IRCloggy #git 2012-08-14

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.

2012-08-14

→ john3 joined00:00
← beneggett left00:01
← cybersphinx left00:03
→ cybersphinx joined00:04
← kart0ffelsack left00:04
→ kart0ffelsack_ joined00:05
← alexim left00:05
→ _huoxito joined00:06
← ehsan left00:08
→ idoru joined00:08
← LekeFly left00:09
→ ChadStudmuffin joined00:10
← boombatower left00:12
← Dimitr left00:12
← jn_ left00:14
← macIvy left00:14
→ mmattice joined00:15
← NimeshNeema left00:15
→ corburn joined00:15
→ macabre joined00:16
← cybersphinx left00:16
→ cagedwisdom joined00:17
← FatDarrel left00:18
← _Vi left00:18
sente_ → sente00:18
→ NimeshNeema joined00:18
→ _Vi joined00:18
AlexWaters FauxFaux: got it, thank you00:19
→ scwizard joined00:19
scwizard when I "git rm -r ." all the files get deleted which is what I want00:20
→ harish joined00:20
← maestrojed left00:20
scwizard but then when I check out a different branch, all the files get deleted from that to :(00:20
← harish left00:20
← arcanis left00:21
→ piotrj joined00:22
scwizard oh it happens if you don't commit the rm00:22
i dunno00:23
piotrj Hi all00:23
gitinfo piotrj: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.00:23
→ oojacoboo joined00:23
← ChadStudmuffin left00:23
piotrj can I specify the gitattributes for paths inside the submodule00:23
to be more specific - I want to use a filter attribute on a file that is in a submodule and was wondering whether it's doable at all00:24
← devth left00:24
oojacoboo I have a funny scenario where, we've been merging a new branch with our development branch over the past 3+ months00:24
the issue is that over this time, somewhere the merges took the wrong results00:24
→ ChadStudmuffin joined00:24
oojacoboo so, now when trying to merge into development instead of the other way around, we're already brought up to a certain point and it won't recognize all the changes which we want to force at this point00:25
→ rehat joined00:25
← opus left00:25
oojacoboo is there something we can do on this branch we're trying to merge in to strip of it of it's history and then try and remerge in, so it find all conflicts again?00:25
maybe this is what rebase does or something along those lines?00:25
← strobert left00:26
← glennpratt left00:27
→ opus joined00:27
← tripolar left00:28
→ TheLZA joined00:28
← TheLZA left00:29
→ TheLZA joined00:29
← jargon- left00:29
← Chappo left00:30
→ glennpratt joined00:32
← scwizard left00:33
rohan32 → rohan|dinner00:33
→ v0n joined00:34
→ iamjarvo joined00:34
→ jargon- joined00:36
→ ology joined00:38
→ xiangfu joined00:38
← gavin_huang left00:38
SeoZ-work[AWAY] → SeoZ00:38
→ aspotashev| joined00:39
← robbyoconnor left00:40
← aspotashev left00:40
→ robbyoconnor joined00:41
← maletor left00:42
← fireh left00:43
← kiffer84 left00:43
← dfr|mac left00:44
→ dreiss joined00:46
← hyperair left00:46
→ AaronMT joined00:46
→ mike007 joined00:47
→ Bosox20051 joined00:49
← corburn left00:50
→ corburn joined00:50
← Destos left00:51
→ tshauck joined00:52
← Araxia left00:55
frogonwheels oojacoboo: This is a tricky one. I think you do want something along the lines of a rebase, replacing the merge commits with the commits from the new branch.00:56
frogonwheels: This of course would !rewrite history, but I suspect that's what you need to do. Start by calling the new development a new name...00:56
gitinfo frogonwheels: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)00:56
← glennpratt left00:56
frogonwheels oojacoboo: ^^^ that was for you. Talking to myself *sigh*00:56
→ cjb joined00:57
← iamjarvo left00:58
cjb Hi! I'd like to run a bisect against the mainline kernel, but I have a stack of ten patches that I'm going to need to apply to each bisection point to be able to exhibit the bug. Is there some concept of "git bisect, but add these patches to each step" I could use?00:58
→ BiggFREE joined00:58
← msmithng left00:58
frogonwheels cjb: Sounds like fun. I'd just make a script to do it.00:59
→ msmithng joined00:59
← SkAzZ left00:59
frogonwheels cjb: you could even use git apply -n {sh1} && git reset if you wanted to get the patches from the repo.01:00
← codeFiend left01:00
← textures2 left01:00
→ iori joined01:00
← kenperkins left01:00
