
→ jpr5 joined19:18
→ injekt joined19:18
→ dennda joined19:18
→ sutto joined19:18
→ ft joined19:18
→ ACfromTX joined19:18
→ ptrb joined19:18
→ UukGoblin joined19:18
→ d1b joined19:18
→ dsop joined19:18
→ Nat_RH_ joined19:18
zelazny.freenode.net set mode: +vv19:18
remur_030 sako: that'd be a usable approach yes, everybody has his own branch (local is enough) and once he want's to commit his stuff to the git host he should checkout his local master, update this, then switchback to his local branch, git rebase on the local master and then push the master19:18
sako remur_030: for now though, how can i reset the branch to the 6th commit down?19:18
← Error404NotFound left19:18
→ Error404NotFound joined19:18
→ FilipeMaia joined19:18
remur_030 sako: i replied, got a bit off in the log19:18
← Yuffster left19:18
← gplocke left19:18
→ delinquentme joined19:19
← __name__ left19:19
Hutch[S]|away → Hutch[S]19:19
→ Crell joined19:19
→ banisterfiend joined19:19
Crell Hopefully simple question. I have a repo where the working copy has a directory, in which a lot of files have been added, updated, or deleted.19:19
→ tvw joined19:19
→ miloshadzic joined19:20
sako remur_030: i just scrolled up dont see it19:20
Crell I know I can stage all of the add and update files with just git add dirname, but how can I also stage the deletion of those files that are now absent?19:20
Or must I do it individually?19:20
→ ben225 joined19:20
remur_030 sako: make a backup of the git repo to be safe first, then 'git reset --hard commitid19:21
sako ah ok19:21
remur_030 commitid is the id of that 6th down19:21
sako and to backup its just a fs copy right?19:21
copy the folder19:21
← kelvie_ left19:21
remur_030 sako: clone is somewhere safe, you could just push that back in case something goes wrong19:21
clone it*19:21
→ kelvie_ joined19:22
albel727_3 Crell: man git-add -u19:22
← Samual left19:22
jast Crell: the 'git-add' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-add.html [automatic message]19:22
albel727_3 → albel72719:22
← Cipher-0 left19:22
remur_030 you shouldn't work on the host git repo anyway19:22
do two clones, git reset hard one of them and push that one upstream19:22
← elmob left19:23
← codespectator left19:23
← _psychic_ left19:24
← alarkin left19:24
→ tek0 joined19:24
remur_030 then tell people to rebase their stuff on origin/master19:24
Crell albel727: Curious. git add will catch new files but not deleted. add -u catches deleted but not added.19:24
remur_030 and then push their stuff19:25
→ codespectator joined19:25
remur_030 sako: btw the linux kernel even goes abit further with a maintainer beeing the entryguard for a repository, so people can't push but he will pull their changes19:26
pretty useful if you have people constantly misbehaving ;-)19:26
albel727 Crell: there's also -A (--all) option. it does both.19:26
Crell Ahso.19:26
← tuvok302 left19:27
← keystr0k left19:27
Crell That's what I was looking for, thanks.19:27
albel727 yw19:27
→ noahcampbell joined19:27
← eletuchy left19:27
← LouisJB left19:27
→ storrgie joined19:27
→ eletuchy joined19:28
→ ISF__ joined19:28
← ISF_ left19:28
sako remur_030: thanks alot.. so on the push now it complains about fast forwards19:29
→ Samual joined19:29
sako if i do the pull though it will rollback changes i made with the reset right?19:29
ISF → Guest8785219:29
sako can i force it?19:29
← Dave^|| left19:30
remur_030 sako: sure, --force19:31
sako remur_030: im gonna lose the 5 commits right?19:31
← tek0 left19:31
sako will be gone from history?19:31
remur_030 sako: yep, you should be safe with the backup you made ;-)19:31
yes19:31
→ krfs joined19:31
→ Dave^| joined19:31
← Dave^| left19:32
→ ustunozgur_home joined19:32
← mabrand left19:32
→ Dave^| joined19:32
← ISF__ left19:32
← Dave^| left19:32
→ Dave^| joined19:32
← Dave^| left19:33
→ ISF_ joined19:33
← ISF_ left19:33
→ Dave^| joined19:33
← guardian left19:33
→ ISF_ joined19:33
sako remur_030: thanks alot!19:33
you git wizard..19:33
← bentob0x left19:34
← sergiomc left19:34
remur_030 sako: lol, far from it19:34
→ huehnts joined19:34
← krfs left19:34
Hutch[S] → Hutch[S]|away19:35
→ jrmuizel joined19:35
remur_030 the basics just need a bit of getting used to, learning by doing ain't that bad, even though this part of the git history is propably nothing somebody wants to bisect later on in case something broke ;-)19:35
sako hah, ya i personally love to use branches19:36
and to be safe will always pull in master to my branch19:36
i like to isolate conflicts19:36
problem is, we have all these frontend guys working on this and they are all coming from svn backgrounds with some bad habits :/19:36
remur_030 habits are hard to change, having some gatekeeper for the git host is propably a good idea to enforce some standards19:37
sako remur_030: so you were recommending to rebase? so say i do a commit, then i would do git rebase master, then git push?19:38
→ _Unode joined19:38
sako remur_030: i love the idea, however, the gatekeeper is for sure going to be me and I am caught up in 3 other projects ;)19:38
remur_030 sako: I think I read somewhere that is the recommended way to do it and it feels pretty natural, that way you only have small changes to master in your branch and git will just have to push your changes and don't get all confused with history issues19:39
→ msieradzki joined19:39
remur_030 and beeing the gatekeeper doesn't really require much work, if something looks like bad style just reject, once people got accustomed you can relax the scheme again19:41
← linusoleander left19:41
← CaptainKnots left19:41
← ashleyw left19:42
remur_030 in svn you were also expected to do constant svn updates to keep in touch with your upstream repo, git rebase is quite similar19:42
→ gebi joined19:43
← ajpiano left19:43
← marcob left19:44
← ChUrP left19:44
→ ChUrP joined19:45
← remur_030 left19:45
← Tommy[D] left19:45
→ nevyn^1 joined19:45
← tvw left19:45
← pheaver left19:46
→ cubuanic joined19:46
← vano left19:47
→ tvw joined19:48
← bkosborne left19:48
← mloki left19:49
← tstclair left19:49
→ Eridius joined19:49
→ bkosborne joined19:50
→ tstclair joined19:50
→ ProLoser joined19:50
→ cilly joined19:50
← goshawk left19:51
→ pheaver joined19:51
← msieradzki left19:51
→ aah joined19:52
← stamina left19:52
← bhosie left19:52
→ SimonP86_ joined19:53
Hutch[S]|away → Hutch[S]19:53
← wagle left19:54
← Samual left19:54
← dfr|mac left19:55
→ SimonP86__ joined19:55
← miloshadzic left19:55
← SimonP86 left19:56
→ bhosie joined19:56
← ISF_ left19:57
→ stiv2k joined19:57
stiv2k hi19:57
→ wagle joined19:57
← SimonP86_ left19:58
stiv2k I'm trying to update a members key on gitosis and I can't push the changes to gitosis-admin repo19:58
it keeps saying ERROR:gitosis.serve.main:Repository read access denied19:58
fatal: The remote end hung up unexpectedly19:58
→ dfr|mac joined19:58
stiv2k what does it mean? how to fix it?19:59
→ SIGBUS joined19:59
→ Tommy[D] joined19:59
→ ajpiano joined20:00
→ LouisJB joined20:00
← dpino left20:00
→ msieradzki joined20:01
← aah left20:02
→ swalberg joined20:02
ChanServ set mode: +v20:02
→ aah joined20:03
← ChUrP left20:03
→ clarkfischer joined20:03
← _Unode left20:03
← Dave^| left20:03
→ ChUrP joined20:04
→ _psychic_ joined20:04
→ remur_030 joined20:05
→ Dave^| joined20:05
stiv2k ?20:06
← rgr_ left20:07
doener what it says, you don't have read access to the repo20:07
→ yngress joined20:07
← Textmode left20:07
stiv2k doener: sure i do, i just set it up yesterday20:07
← clarkfischer left20:07
stiv2k doener: all im doing it updating the key of a different user20:07
← banisterfiend left20:08
← bkosborne left20:08
→ rgr joined20:08
← bhosie left20:08
→ vitor-br joined20:08
← joshsmith left20:09
Hutch[S] → Hutch[S]|away20:09
Kyso → KyleXY[2]20:10
swalberg I cloned a github repo, made some 3 commits locally, and pushed. Now I have to rebase the 3 commits into 1. I did git rebase -i HEAD~3 and a push. How do I fix the github repo? It shows 3 commits, including the "right one"20:10
← rgr left20:10
stiv2k doener: how can i override or double check this20:10
← spaceonline left20:10
ProLoser anyone use msysgit?20:11
Eridius swalberg: why did you push without verifying your local history?20:11
swalberg: also, `git rebase HEAD~3` isn't going to do anything at all unless there's a merge within the last 2 commits20:11
← Crell left20:11
swalberg Eridius: The short answer is "I don't know what I'm doing"20:11
:)20:11
→ Axsuul joined20:12
→ Bass2 joined20:12
delinquentme ^^ takes balls20:12
swalberg Eridius: OK, it's possible the 3rd commit was sent as is (not merged), it's all just whitespace changes20:12
stiv2k Eridius: can you help me figure out why i'm getting a repo read access denied when i shouldn't be :/ (gitosis-admin)20:13
Axsuul has anyone had trouble accessing a repo behind a corporate firewall?20:13
→ Daniel_WMDE joined20:13
Daniel_WMDE → Duesentrieb20:14
Duesentrieb hi all! i'm just getting me feet wet with git, so I have a few noob questions while i play with it.20:14
i won't be mad if you point me to TFM when appropriate :)20:14
← henrikbjorn left20:15
delinquentme swalberg, did you squash anything in the rebase?20:15
swalberg delinquentme: I picked it20:15
delinquentme Duesentrieb, ask away20:15
Duesentrieb so, first question... what exactly is the difference between pull und checkout?20:15
← JM left20:15
Duesentrieb is it right to say that checkout is to commit as pull is to push?20:15
Axsuul checkout from svn?20:16
Duesentrieb i am coming from svn, so i'm20:16
delinquentme Duesentrieb, a checkout is switching between branches20:16
← wonder95 left20:16
Duesentrieb ...entangled in svn concepts20:16
but i meant git checkout.20:16
delinquentme yeah20:16
→ guardian joined20:16
Axsuul One big difference is that everything is local20:16
remur_030 sako: btw, I suggest checking out gitolite, it's made with corporate needs in mind20:16
delinquentme git checkout yadda20:16
will switch you the the yadda branch ..20:16
Axsuul when you commit something, you are committing to the local repository20:16
Duesentrieb Axsuul: i got that much. very handy20:17
nice for working offline20:17
← eletuchy left20:17
Duesentrieb delinquentme: so, git checkout is kind of like svn switch?20:17
remur_030 you can prohibit people from doing 'git reset --hard' on the git host for example so no one can rewrite history ;-)20:17
delinquentme Duesentrieb, i havnt used svn :D20:17
Duesentrieb heh :)20:17
← jonanin left20:17
delinquentme PULL is a command youd use to update your local version with whats on the server20:18
so you start out with all your branches being up to date .. your local , your buddys local .. and the repo20:18
→ pigdude joined20:18
delinquentme all have the same code20:18
Duesentrieb so it's like "svn up"20:18
delinquentme you BUDDY edits you "foo" branch20:18
← codespectator left20:18
→ fedesilva joined20:18
delinquentme and pushes it up to the repo ( github for many ppl )20:18
→ juangiordana_ joined20:19
→ SeanLazer joined20:19
← guardian left20:19
delinquentme then lets say you want to work on the foo branch20:19
← juangiordana left20:19
→ Escherial joined20:19
delinquentme before doing any edits .. its a good practice to "git pull origin foo"20:19
origin specifies that its up on github20:19
and that will update your local code20:19
Axsuul origin is basically just an alias for a url20:19
delinquentme with your buddys changes20:19
Axsuul, god to know :D20:20
ChanServ set mode: +v20:20
← curtana left20:20
SeanLazer hey guys, is it possible to format a patch from x commit to y commit but only using commits made by one specific user?20:20
Escherial right, although if you want more control over the process, use git fetch followed by git merge if you want the changes20:20
delinquentme SeanLazer, a "patch"?20:20
Duesentrieb delinquentme: right, thanks20:21
here are my two use cases: 1) use git-svn so i can do local commits when working offline, and push out change sets to the svn repo later.20:21
SeanLazer delinquentme: using git format-patch20:21
→ codespectator joined20:21
Duesentrieb 2) sync config files and personal documents between two laptops20:21
Escherial i'm having that classic 'don't want the current state of the settings file to be committed, but do want the old version to be in the repository' issue20:21
Duesentrieb ideally without a central repo20:21
← Buglouse left20:21
Escherial what's the latest opinion on how to deal with that? :\20:21
Axsuul Duesentrieb: git can also be a central repo20:21
stiv2k Hello yesterday I set up gitsosis it worked fine now I'm trying to check out the gitosis-admin repo and update some keys and it keeps giving me ERROR:gitosis.serve.main:Repository read access denied I can't figure out why20:21
swalberg If I'm reading correctly, I think I want to git checkout master; git rebase the_branch which will give me a single commit on master, which I can push to github and send a pull request with that?20:21
stiv2k anyone can help me?20:21
Duesentrieb Axsuul: i know, but i don't want that for the second use case.20:21
SeanLazer delinquentme: i need to move some changes from one repo to another very similar repo, but two guys were making commits to the master branch in tandem and i only want one of the guy's changes20:21
FauxFaux Escherial: man git-update-index --assumed-unchanged20:21
jast Escherial: the 'git-update-index' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-update-index.html [automatic message]20:21
Escherial i've heard people say bad things about --assume-unchanged20:22
Duesentrieb Axsuul: for the first use case, i don't control the repo. it's wikimedia's place for developing mediawiki :)20:22
SeanLazer obviously they should have branched separately while they did their changes but what's done is done and this is what i'm stuck with20:22
swalberg stiv2k: did you try loglevel = DEBUG in gitosis.conf?20:22
Escherial not sure what specifically, but they argue that you should only use that if the file is indeed unchanged20:22
swalberg stiv2k: I had a similar problem, and it turned out I was using commas to separate users instead of spaces20:22
stiv2k swalberg: i cant change my gitosis.conf anymore, it keeps giving me the same error20:22
swalberg: i cant do anything to the gitosis-admin repo, totally locked out20:23
Escherial i think there was some other workaround that involved using a server-side hook to reject commits of the settings file specifically, but i can't remember how it works20:23
swalberg stiv2k: can you edit it directly inside the repo?20:23
stiv2k swalberg: i dont know?20:23
→ EricInBNE joined20:23
← mastro left20:23
swalberg stiv2k: http://jeetworks.org/node/18 looks promising20:24
← TheFox21 left20:25
delinquentme SeanLazer, im not sure how to use that command .. the ONLY thing i could begin to suggest is removing all the commits from the one guy with "git rebase -i "20:25
→ bkosborne joined20:25
delinquentme but that might not be the best solution20:25
← anon33_ left20:25
← Fullmoon left20:25
→ clarkfischer joined20:25
SeanLazer delinquentme: yeah probably not... i think i'm just going to manually pick out all of the commits i want and format individual patches from each of them, there's only about 10020:25
stiv2k le20:25
swalberg: thats impossible20:25
Escherial er, let's assume i'm going to use --assume-unchanged...is that branch-specific, or does that apply to all branches? and if i've already changed the file, will that remove it from my "changed but not updated" list?20:26
stiv2k swalberg: i edited someone elses key not my own, so i should still have access20:26
i still saw it in the keydir20:26
swalberg stiv2k: the debug line will tell you why the user is not getting matched20:26
stiv2k swalberg: shit20:27
→ mrts joined20:28
→ danlucraft joined20:29
→ khmarbaise joined20:29
mrts is there a less obscure way to create a remote branch than `git push origin origin/base-branch:refs/heads/new-branch`?20:29