cjb frogonwheels: oh, my git-apply doesn't have a "-n" option01:01
what's that do? I'm on 1.7.11.2.01:01
frogonwheels cjb: bleh. I mean git cherry-pick -n would do it.01:01
cjb: git apply takes a patch - and doesn't commit it anyway01:02
cjb: you can use git apply for a patch tho'.01:02
→ flaguy48 joined01:02
frogonwheels cjb: can this bug be checked by a git bisect run ? or does it require manual intervention?01:03
← cmullican left01:03
← artnez left01:04
→ Davey joined01:04
→ dsirijus joined01:05
← hellopat left01:05
← dreiss left01:05
← utdemir left01:08
← Davey left01:10
→ cooloney joined01:10
← iliketurtles left01:11
→ glennpratt joined01:12
← glennpratt left01:13
→ jobin joined01:13
→ fireh joined01:15
→ gumnos joined01:15
← jjbohn left01:16
rohan|dinner → rohan3201:16
→ ehsan joined01:16
→ codeFiend joined01:18
→ Davey joined01:18
← _Vi left01:18
→ _Vi joined01:19
← palm_vue left01:20
← bizhanMona left01:20
← Lgb left01:20
← towski left01:20
→ hyperair joined01:22
→ palm_vue joined01:22
→ Lgb joined01:22
→ bizhanMona joined01:22
← metcalfc left01:23
→ intripoon_ joined01:23
← corburn left01:24
→ erynofwales joined01:24
→ cakehero joined01:24
→ corburn joined01:26
← intripoon left01:27
← mcluskyd_ left01:28
← hfutrell left01:29
← intripoon_ left01:32
limpc anyone happen to know of a great company located in Austin, TX that provides SIP trunking?01:33
er wrong channel01:33
→ intripoon joined01:34
oojacoboo frogonwheels gitinfo we don't really care about the history at this point, the problem is that, the issue branch we have has continual merges that are flat out incorrect, and we need a way to merge this branch against our known good copy, the development branch, to force it to throw conflicts in an attempt to create a new clean merge. We basically want to make it forget it's ever done a merge before01:34
→ Turicas joined01:35
→ Vampire0_ joined01:36
→ tqmz___ joined01:37
→ russt joined01:38
→ zdk joined01:39
← Vampire0 left01:40
← tqmz left01:40
← neurodrone left01:40
← RaceCondition left01:40
→ reds joined01:40
→ nglorioso joined01:41
reds → Guest5070901:41
← ehsan left01:42
gitinfo set mode: +v01:42
nglorioso → yuriismaster01:42
← chihjen left01:43
← opus left01:44
→ opus joined01:44
← Guest50709 left01:45
← hwrd|wor_ left01:47
gitinfo set mode: +v01:47
← hyperair left01:47
← Sym3tri left01:48
→ zivester joined01:49
→ hyperair joined01:49
→ Chib joined01:50
→ subbyyy joined01:51
← shruggar left01:53
← sorahn left01:55
→ iamjarvo joined01:56
→ efreese joined01:57
← duckx left01:57
← FernandoBasso left01:58
← jonshea left01:58
→ intripoon_ joined01:58
← TheJH left01:58
→ xMYTHICx joined01:58
← efreese left01:59
→ flebron joined01:59
flebron So, say I made a commit, and pushed to github, with the wrong commit message. What's an acceptable solution?02:00
← flavius left02:00
→ vitor-br joined02:00
← intripoon left02:00
← apok left02:01
corburn git revert creates a new commit that undoes what the last one did.02:02
← helgikrs left02:03
← graphitemaster left02:03
← raz left02:03
flebron corburn, so I should revert, then re-commit with a correct message?02:03
→ nothingbutlove joined02:04
← john3 left02:05
← asakura left02:05
→ john3 joined02:05
→ maletor joined02:06
← xMYTHICx left02:06
→ asakura joined02:07
→ igotnolegs joined02:09
← pdtpatrick left02:10
→ hfutrell joined02:10
→ hwrd|work joined02:11
← hyperair left02:12
flebron Stupid question: Will reverting the commit revert my edits locally? That is, will I revert my local repo to before the commit?02:12
→ xMYTHICx joined02:13
EugeneKay !revert02:14
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [git checkout -- .]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?02:14
← Marbug left02:14
→ ehsan joined02:14
← _huoxito left02:14
EugeneKay flebron - if you're relatively sure that nobody has fetched your commit from github you can use `git commit --amend` to change the commit message of the last commit(or add more files to it!). If it's farther than back than that you will need to hide the !sausage02:14
gitinfo flebron: [!sausage_making] Some developers like to "hide the sausage making", transforming their commits before presenting them to the outside world. See http://sethrobertson.github.com/GitBestPractices/#sausage and !perfect02:14