Escherial didn't realize that was even possible, heh...thank you for letting me know :)20:29
mrts my pleasure :)20:30
→ ablemike joined20:30
ablemike having trouble with git log20:30
→ CaptainKnots joined20:30
ablemike i am trying to do | subject | author | size of commit | date |20:31
→ rgr joined20:32
→ joel_falcou_ joined20:32
→ psoo joined20:32
← rgr left20:32
mrts and on another but similar issue: if you have a local branch and decide to share it later, you do push it to origin. Can you automagically change that branch to track the remote branch from now on?20:32
wereHamster Escherial: the assume-unchanged bit is stored in the index, and the index doesn't belong to any branch20:32
← joel_falcou left20:32
joel_falcou_ → joel_falcou20:32
→ bhosie joined20:33
Escherial wereHamster: logical, thank you. even after reading pro git, these concepts are still hazy to me.20:33
→ miloshadzic joined20:33
mrts i.e. achieve an effect similar to: git push origin the-branch; git branch -d the-branch; git checkout --track origin/the-branch ?20:33
→ rgr joined20:33
wereHamster mrts: git push --set-upstream origin the-branch20:33
or branch --set-upstream20:33
mrts wereHamster, thanks a lot (bow)20:33
← clarkfischer left20:34
swalberg OK, I rebased the 3 branch commits onto master, rebase -i, pick/squash/squash, and it all seems to work fine. Pushed to github and I can send a pull request for one commit. Do I torch the remote branch? I remember reading that rebasing after pushing is evil.20:34
← rgr left20:35
← bhosie left20:35
→ bhosie joined20:35
← zenazn left20:35
← mrts left20:35
→ rgr joined20:35
→ v0n joined20:36
← bhosie left20:36
→ Samual joined20:36
Escherial i must wonder if the essential problem with git is that people aren't willing to put in the time to learn it20:36
→ bhosie joined20:37
Escherial er, more specifically, my essential problem with it20:37
← codetonowhere left20:37
swalberg Escherial: With most projects you can learn as you go. Everything about git out there seems to be about trying to explain graph theory. git is a lot different than everything else out there, but I've found that all the explanations assume that if you were to only learn this graph stuff, all will fall into place20:38
Escherial: I've liked git so far, but man... I lost several commits switching over from svn and I'm never sure if I'm doing things right.20:39
← FilipeMaia left20:39
Escherial +swalberg: oh, and as for rebasing after pushing being evil, you may have seen that at the bottom of this chapter: http://progit.org/book/ch3-6.html20:39
heh, i know exactly what you mean20:39
→ vitorpc joined20:39
→ pantsman joined20:39
Escherial i agree...the metaphor is deceptively simple, but there are a lot of gotchas. the way it's put together reminds me vaguely of perl20:39
swalberg Escherial: Yea, it's also in huge type on http://learn.github.com/p/rebasing.html... But I don't know if I just did something bad or not until it's too late.20:40
Escherial i wonder how people generate these pretty diagrams20:40
swalberg dot?20:41
er, graphviz?20:41
Escherial ah, yeah, that makes sense20:41
← rgr left20:41
← vitor-br left20:41
Escherial i've used it before, but i wouldn't know how to go about getting everything to line up like they do20:41
i kind of just throw a soup of nodes at it and pray that it comes out looking good20:42
→ rgr joined20:42
Escherial rarely disappointed :)20:42
swalberg Escherial: That's the limit of my experience. "Why's that node called GaGGrnm? - Because spelled out, dot made the graph look terrible"20:43
wereHamster Escherial: or keynote20:43
Escherial +swalberg: heh20:43
→ Buglouse joined20:43
ablemike Escherial: one would definitely get that opinion about any technology in the respective support channel20:44
Escherial very true20:45
→ kipibenkipod joined20:48
→ codetonowhere joined20:48
← Zephyrus left20:49
← rgr left20:49
→ rgr joined20:50
sako hey guys, so how would you patch from a git diff? is it as simple as git diff commitid > file.patch20:51
← rgr left20:51
sako how do you patch after that? is there any special flags or anything?20:51
Escherial naive sanity check: say i'm in a branch and i accidentally change a file that i'd really prefer to be on a new branch...should i commit everything but that file, stash that file, then use git stash branch to create a new branch with that changed file in it? or is this ridiculous?20:51
teuf sako: patch -p1 <file.patch20:51
sako teuf: thanks!20:52
← cilly left20:52
→ rgr joined20:52
sako teuf: to make the patch i just use git diff right?20:52
teuf Escherial: git checkout -b newbranch; <commit your file>; git checkout oldbranch20:52
sako: git show20:52
→ FilipeMaia joined20:52
→ Mage joined20:53
← bhosie left20:53
← rgr left20:53
→ versatilemind joined20:53
Escherial can i do that if i have uncomitted changes in my current branch?20:53
→ s0ber_ joined20:53
← ProLoser left20:53
wereHamster sure20:54
→ bhosie joined20:54
Escherial i was under the impression that you either had to commit everything or stash if you wanted to switch branches20:54
→ rgr joined20:54
wereHamster that's true only under certain circumstances20:54
teuf Escherial: at worst, git will refuse to switch branch20:55
← s0ber left20:55
← cubuanic left20:55
teuf so you don't have anything to lose by trying any time you want :)20:55
s0ber_ → s0ber20:55
teuf and to fallback to this stash dance if it doesnet work20:55
Escherial true...i'm always a little afraid that i'll do something irrevocable, though, heh...20:56
← rgr left20:56
teuf Escherial: this is https://git.wiki.kernel.org/index.php/GitFaq#How_do_I_remove_my_uncommitted_changes_from_branch_A_and_add_them_to_a_new_branch_B.3F fwiw20:56
sako teuf: i get patch: **** Only garbage was found in the patch input.20:57
→ brandFU joined20:57
Escherial well, that's useful :) thank you all20:57
→ p3rg2z joined20:58
teuf sako: git show HEAD >toto.diff; git checkout HEAD^ ; patch -p1 <toto.diff just worked here20:58
→ ashleyw joined20:59
Escherial unrelated, but i do love it when people inexplicably add tables that reference mine, but neglect to foreign key them, leaving a ton of junk data in their tables that causes exceptions when they try to follow references that no longer exist :\20:59
→ borwar joined21:00
→ tunedal joined21:00
→ kylehayes_ joined21:00
← bkosborne left21:01
← ashleyw left21:01
wereHamster teuf: why would you do that?21:01
→ sergio joined21:02
sako teuf: it was some weird fast forwards issue.. thanks for your help121:02
!21:02
← kylehayes left21:02
→ Antisoche_ joined21:02
teuf wereHamster: just for a quick confirmation that patch accepts the output of git show21:02
← Tronic left21:03
→ Tronic joined21:03
sergio In "master" I have a folder called "folder", when I checkout branch "branch", I still see that folder. How can I make git-checkout delete it ?21:03
← Kobaz left21:03
→ Kobaz joined21:03
wereHamster sergio: git checkout branch; git rm -rf folder; git commit -m 'remove folder'21:03
and then make sure there are no untracked files in taht folder21:04
← v0n left21:04
sergio wereHamster, The folder doesn't exist in branch "branch", it's already untracked21:04
← Guest48246 left21:04
wereHamster sergio: are there any untracked files in it?21:04
← kylehayes_ left21:05
→ Godslastering joined21:06
sergio wereHamster, yes, it has untracked files in it ( files that belong to master )21:06
Godslastering → Guest5500121:06
wereHamster sergio: fi you checkout master, are there any untracked files?21:06
sergio wereHamster, ah, i see what you mean.21:07
← pigdude left21:07
→ mmw joined21:07
sergio wereHamster, git will refuse to delete any untracked file while changing branch21:07
thanks21:07
→ v0n joined21:07
← swalberg left21:08
→ alex__c2022 joined21:09
← FilipeMaia left21:09
→ clarkfischer joined21:09
← vitorpc left21:10
→ drubies joined21:10
← sergio left21:10
→ goshawk joined21:10
→ ashleyw joined21:13
brandFU → brandFU|away21:14
→ vitor-br joined21:14
← delinquentme left21:15
← kipibenkipod left21:15
→ kpreid joined21:15
← brandFU|away left21:15
→ ereslibre joined21:15
→ spaceonline joined21:16
→ brandFU joined21:16
→ timj_ joined21:18
← timj left21:18
ereslibre arg ! I think git just ate an important document. workflow => git add file.txt; "oh I added on the wrong branch !"; git reset --hard; git checkout goodBranch21:19
→ carpii joined21:19
ereslibre now goodBranch has a folder "lulz" that in the old branch didn't exist and where this document was living. It seems my document dissappeared21:19
and extundelete cannot retrieve it :'(21:19
carpii is there a way to stage a subset of lines from command line? Ive only ever seen it done in graphical tools21:19
teuf carpii: git add -p21:20
(or git add -i which I find more complicated)21:20
carpii oh wonderful21:20
thanks :)21:20
Escherial idling in this channel is like taking a class in git :) useful tip there21:21
→ _ollie joined21:21
carpii yeah, i learnt more from this chan than reading many of the terse git docs :)21:21
→ darkredandyellow joined21:23
teuf carpii: git-reset and git-checkout also takes a -p though I never used these :)21:23
→ tek0 joined21:23
NfNitLoop ereslibre: That's what 'git reset --hard' does.21:24
you probably meant to 'git reset'.21:24
→ eletuchy joined21:24
NfNitLoop you specifically gave git an option to delete the file.21:24
carpii teuf, you can reset a partial diff ?21:24
→ Zer joined21:25
→ _Unode joined21:25
stiv2k is it possible to edit the git repository directly? without checking it out21:25
teuf carpii: something like that, as I said I never used it, so I'm not exactly sure what this does :)21:25
ereslibre NfNitLoop, yeah... git reset --soft. okay, this is really wrong. it wasn't even committed, I think I have to realize I just lost it21:25
NfNitLoop, thx anyway21:25
carpii ok thanks, i dont think ive never needed that either, but good to know :)21:25
everything in git fits together so well, linux is just too damn smart21:25
linus21:26
ereslibre carpii, unless you run because of being late a git reset --hard and your party goes on !21:26
carpii hm, was that meant for ereslibre maybe? :)21:27
Escherial heh, he meant it for himself? :321:28
ereslibre carpii, hehe no, ereslibre is me xD21:28
carpii haha21:28
by jove, youre right!21:28
← ThomasLocke left21:28
Escherial i think it was a mutation of "everything is fantastic until catastrophe strikes"21:28
NfNitLoop But who's on first?21:28
→ hasrb joined21:28
ereslibre carpii, it was just a 'funny' thing because of your statement "everything in git fits together so well, linux is just too damn smart"21:28
carpii ah :)21:28
ereslibre Escherial, that is...21:28
carpii, i just lost a damn important document.21:29
carpii ouch :(21:29
ereslibre and yeah, it was my fault, not git's...21:29
carpii often the way :D21:29
Escherial git is like an affectionate pet that gets its tail stepped on a lot :\ eventually you end up running it over with the sedan as it attempts to follow you grocery shopping21:30
→ rgr joined21:30
Escherial many tears are shed, but it's really nobody's fault but your own21:30
carpii svn is more like the rabid dog you swerve in order to hit, then reverse over it to make sure21:31
← TrineLocke left21:31
Escherial haha21:31
→ cptG joined21:31
Escherial by this standard, source safe is the pit bull to which you are chained21:31
carpii hehe21:32
→ ISF_ joined21:32
→ Rickardo1 joined21:32
doener ereslibre: what kind of document?21:34
ereslibre: since you ran "git add" on it, chances are that you'll be able to restore it21:34
ereslibre doener, odt. I know this kind of documents should not be added to git, but it is an exception21:34
doener, really ?21:34
doener ereslibre: "git add" creates the blob object, so the index can connect the path to that blob21:35
ereslibre: and that blob should still be around21:35
← skoop left21:35
tek0 can gitolite be configured to accept all repos pushed by a user and automatically add them to the config?21:35
ereslibre doener, i am all ears. how could I find that blob ?21:35
doener ereslibre: first step: git fsck --lost-found21:35
← rgr left21:35
← Tianon left21:35
ereslibre doener, some blobs yeah21:35
stiv2k hi, i somehow locked myself out of the gitosis-admin repository, it keeps giving this error: ERROR:gitosis.serve.main:Repository read access denied ... any ideas how to fix this?21:35
ereslibre 8 to be exact21:35
→ banisterfiend joined21:36
→ tal2 joined21:36
carpii ssh into server ?21:36
← dr_win_ left21:36
ereslibre doener, are they ordered by date or something ?21:36
→ SirFunk_ joined21:36
stiv2k carpii: i am sshed... but i cant check out the gitosis-admin repo21:36
doener ereslibre: no. wait a sec21:36
ereslibre doener, yeah np.21:36
carpii cant you find the 'raw' config file ?21:36
on the server21:36
stiv2k carpii: im not sure, where would it be21:37
carpii did you create a dedicated git user account for gitosis ?21:37
Hutch[S]|away → Hutch[S]21:37
Zer Has anyone here worked with the JGit library?21:37
stiv2k carpii: i did not, but ubuntu did21:37
← danlucraft left21:37
carpii is there a /home/git ?21:37
or a /home/gitosis21:37
Zer I'm looking for the equivalent of the git rm --cached command and not finding it21:38
← nadavoid left21:38
stiv2k carpii: i think its placed in /srv/gitosis21:38
carpii oki21:38
there should be a 'respoitories' directory there i think21:38
repositories21:38
and in your gitosis dir, you should find gitosis.conf21:38
stiv2k looks21:38
stiv2k cool21:39
now how do i enable debug logging21:39
so i can see why it keeps locking me out21:39
← kelvie_ left21:39
→ rgr joined21:40
→ Tianon joined21:40
← Tianon left21:40
→ Tianon joined21:40
ereslibre doener, I LOVE YOU !21:40
doener, git show SHA1-of-blob until i found a binary one21:41
doener, git show SHA1 > $HOME/lol.odt21:41
← alberto56 left21:41
ereslibre there it is !21:41
doener, THANK YOU ! I lost all my faith !21:41
← rgr left21:41
doener ereslibre: ah, ok, for 8 blobs that's even doable, I wrote a small for loop in the meantime to write the blobs into some directory21:41
ereslibre: you're welcome!21:41
ereslibre doener, really, thank you !21:41
*phew* that was close !21:41
← brandFU left21:42
ereslibre carpii, see ? git is f**** awesome !21:42
xD21:42
Escherial indeed, git (and doener) inexplicably save the day :D21:42
→ rgr joined21:42
stiv2k carpii: where does the log get written to?21:42
Escherial lessons about --hard are learned and life continues21:42
ereslibre literally. a the-whole-day document...21:42
carpii which log?21:42
→ kelvie_ joined21:42
→ Chaser joined21:42
→ boombatower joined21:43
← p3rg2z left21:43
← rgr left21:43
← versatilemind left21:43
→ ProLoser joined21:44
stiv2k carpii: nevermind21:44
it output to the cli21:44
← WebDragon|away left21:44
← Zer left21:44
← _psychic_ left21:44
→ rgr joined21:44
stiv2k ok21:44
here is my problem: http://fpaste.org/c4e7/21:44
carpii edit the raw config on server, to give you write access21:45
then abandon your config changes on local copy21:45
stiv2k carpii: i should already have it though21:45
DEBUG:gitosis.access.haveAccess:Access check for 'stiv2k' as 'writeable' on 'gitosis-admin.repo'...21:45
doener stiv2k: is that .repo suffix correct?21:45
stiv2k: looks weird to me21:45
→ versatilemind joined21:46
remur_030 stiv2k: you just setting gitosis up?21:46
stiv2k uh21:46
remur_030: i set it up yesterday, today i have problems21:46
← _ollie left21:46
stiv2k doener: lets see21:46
wereHamster stiv2k: !gitosis21:46
jast stiv2k: gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite [automatic message]21:46
remur_030 stiv2k: then you propably have nothing important in there yet, gitosis is unmaintained but has a nice replacement, gitolite21:46
stiv2k doener: /srv/gitosis/repositories/gitosis-admin.git21:47
remur_030 it has a good doku and is maintained21:47
stiv2k wereHamster: really?21:47
doener stiv2k: yeah, no .repo there21:47
stiv2k ok21:47
ill try gitolite21:47
i didnt even know gitosis was old21:47