EugeneKay flebron - if somebody HAS fetched the history then you should just live with the bad commit message, because !rewriting will incur the wrath of the Git Gods02:15
gitinfo flebron: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)02:15
flebron Haha, yeah, the commit was made just now, and I'm the sole committer in the repo :)02:15
EugeneKay flebron - when you push the new commit up to github you will need to use the -f flag02:15
SethRobertson flebron: the question isn't so much who commits, but who pulls02:16
flebron Ah, yes. That's true.02:16
EugeneKay Like I said, so long as you're pretty sure nobody has fetched ;-)02:16
← JStoker left02:17
EugeneKay Even on relatively busy repos a push -f within 30 seconds probably won't end up in any bad places02:17
→ artnez joined02:17
EugeneKay But get in the habit of checking your history BEFORE you push02:17
← ketas left02:18
→ towski joined02:18
→ ketas joined02:18
flebron Yay, that worked. :)02:18
← _Vi left02:18
flebron What do you mean by "checking m y history"?02:18
*my history02:19
→ _Vi joined02:19
EugeneKay man git-log02:19
gitinfo the git-log manpage is available at http://git-scm.com/docs/git-log.html02:19
EugeneKay Make sure you did what you thought you wanted to did done do dang02:19
→ molixiaoge joined02:20
→ boombatower joined02:21
→ hypera1r joined02:21
flebron Yeah, looks about right :) Thanks!02:21
← msmithng left02:21
→ Marbug joined02:21
← hfutrell left02:24
→ harish joined02:24
frogonwheels oojacoboo: here's a possible way: Create a new devel branch. git checkout devel && git checkout -b devel-v2 Then git rebase -i {sha1} where this is a commit back in history where the other branch started.. and then go and find all the merge commits and replace them with picks that get the commits from your new branch that would have been introduced at that point. OR, you could just delete all the merges from that branch and put all the commits02:24
from your new branch02:24
oojacoboo: !backup02:24
← maletor left02:24
← tjholowaychuk left02:24
gitinfo oojacoboo: Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups02:25
→ jso_ joined02:26
← goshawk left02:27
oojacoboo reading...02:27
← jso_ left02:27
← pxdr left02:27
→ pxdr joined02:28
→ ehsan_ joined02:28
frogonwheels oojacoboo: creating a separate branch also gives you some protection02:28
oojacoboo: actually quite a lot :)02:28
oojacoboo frogonwheels, deleting the merges sounds good02:28
how can you delete the merges on a branch?02:28
← ehsan left02:28
frogonwheels oojacoboo: the problem then is that you lose the fact that you merged.02:28
oojacoboo I don't want any of the merged changes on this branch, they are the whole problem right now02:29
b/c some of them aren't right and I don't know which ones are and aren't02:29
frogonwheels oojacoboo: remember if you do it this way,you're either doing a !rewrite of history, or dumping your old branch and creating a new one.02:29
gitinfo oojacoboo: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)02:29
oojacoboo this is a private issue branch on a colleagues system02:30
we don't care about any history of merges and infact want the merges to have never existed02:30
however, we'd like to keep all the changes made to that branch, just minus the merges02:30
← molixiaoge left02:31
→ molixiaoge joined02:31
→ dhruvasagar joined02:34
← molixiaoge left02:35
→ molixiaoge joined02:36
→ hfutrell joined02:36
→ JStoker joined02:37
← codeFiend left02:37
reactormonk flebron: git commit --amend -m 'new message' && git push --force02:37
→ codeFiend joined02:37
→ _class_ joined02:38
← toothrot left02:38
→ Araxia joined02:40
← codeFiend left02:41
frogonwheels oojacoboo: git rebase -i wil do that for you .. you can just remove the 'pick ' lines for the merges.02:41
← Eridius left02:42
← opus left02:42
→ Who joined02:43
← ehsan_ left02:43
Who Hello, anyone around ?02:43
_class_ → __class__02:43
EugeneKay Who - !hi02:43
gitinfo Who: [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.02:43
← orafu left02:43
← mike007 left02:43
→ orafu joined02:43
Who Hey guys, I heard git is one of the best designed software around. I want to read the source code to find out more but it looks intimidating02:44
→ rchavik joined02:44