doener stiv2k: git clone gitosis@localhost:gitosis-admin -- should work, but as the others said, switch to gl21:47
stiv2k will do21:47
remur_030 stiv2k: doener is right, always use the git user, not your personal user21:48
stiv2k remur_030: i do21:48
remur_030 it's one of the usual mistakes one starts out with21:48
← dfr|mac left21:49
stiv2k doener: i took off the .repo and its still doing it21:49
:/21:49
http://fpaste.org/VGCO/21:50
← jaql left21:52
← doener left21:53
← ashleyw left21:55
carpii git clone gitosis@localhost:gitosis-admin.git21:56
← tewecske left21:57
→ tewecske joined21:57
→ montylounge_ joined21:57
← Pewpewarrows left21:59
← frenzz left21:59
← spaceonline left22:00
← montylounge left22:00
montylounge_ → montylounge22:00
→ Smirnov joined22:00
← versatilemind left22:00
Smirnov if i want to squash multiple commits in a branch into 1 commit before merging to master, is there an easier way than just using git rebase + squash before doing a merge?22:00
← mahata left22:00
← jmil left22:00
← alester left22:01
carpii you could git reset back to the point of the first commit, in your branch22:01
then git add with a single commit22:01
← SeanLazer left22:01
→ guardian joined22:01
carpii then merge into master22:01
← TheMoonMaster left22:01
carpii not sure if thats any easier though, really22:01
← guardian left22:01
→ SeanLazer joined22:02
hagebake → hagabaka22:02
stiv2k carpii: ill just start over with gitolite22:02
:)22:02
carpii sti, did you try adding .git onto the end of the git clone ?22:03
stiv2k carpii: i got that msg from you after i removed gitosis22:03
carpii ah22:03
too bad :)22:03
im pretty sure that was the problem22:03
Escherial anyone happen to know of a nice tool for visualizing history across multiple branches? something like those diagrams in the pro git book22:03
carpii what os ?22:03
→ frenzz joined22:03
Escherial windows preferably :\22:03
stiv2k carpii: ubuntu server edition22:03
→ benishor_ joined22:03
stiv2k oh22:03
Escherial oh...heh.22:03
carpii sorry stiv2k, i meant esch :)22:04
esch, you could try smartgit22:04
but the nicest ive seen is gitx, but thats mac only22:04
Escherial smartgit looks interesting, thanks :D ah, and yeah, gitx does look neat based on the screencasts...22:05
← Chaser left22:05
→ Chaser joined22:06
← SeanLazer left22:06
→ FilipeMaia joined22:06
← FilipeMaia left22:06
→ FilipeMaia joined22:06
→ jmil joined22:07
← benishor left22:07
← mmw left22:07
← ch077179 left22:08
→ FilipeMaia_ joined22:09
← FilipeMaia_ left22:10
→ aeurielesn joined22:10
← FilipeMaia left22:10
← montylounge left22:10
→ FilipeMaia joined22:10
← funsheep left22:10
aeurielesn Hi guys, is there any "standard" filename (as .gitignore) to place in empty dirs which need to be must be committed?22:11
ereslibre aeurielesn, .keep I think22:12
← ymasory_ left22:12
← cptG left22:12
→ tty2 joined22:12
← ISF_ left22:12
← tty1 left22:13
→ ISF joined22:13
→ SeanLazer joined22:13
wereHamster aeurielesn: no22:13
jast aeurielesn: most people use either .gitignore or .keep22:14
→ bkosborne joined22:14
aeurielesn what about .gitkeep? sounds any good?22:15
Hutch[S] → Hutch[S]|away22:17
wereHamster aeurielesn: I'd use .orly or .omgwtflol. But ultimately it's your choice22:17
Escherial kind of a long shot, but has anyone used egit (eclipse plugin for git) to clone a project from github?22:18
running into some issues, probably because whoever maintains the project doesn't use eclipse themselves (thus the project doesn't have the eclipse metadata describing it as a project)22:18
aeurielesn wereHamster: lol22:19
wereHamster aeurielesn: no, .lol, so that it's hidden22:19
aeurielesn thanks ereslibre, wereHamster, jast22:22
← aeurielesn left22:22
← _Unode left22:22
← SimonP86__ left22:24
← ablemike left22:24
→ rpbertp13 joined22:25
→ kuhrt joined22:26
→ pimeja joined22:27
← bkosborne left22:29
→ p3rg2z joined22:29
→ krfs joined22:29
← krfs left22:30
← SeanLazer left22:30
← kelvie_ left22:30
← Mage left22:30
← pantsman left22:30
← Samual left22:30
← LouisJB left22:30
← frogonwheels left22:30
← enherit left22:30
← ph^ left22:30
← DiaPo left22:30
← Axsuul left22:30
← Schmallon left22:30
← fr0sty-away_ left22:30
← charon left22:30
← gregglind left22:30
← athit left22:30
← jones_ left22:30
← paulproteus_ left22:30
← PerlJam left22:30
← Will| left22:30
← eean left22:30
← larsrh left22:30
← stillLotR left22:30
← JEEB left22:30
← przemoc left22:30
← joevano left22:30
← ericholscher left22:30
← jpieper left22:30
← jessfishenden left22:30
← Nat_RH_ left22:30
← dsop left22:30
← d1b left22:30
← dan2 left22:30
← ft left22:30
← sutto left22:30
← UukGoblin left22:30
← injekt left22:30
← jpr5 left22:30
← blast_hardcheese left22:30
← kokx left22:30
← Rhonda left22:30
← dennda left22:30
← MadCoder left22:30
← nothingmuch left22:30
← Stravy left22:30
← crab left22:30
← Rapp left22:30
← ACfromTX left22:30
← mnemoc left22:30
← ptrb left22:30
← cworth left22:30
← foom left22:30
← fr0sty_ left22:30
← mett` left22:30
ustunozgur_home → ustunozgur22:30
→ pfrog joined22:31
→ _Unode joined22:31
→ Cro_Crx joined22:31
pfrog is there a nice way of extracting a few files and their history from a repository in to a new repository? maybe a filter or something?22:31
wereHamster pfrog: filter-branch22:32
→ angelsl joined22:32
→ bmahe joined22:32
pfrog cool, yeah thats what I was looking for thanks22:32
ustunozgur → 52AAADOAJ22:32
→ SeanLazer joined22:33
→ kelvie_ joined22:33
→ Mage joined22:33
→ pantsman joined22:33
→ Samual joined22:33
→ Axsuul joined22:33
→ LouisJB joined22:33
→ frogonwheels joined22:33
→ enherit joined22:33
→ ph^ joined22:33
→ DiaPo joined22:33
→ Schmallon joined22:33
→ fr0sty-away_ joined22:33
→ charon joined22:33
→ gregglind joined22:33
→ athit joined22:33
→ jones_ joined22:33
→ paulproteus_ joined22:33
→ PerlJam joined22:33
→ Will| joined22:33
→ ericholscher joined22:33
→ eean joined22:33
→ larsrh joined22:33
→ stillLotR joined22:33
→ JEEB joined22:33
→ dan2 joined22:33
→ przemoc joined22:33
→ joevano joined22:33
→ foom joined22:33
→ Rapp joined22:33
→ jpieper joined22:33
→ jessfishenden joined22:33
→ Nat_RH_ joined22:33
→ dsop joined22:33
→ d1b joined22:33
→ UukGoblin joined22:33
→ ptrb joined22:33
→ ACfromTX joined22:33
→ ft joined22:33
→ sutto joined22:33
→ dennda joined22:33
→ injekt joined22:33
→ jpr5 joined22:33
→ blast_hardcheese joined22:33
→ kokx joined22:33
→ Rhonda joined22:33
→ MadCoder joined22:33
zelazny.freenode.net set mode: +v22:33
→ nothingmuch joined22:33
→ Stravy joined22:33
→ crab joined22:33
→ cworth joined22:33
→ mnemoc joined22:33
→ fr0sty_ joined22:33
→ mett` joined22:33
zelazny.freenode.net set mode: +v22:33
→ unhygienix joined22:33
← rgr left22:33
← ExtraSpice left22:33
← unhygienix left22:33
→ rgr joined22:35
→ PigeonFriend joined22:35
← ereslibre left22:36
← fr0sty_ left22:36
← Cro_Crx left22:36
← LouisJB left22:36
→ airborn joined22:36
← rgr left22:36
← kelvie_ left22:36
→ kelvie_ joined22:36
→ rgr joined22:37
← fedesilva left22:37
← Maletor left22:37
← rpbertp13 left22:38
← booginga left22:38
→ agile joined22:38
← khmarbaise left22:39
PigeonFriend hey guys. Im pretty new to git, so i bet this question has been answered *loads* of times; sorry if it's a pain. I'm a long term SVN user and use SVN:externals all the time for core code that i want to keep in sync with 3rd party repos (or my own centralised repo for many projects). What is the equivelent for git? If i want to have a copy of someone elses code (eg: jquery) in my project, how would22:40
i do that? I'm looking for it to behave in the same way as svn:externals (in that i can just grab the latest pushed data when i like)22:40
→ fedesilva joined22:41
→ _psychic_ joined22:41
→ SimonP86 joined22:41
carpii youve a few choices i guess, but git doesnt handle this too well22:44
im assuming the third party components are in a subdir >22:44
?22:44
Eridius PigeonFriend: the two standard options are submodules or git-slave22:45
wereHamster submodules, or subtree merge, or gitslave (third-party project) or simply keep a copy of the code22:45
← nevyn^1 left22:45
Eridius in both cases, you need to create a new commit in your repo in order to pull in changes from the other projects22:45
PigeonFriend carpii: sorry, yer, in a sub dir is best22:45
carpii theres a third option, which is to creata a seperate repo in your subdir, which is a repo in itself22:45
← jmil left22:45
carpii then ignore this subdir from your main repo22:45
but it does mean you have to explciitly go in and pull it22:45
← bmahe left22:46
← chrischambers left22:46
PigeonFriend carpii: sounds fairly reasonable22:46
→ TheMoonMaster joined22:46
Escherial i'm off, but thank you all for the help and advice :)22:46
← Escherial left22:46
carpii look into submodules too though, ive never used them, but if youre purely read-only on the code youre pulling, they might work nice22:46
← SeanLazer left22:46
scelis PigeonFriend: I would recommend using submodules. they are more complicated than externals and harder to learn at first, but they have a ton of benefits22:48
PigeonFriend i've been reading gitpro.org and (maybe wrongly) got the impression i could checkout a copy of jquery (or fork it?!) and then i can have it tracked in my project. I could then make changes to the core code of jquery and just keep that going in parallel with changes the core team makes. then (if i feel like it) i can push my changes into jquery. [nb: jquery here is just an example]22:48
→ illustir joined22:48
PigeonFriend scelis: sell submodules to me, what is it? :)22:48
seme hey guys... I had an existing project and created a git repo with it by doing git init, adding everything, committing it and then I copied that directory to a remote ssh location... then I did a git clone onto my machine from that remote ssh location, branched and again committed my work... I'm trying to merge the work from my branch into the master branch but it is saying it won't update a checked out branch22:48
scelis PigeonFriend: Basically, your main repo contains "pointers" to other git repositories22:49
SethRobertson PidgeonFriend: Gitslave might work for you as well.22:49
← FilipeMaia left22:49
seme I'm new so I'm not sure if I'm doing this right22:49
scelis PigeonFriend: these pointers point to individual commits in the other repos22:49
illustir I have no clue, git pull used to work and now it starts complaining I need to tell it which branch to merge with;22:49
scelis PigeonFriend: so you update the pointers periodically to point to new commits as your dependencies change upstream22:49
← cwebber left22:49
→ FilipeMaia joined22:50
seme it also mentions something about a non-bare repo... not sure what that is22:50
→ bmahe joined22:50
scelis PigeonFriend: it is like using svn:externals but always referencing actual revision numbers in those externals22:50
seme thanks for any help you can offer22:50
PigeonFriend scelis: sounds good tbh. how does it compare to my little theory or letting me add my own changes?22:50
SethRobertson If you want to keep yourself at the tip of the remote subproject branches (unless you checkout a specific tag on all repos), then gitslave is better22:50
scelis PigeonFriend: the benefit to this is that when you tag, branch, etc, that always remembers the old commits to those dependencies22:50
PigeonFriend: so you can checkout a tag, and build *EXACTLY* what you had before when you first created that tag. you dont have to remember what versions of each dependency you used22:50
PigeonFriend: yep, you can commit to the submodules. they really are just git repos within your parent repo22:51
← tango_ left22:51
PigeonFriend ahhh, scelis: i like, i must say thats the one thing that frustrates me about svn:externals. you cant link a commit to an exact external revision22:51
← pgib left22:51
scelis PigeonFriend: I think svn added that in 1.622:52
PigeonFriend scelis: perhaps for what i want i should 'fork' (i dont really know what that is yet - sorry) and then merge in the core changes every so often. giving me the freedom to make changes to core if i need to? or am i misunderstanding?22:52
SethRobertson Then you do not want gitslave.22:52
PigeonFriend ah, submodules sounds good then22:52
→ |Bboy| joined22:52
scelis PigeonFriend: here is my workflow, if it is helpful...22:52
PigeonFriend scelis: ok, im running 1.5 because i'm running RHEL on my live server and i don think their repos support 1.6 yet22:52
scelis PigeonFriend: I have a server with bare repositories (both my own projects, as well as dependencies I use as submodules)22:53
← SandCube left22:53
→ tango_ joined22:53
scelis PigeonFriend: these dependencies are "forks" of other repositories (a fork is just a clone of the repo that you decide you want to commit to)22:53
← lvianna left22:53
← joel_falcou left22:54
scelis PigeonFriend: the submodules point to my own URLs for these dependencies, allowing me to easily push changes to them if I need to, or branch / tag what I need22:54
PigeonFriend: every once in a while, I merge in all upstream changes into this repo, and push it to my fork on my server22:54
seme do I have to do the merge from my test-branch to master on the remote host?22:54
PigeonFriend scelis: ok, makes sense. but from the perspective of a complete novice, it does seem slightly complex and daunting.22:55
scelis PigeonFriend: so yeah, this form becomes your repo to do with as you see fit. you can make any changes you want, branch, commit, merge, etc etc22:55
← enherit left22:55
scelis PigeonFriend: git is definitely a complex piece of software. much much much more complex than svn. but it is definitely worth learning. it's amazing22:56
PigeonFriend as i said, i've been reading progit.org and git seems awesome, but i'm slightly suprised there is no easy way to have 'externals'.22:56
← tewecske left22:56
scelis PigeonFriend: thats externals. :) its as easy or complex as you want it to be depending on the workflow you desire22:56
← banisterfiend left22:56
PigeonFriend yer, it does seem excellent. but im always more comfortable with what i know, and i know SVN pretty well. got an awesome set-up at our company22:57
→ tewecske joined22:57
→ SeanLazer joined22:57
scelis PigeonFriend: one helpful piece of advice: when you commit changes to the submodule pointers, do so in their own commits. don't change both source code and a submodule pointer in a single commit. not everyone agrees with me on this, but I find it makes merging, rebasing, and cherry-picking much easier and less prone to conflicts22:57
← SeanLazer left22:57
← ChUrP left22:58
scelis there is no such thing as an awesome svn setup. soon you will see :)22:58
PigeonFriend scelis: ok. perhaps once i get things set up i'll be able to make more use of that advise!22:58
scelis: hehe, yer. i admit there are limitations, but for what i know, its a good set up22:59
scelis PigeonFriend: but yes, progit is an excellent place to start22:59
PigeonFriend scelis: i'll do some thinking and make note of this. perhaps coming back later to get more advice from anyone that's willing to share. thanks for your help. our company isnt moving to git just yet, will be in the coming months22:59
scelis PigeonFriend: good luck!22:59
PigeonFriend its basically because SilverStripe (our CMS/Framework that we dev with) is moving, so we are following. sheep, eh?22:59
← MacGyverNL left23:01
← seme left23:02
→ brandFU joined23:02
→ t0rc joined23:03
PigeonFriend scelis: thanks, hopefully its not too difficult. i already love the fact you can make alias commands with git. im slowly making my git set-up to have all the svn commands! oh god; i'm doing what progit said not to do... it said 'forget everything you know about SVN' oopsey!23:03
← _Unode left23:03
← Guest26619 left23:04
← mw left23:04
scelis PigeonFriend: haha :)23:05
← d0k left23:05
→ MacGyverNL joined23:06
← ajpiano left23:07
← Dave^| left23:07
→ Cro_Crx joined23:07
→ lepine_ joined23:08
← MacGyverNL left23:09