← rchavik left02:44
→ rchavik joined02:44
EugeneKay "best designed" is very questionably; it was really more of an evolutionary process.02:44
The object model is a very solid piece of work, but the plumbing commands leave a lot to be desired.02:45
milki http://www.bonkersworld.net/images/2011.11.15_life_of_a_swe.png02:45
like one of these?02:45
EugeneKay Are you approaching this as a research project?02:45
→ project2501b joined02:45
EugeneKay milki - exactly like02:45
Who EugeneKay: no I am a programmer (Ruby on Rails kind to be honest, though I used contribute to KDE some time back), I am curious on how do you design good software.02:46
milki ya, dont look at git >.>02:46
EugeneKay Well, git is really not a software design tool. It's just a piece in the process02:47
milki its rather better to read documentation to see how things work together02:47
EugeneKay There are many books on software design, architecting, and team management. Most of them were written by idiots.02:47
milki design first at the conceptual level before you dive into the implementation02:47
→ neurodrone joined02:47
← neurodrone left02:47
→ neurodrone joined02:47
← AaronMT left02:47
Who EugeneKay: That's why I thought of reading source of a project then a book02:48
→ ehsan joined02:48
EugeneKay It's hard to design a ship by looking at the bilge ;-)02:48
← rehat left02:49
milki but it can be done!02:49
EugeneKay You'll end up with a definite focus on pumps and green things02:49
milki soo goood02:49
oojacoboo frogonwheels, so is it best to just rebase on the development branch each time you do a checkout/start coding?02:50
I'm not sure I quite understand the best workflow02:50
we were basically doing a development branch checkout, then merging it on our issue branch02:50
← BiggFREE left02:50
EugeneKay oojacoboo - depends upon your workflow. I believe that merge commits were a gift from the gods, not the invention of the devil as the SVNites would have you believe.02:50
oojacoboo EugeneKay, yea, well, I come from an SVN background, I'm trying to understand this here02:51
I guess had he actually done proper merges, it wouldn't have been an issue02:51
but since he screwed up at least one if not more of the merges somehow, it's a huge mess02:51
→ jackh joined02:52
Who okay I am curious about how Git achieves the desired performance compare to other VCS, if possible I would just like to read about that02:52
→ hellopat joined02:52
← etcetera left02:53
EugeneKay Pixie dust and unicorn farts02:54
corburn Linus has commented a lot of micro-optimization goes into git.02:54
← hpa left02:54
← rohan32 left02:56
← Who left02:56
→ hpa joined02:57
→ etcetera joined02:57
→ Who joined02:58
Who I am sorry, I got disconnected, last message I got was corburn: Linus has commented a lot of micro-optimization goes into git.02:58
← yuriismaster left02:58
milki learn2screen02:59
← jh3 left02:59
→ jh3 joined02:59
oojacoboo milki, not everyone wants to read irc in their terminal02:59
I sure don't02:59
milki o....02:59
milki forgot those people existed02:59
oojacoboo ;)02:59
chuckf milki: those people don't exist. They are figments of imagination and bots03:01
← Who left03:01
chuckf :)03:01
→ Chappo joined03:01
← dstorrs left03:03
→ opus joined03:03
sharpobject git's default behavior seems to be to replace LF with CRLF on Windows03:04
→ rehat joined03:05
← eletuchy left03:05
sharpobject but to present a merge conflict spanning the entire file if a CRLF revision is merged with a LF revision03:05
this is basically terrible03:05
milki ya, windows -.-03:06
sharpobject it's not a Windows issue, git doesn't have to add CRLF -_-03:06
milki :P03:06
well, people dont complain about it much so...you sure its default?03:06
sharpobject yes03:06
← etcetera left03:08
← tyoc213 left03:08
→ ReekenX joined03:10
sharpobject no03:11
ahaha03:11
EugeneKay AFAIK git on windows uses the "native" setting as the default, which works just fine for me. No CRLFs added. I use Notepad++ for my editor.03:11
← xMYTHICx left03:11
sharpobject how can I see what the config core.autocrlf's current value is?03:11
I want to check this on another machine03:11
EugeneKay man git-config; see --get ;-)03:11
gitinfo the git-config manpage is available at http://git-scm.com/docs/git-config.html03:11