pasky I need to find out easily if any line touched by commit X has been changed in any followup commit - how to go about that?23:10
← Agos left23:10
Eridius blame the relevant lines?23:10
pasky how do I do that?23:11
→ MacGyverNL joined23:11
pasky aside of manually getting all line ranges; the patch is rather large, so it's not too convenient unless i script it :)23:11
← lepine_ left23:12
pasky actually, only single -L parameter may be given so this is not an option at all23:12
→ lepine joined23:12
scelis pasky: you could try to revert commit X and see what conflicts you get...23:13
lepine Is there an easy way to open/see all revisions of a file after <hash> ? no matter which branch the commit is on23:14
← frogonwheels left23:14
Eridius pasky: blame the whole file, then filter to the relevant lines?23:14
→ banisterfiend joined23:14
pasky Eridius: well, again, how would i do that easily?23:14
→ frogonwheels joined23:14
lepine I'm trying to see why a change isn't present in my current branch. I've been merging alot lately; chances are i screwed up somewhere23:14
Eridius pasky: easily? I don't know23:14
pasky scelis: it's likely that i will get many and it will be specific to whole chunks rather than lines, so a lot of followup work23:15
frogonwheels pasky: you look at man git-blame ?23:15
jast pasky: the 'git-blame' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-blame.html [automatic message]23:15
pasky frogonwheels: yes23:15
frogonwheels: anything i missed there?23:15
→ pr0wler joined23:15
scelis pasky: this commit only touched one file?23:15
→ spaceonline joined23:15
← brandFU left23:15
frogonwheels is pondering the question.23:16
scelis pasky: i can't think of anything other than looking at the git-revert conflicts. not unless you want to write a script to help you out23:16
← Duesentrieb left23:17
→ newbiz joined23:17
→ recruit0 joined23:18
frogonwheels pasky: I'm thinking parse the diffs (maybe using a narrower context) and git blame --reverse for all ranges in the hunks23:18
recruit0 Is it okay for a team to edit files in a shared directory and git will keep track appropriately?23:19
← tom_ left23:19
← dreiss left23:19
frogonwheels pasky: as in do a diff on the commit in question, parse all the file/range info in each hunk of the diff and git blame --reverse it23:19
recruit0 Or should we edit stuff in our own directories then merge into master branch?23:19
wereHamster no23:19
→ dreiss joined23:19
wereHamster yes23:19
recruit0: ^^^23:20
← bhosie left23:20
frogonwheels pasky: and possibly use the ermm.. --porcelain option, or whatever just reports the things as it finds them.23:20
pasky hmm23:20
← ludde left23:20
recruit0 I initialized git on our shared folder "orange"23:20
pasky I suppsoe I could run git blame for each range in turn, yeah23:20
recruit0 to keep track of everything...23:20
→ ablemike joined23:20
pasky and --reverse is a nice trick!23:21
frogonwheels: thanks23:21
→ Mannequin joined23:21
← ablemike left23:21
← tunedal left23:21
frogonwheels pasky: assuming you can't put multiple -L ?23:21
→ _ollie1 joined23:21
recruit0 So I should lock down our master git repo?23:21
So only 1 person (me) can commit to it...23:22
frogonwheels recruit0: the master git repo should be bare (no working directory)23:22
recruit0 -.- I'm still confused23:22
→ warlock_mza joined23:23
→ bhosie joined23:23
frogonwheels recruit0: and you don't _need_ to lock it down, people push up to it, and since you can't push a non fast-forward commit, that gives you your point of 'serialisation' of commits23:23
← _ollie1 left23:23
← |Bboy| left23:23
recruit0 Multiple people are editing directories though23:24
example: we have a public_html directory where we have multiple webmasters writing PHP scripts for our website23:24
← MattDiPasquale left23:25
→ tunedal joined23:25
carpii recruit, youre thinking in terms of svn i think23:25
git is a lot more flexible23:25
frogonwheels recruit0: yeah.. so they make the change locally, and commit it to their local repo, then push it up to the master23:25
← rgr left23:25
← tango_ left23:25
← ProLoser left23:26
← stiv2k left23:26
recruit0 carpii: I know it's more flexible, thats why I/we chose it but I'm trying to understand how to use it and stuff23:26
frogonwheels recruit0: if somebody else has made a change, the push ill be rejected and they will have to fetch and merge/rebase their changes onto the other person's (ie pull) change23:26
← tvw left23:26
→ rgr joined23:26
carpii i know, it wasnt a criticism :)23:26
frogonwheels s/ill/will/23:26
→ tango_ joined23:27
→ bkosborne joined23:27
← LiamH left23:27
recruit0 Oh I think I understand...23:27
← bkosborne left23:27
← HG` left23:27
carpii but git is pretty good at stopping people screwing stuff up23:27
so if 5 people change the same file and push it to master23:27
← fedesilva left23:27
carpii itll either incorporate all 5 changes, or bork when one user tries to push, and let him know why23:27
recruit0 See I understand the repo control part but its from the file system viewpoint that is confusing me23:28
carpii got an example?23:28
recruit0 Ye, typing right now lol23:28
→ Zer joined23:28
carpii hehe23:28
Zer Has anyone here used JGit?23:28
← matjas left23:28
Zer I'm wondering if anyone knows what the equivalent command is in it for git rm --cached23:29
← v0n left23:29
→ matjas joined23:29
← matjas left23:29
← tizbac left23:30
recruit0 So, if Person C pulls, then Person A edits blah.php, saves, then Person B edits blah.php, saves, then Person C pushes, then C's push will fail?23:30
← pantsman left23:30
carpii its not about saving23:30
youre all working on a local copy, right ?23:31
because you checked the git repo out23:31
Zer I believe he'll have to pull first23:31
recruit0 Saving = changing file tho...23:31
carpii ok, maybe by save you meant commit?23:31
recruit0 meaning C's repo will be wrong23:31
no they haven't committed yet23:31
← alexmchale left23:31
recruit0 A/B23:31
carpii theyve staged their changes (ie git add) ?23:31
frogonwheels recruit0: you commit locally, then push the change23:31
← goshawk left23:31
→ pheaver_ joined23:31
recruit0 No I mean they're messing with the files on the server directly23:32
carpii yeah they shouldnt23:32
→ ckrailo joined23:32
carpii else why use git23:32
→ alexmchale joined23:32
recruit0 so then I should lock that down so they cant break it23:32
or something...23:32
carpii ok typically, youd set it up like this23:32
scelis recruit0: you shouldn't share one working copy with multiple users. git wasn't designed for this23:32
carpii you have a central repo (which isnt the servers working copy)23:32
← ph^ left23:32
recruit0 Because there's no "locking mechanism" stopping them from breaking it23:32
carpii everyone pulls and pushes their changes to this repo23:32
recruit0 We're all new to git23:32
carpii and on the server, you pull from the repo, to get the clean committed codebase23:33
← alexmchale left23:33
frogonwheels recruit0: have a look at this: http://toroid.org/ams/git-website-howto23:34
recruit0: I'm not saying it's perfect, but you probably want a variation on that23:34
Chillax → Chillaway23:35
→ pheaver__ joined23:35
← pheaver_ left23:35
← pheaver left23:35
pheaver__ → pheaver23:35
recruit0 So basically, no one should be accessing the files directly in our shared folder23:36
I say shared folder because we are working on a college project as a team23:36
carpii what shared folder ?23:36
the git repo is the 'shared' bit23:36
← ormaaj left23:36
→ tvw joined23:36
→ dr_win joined23:37
recruit0 the shared folder is shared so that everyone in the team can write to it...23:37
scelis recruit0: you should have a "master" bare repository that everyone pushes their changes to and pulls updates from23:37
carpii usually changes should only get put onto the server, by doing a gut pull form the server23:37
→ v0n joined23:37
scelis recruit0: that is the "shared" bit. but as far as doing work is concerned, ever user should clone the repository to their own computer or home directory and work on it from there23:37
← andrewvos left23:37
recruit0 So basically, no one should be writing in our master repo folder23:38
e.g. repo ".git" folder is in "public_html" dir23:38
→ ormaaj joined23:38
recruit0 then no one should be writing directly (e.g. nano -w file.txt) in "public_html"23:38
scelis recruit0: that is basically correct23:38
→ _Unode joined23:38
recruit0 ok that's wuts been confusing me XD23:38
← tango_ left23:39
→ luxurymode joined23:39
recruit0 does it matter who runs the commit command on the master repo?23:39
e.g. "cd public_html"23:39
scelis recruit0: though the "master" repository should be somewhere else... for example, you should have a bare, "master" repository in /home/git/myproject.git23:39
→ klunky joined23:39
scelis recruit0: then, in public_html you could clone that master repository and periodically "pull" in the latest changes23:40
recruit0: you would not ever commit in the public_html repo23:40
→ t3i joined23:40
→ tango_ joined23:40
recruit0 ok23:40
cbreak pull can involve committing23:40
recruit0 does it matter who runs "pull"?23:40
scelis recruit0: no23:41
→ Tuller joined23:41
recruit0 i.e. does it have to be whoever init'd the repo on "public_html"23:41
scelis cbreak: don't confuse the poor guy. :) it won't commit if he follows a normal workflow23:41
→ tylerl joined23:41
cbreak it will23:41
dependingon where he pulls from23:41
← rgr left23:41
tylerl can you specify a specific branch in your .gitmodules file?23:41
cbreak it matters who pulls because of the file permissions23:41
make sure to create it with shared permissions23:41
frogonwheels recruit0: you don't 'commit' anything to the public repo.. you push your changes/commits to it.23:42
cbreak tylerl: no23:42
recruit0 we have group permissions activated?23:42
tylerl cbreak: :(23:42
scelis cbreak: if he only pulls from origin and never commits, then pull shouldn't create commit objects unless someone pushes with -f and rewrites history23:42
tylerl thanks23:42
cbreak tylerl: submodules do not track branches23:42
← Wicla left23:42
lepine cbreak: is that a planned feature?23:42
cbreak no23:42
SethRobertson tylerl/lepine: gitslave does23:42
← ruskie left23:42
cbreak there's no point23:42
→ rgr joined23:42
← malesca left23:42
→ Wicla joined23:42
cbreak git submodule binds to a commit23:42
that's it's purpose23:42
← Tuomas_ left23:42
→ malesca joined23:43
SethRobertson gitslave binds to a branch (or tag). that's its purpose23:43
→ Tuomas joined23:43
cbreak just pick your tool23:43
recruit0 So... my teammates can "git pull" in "public_html"?23:43
lepine so it's up to the developer to track (and opssibly automate) submodule version selection in the super project23:43
SethRobertson For git-submodules, yes23:44
tylerl SethRobertson: ... but in your .gitmodules file, you don't specify a commit, you just specify a repo... right?23:44
lepine are there any preferred/standardized methods?23:44
cbreak you get to chose exactly which commit it should be23:44
← remur_030 left23:44
cbreak the revision is recorded in the tree object23:44
scelis recruit0: if permissions are all setup properly, yes, though it might be easier to use rsync or something similar to get files to public_html23:44
SethRobertson tylerl: yes, but when you actually active it, you git add a specific revision the submodule is at23:44
cbreak each commit knows exactly which commit in the submodule to track23:44
← Nikoli left23:44
lepine cbreak: ah! so it's done, just transparently23:45
← dean[w] left23:45
← sitaram left23:45
← dec left23:45
cbreak what is done?23:45
← rgr left23:45
lepine to tell submodule to use such commit, i just check some revision out23:45
→ Nikoli joined23:45
SethRobertson It is very transparent. You get to see the ugly guts and you must perform surgery to change anything23:45
cbreak yeah. but it will bind to the commit, not the branch23:45
← SirFunk_ left23:45
← noahcampbell left23:45
← Vampire0 left23:45
← mjv left23:46
cbreak it's probably the only way to get reproducable history23:46
→ rgr joined23:46
tylerl SethRobertson: so how do I specify which branch to activate it on? `git submodule init` just pulls the HEAD23:46
lepine cbreak: right, so it won't update itself. which is both great and sad. better safe than sorry, though23:46
← workmad3 left23:46
← Thumper_ left23:46
← rgr left23:46
cbreak tylerl: as I said a few times already23:46
→ dec joined23:46
→ sitaram joined23:46
cbreak it does not track branche23:46
← uau left23:46
scelis recruit0: the websites i have in git just have a publish.sh shell script that rsyncs all of the files to the public_html dir on the server23:47
← ndim left23:47
→ ruskie joined23:47
→ mjv joined23:47
→ ndim joined23:47
→ numbah joined23:47
scelis recruit0: this avoids issues like your webserver serving .git files23:47
cbreak init won't pull anything23:47
lepine I approve of it not tracking a branch.23:47
→ zomg joined23:47
SethRobertson tylerl: You check out what version you want (branch, tag, revision, etc) on the submodule, then you go to the superproject and git-add that submodule23:47
scelis recruit0: and makes it easy for anyone with rsync access to update everything23:47
→ rgr joined23:47
cbreak update will check out which ever commit is recorded in the tree23:47
← rgr left23:47
← Ryback_ left23:47
← codespectator left23:48
→ rgr joined23:48
→ taxilian joined23:48
SethRobertson If you want to track a branch instead of a commit (meaning during development you want to be at the tip of the branch named $branch on all subproject repos and the superproject, then you want gitslave. If you want to track a specific revision in the subproject (so if the subproject changes *your* project will not see that change until you specifically ask for it) then you want git-submodules23:49
← msieradzki left23:49
cbreak git submodules will give you verifyable and exact history23:49
→ noahcampbell joined23:50
cbreak which is useful if the submodule is some kind of library23:50
that can change the API23:50
SethRobertson Quite so. Gitslave will only correspond right NOW (at the tip of a branch) or at specific tags.23:50
← recruit0 left23:50
cbreak but the cost is that you have to specify the exact history23:50
← 52AAADOAJ left23:51
← pheaver left23:52
← kpreid left23:52
→ pheaver joined23:52
→ gilaniali joined23:52
→ uau joined23:53
← malesca left23:53
← illustir left23:54
→ Vampire0 joined23:54
← tvw left23:54
→ malesca joined23:54
→ Thumper_ joined23:55
← abetaha left23:55
→ whalesalad joined23:56
→ murr4y joined23:56
Zer Is dircache just another name for the index?23:56
→ fedesilva joined23:56
cbreak I have never heard that word23:56
whalesalad Hey guys, working with a coworker and somehow the css file in this project is being stored as binary. No clue how to fix this.23:56
← Tuller left23:56
cbreak dircache... no idea what that is23:56
← harv left23:57
cbreak whalesalad: ignore it, git stores everything as binary23:57
← orospakr left23:57
whalesalad cbreak: but... I can't diff it23:57
Zer Perhaps it's the byte order marker or such?23:57
→ sagsousuke joined23:57
cbreak hmm... or a .gitattributes override23:57
← oracal left23:58
whalesalad Basically, he made a ton of edits on a live server and now I have to figure out what he changed because he broke a lot of other things. I have all his other changes in another branch that I just merged into the master (probably shouldn23:58
→ dean[w] joined23:58
cbreak it should autodetect text at runtime23:58
← sagsousuke left23:58
→ harv joined23:58
← cannonball left23:58
Zer Well, I know on JGit at least, having a null byte in the first 4096 would make it detect as binary. You could check for null bytes, if ordinary Git does the same anyway23:58
whalesalad Dang... when viewing a diff of the css file this is what I get "Binary files a/b4/css/main.css and b/b4/css/main.css differ"23:58
→ SirFunk_ joined23:59
→ _ollie joined23:59

Logs Search ←Prev date Next date→ Channels Documentation