sharpobject I think github for windows may not use git's default as its default03:12
EugeneKay github for windows is a terrible app that eats babies03:12
sharpobject helpfully, core.autocrlf is set to nothing03:13
on my other install03:13
→ Yuuhi` joined03:13
sharpobject so I'm not sure what the default behavior is03:13
milki EugeneKay: but it looks so nice03:14
sharpobject hmm, after setting core.autocrlf to false, git reports that main.lua has been modified, but git diff main.lua outputs nothing03:15
→ etcetera joined03:15
sharpobject this is weird03:15
← Yuuhi left03:15
sharpobject is there some command other than git diff I can use to see the difference?03:17
← __class__ left03:17
← dsirijus left03:17
CareBear\ sharpobject : if you mess around with autocrlf you really should also make a new checkout03:18
sharpobject alright03:18
it seems strange that there is no command to see the diff though >_>03:19
← _Vi left03:19
→ __class__ joined03:19
← etcetera left03:19
→ _Vi joined03:19
← hellopat left03:20
→ dfr|mac joined03:21
→ gnoel joined03:21
EugeneKay sharpobject - you can use any difftool you like03:23
→ dreiss joined03:23
EugeneKay sharpobject - see man git-config; diff.tool03:24
gitinfo sharpobject: the git-config manpage is available at http://git-scm.com/docs/git-config.html03:24
→ arosen joined03:26
sharpobject I see03:26
so it's not a bug for the VCS to report both that a file is different from another and that it is identical to another?03:26
or rather, that the two files both differ and are equal03:27
arosen Hi, i have a patch that i want to apply but git apply patch; told me there was some errors applying it and it couldn't do it. I was wondering if there was anyway to make it apply and I'll go a head and fix the conflicts myself.03:27
Or is it just not possible to even apply?03:27
EugeneKay sharpobject - as far as the built-in diff tool is concerned, a single character on each line indicates a changed line.03:27
sharpobject that does not seem to be true03:28
← drougge left03:28
sharpobject git diff prints 0 characters, but sit status indicates that a file has been modified03:28
EugeneKay !repro03:28
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session, or at least explain exactly what you did that led up to the problem. This will help immensely with troubleshooting.03:28
→ drougge joined03:28
→ rohan32 joined03:30
← dfr|mac left03:30
→ brettgoulder joined03:31
→ codeFiend joined03:31
← Lanks left03:31
gitinfo set mode: +v03:31
← ChadStudmuffin left03:32
sharpobject my first attempt at reproducing this is failing03:33
I replaced the repo with a fresh one, so I don't have one with this problem any more03:33
EugeneKay If I had to guess, you added the files to the index. `git diff` with no further arguments compares the work-tree to the index. `git diff HEAD` shows what is to be commited(index vs HEAD)03:34
sharpobject git status was reporting it as unstaged03:35
→ graphitemaster joined03:35
← luke-jr left03:35
→ luke-jr joined03:36
← kukks left03:36
← jobin left03:36
← Chib left03:36
← tshauck left03:38
← opus left03:38
← simesy left03:40
← cakehero left03:43
→ w09x joined03:45
← ehsan left03:45
→ neurodrone_ joined03:46
← jargon- left03:47
← igotnolegs left03:48
→ igotnolegs joined03:49
← TomyLobo left03:50
→ private joined03:52
sharpobject got it03:52
let me see if I can get it shorter03:52
private → Guest4931303:52
→ etcetera joined03:53
← etcetera left03:53
← codeFiend left03:54
→ aeontech joined03:54
→ etcetera joined03:54
← iamjarvo left03:56
← gnoel left03:56
← etcetera left03:56
→ iamjarvo joined03:57
→ xMYTHICx joined03:57
← OOPMan left04:00
sharpobject hm, the trick didn't work on OSX04:01
but it works on Windows04:01
→ helgikrs joined04:01
→ duckxx joined04:01
sharpobject EugeneKay: http://pastie.org/447103204:04
EugeneKay Well that is interesting.04:05
← duckxx left04:06
← neurodrone left04:06
← hypera1r left04:06
← intripoon_ left04:06
← Lgb left04:06
← bizhanMona left04:06
← smarter_ left04:06
← SJS left04:06
← john_anderson left04:06
← dbr left04:06
← ajpiano left04:06
← reactormonk left04:06
← kpreid left04:06
← jstemmer left04:06
← Swathe left04:06
← bronsen left04:06
→ OOPMan joined04:07
sharpobject git version 1.7.11.msysgit.104:08
EugeneKay Post to the msysgit mailing list? :-p04:08
sharpobject haha04:08
probably04:08
← vitor-br left04:09
← v0n left04:10
← dhruvasagar left04:11
← nothingbutlove left04:12
→ dfr|mac joined04:12
→ Eridius joined04:12
← OOPMan left04:13
→ hyperair joined04:13
→ dhruvasagar joined04:13
→ ferdna joined04:14
→ maestrojed joined04:15
← artnez left04:18
← _Vi left04:19
→ _Vi joined04:20
→ beautifulmind joined04:20
← dreiss left04:20
← towski left04:21
← centipedefarmer left04:23
→ juvenal joined04:24
← Davey left04:25
→ gavin_huang joined04:26
→ OOPMan joined04:28
→ pretty_function joined04:28
← ZyX-I left04:29
→ noyb joined04:29
→ opus joined04:31
← brettgoulder left04:31
← diegoviola left04:31
← beautifulmind left04:32
← hfutrell left04:32
← flebron left04:32
→ wereHamster joined04:32
→ wolftankk joined04:34
← xMYTHICx left04:36
← macabre left04:37
→ beautifulmind joined04:38
→ hobodave joined04:39
← aeontech left04:40
→ codeFiend joined04:41
← erynofwales left04:42
oojacoboo wow, really not having much luck here04:43
git def makes things complicated it seems04:43
I've spelled out the issue in here earlier04:43
basically, there is a branch that has a bunch of merges with another branch and I want to trash all those merges04:44
← beautifulmind left04:44
oojacoboo or… leave them in place somehow and merge it again, but I don't want the git to think it's already merged04:44
→ etcetera joined04:45
oojacoboo basically, I just want to have it try and do a fresh merge again so it can find changes and conflicts without assuming everything is just hunky-dorey, which it is not04:45
→ beneggett joined04:45
EugeneKay Do you care about published history?04:45
eg, will you be burned at the stake for doing a force-push?04:45
oojacoboo no04:46
← rohan32 left04:46
oojacoboo I don't care about anything but the development branch at this point04:46
→ Davey joined04:46
oojacoboo this issue branch just needs to forget that it has anything done to it except for the status of the files04:46
EugeneKay use `git reset --hard $SHA` on each branch involved to put it back to it's pre-merge state. Then just do your merges cleanly, creating new merge commits. Use the -n flag on git-merge to inspect the results of the automerge before you commit04:46
oojacoboo then I want to merge it onto development so it can perform it's diff04:46
EugeneKay Use the !lol to find the right SHA to reset to for your issue branch04:47
gitinfo git config --global alias.lol "log --oneline --graph --decorate"04:47
oojacoboo mmm04:47
my god at the flags04:47
EugeneKay You should probably also reset your "development" branch to just before you started merging into it04:47
oojacoboo EugeneKay, each branch involved means what?04:48
EugeneKay The branch(es) that you're merging into the "development" branch.04:48
oojacoboo ok, so the single branch I'm referring to here then04:48
EugeneKay Sure04:48
oojacoboo and the sha I'm going to get by running that UN of flags command?04:49
EugeneKay Also the "development" branch itself04:49
How you get the SHA is immaterial. `git log --oneline --graph --decorate` is just a quick, easily digestible view.04:50
oojacoboo gotcha04:50
← piotrj left04:50
oojacoboo I have to reset the development branch as well?04:50
EugeneKay I would, so the bad merge disappears into the mists of the objects/ dir.04:50
You want to forget it anyway, so just kill it entirely.04:50
oojacoboo doesn't that kill the history though?04:51
I want to maintain that on development04:51
EugeneKay Yeah, but only that merge commit.04:52
You can create a new commit on development that puts it back at the exact state it was before the merge.04:52
`git checkout $SHA -- .`, then commit with a message of "Going back to pre-merge state" or something.04:53
Where $SHA is the one just before the merge04:53
oojacoboo TMI man04:53
EugeneKay That means you're learning ;-)04:53
→ Chibby_ joined04:54
oojacoboo yes, that's one way of putting it04:54
→ macIvy joined04:54
← opus left04:56
← beneggett left04:58
→ maleknet joined04:58
← etcetera left04:59
→ dstorrs joined04:59
→ duckxx joined04:59
→ intripoon_ joined04:59
→ bizhanMona joined04:59
→ Lgb joined04:59
→ smarter_ joined04:59
→ SJS joined04:59
→ john_anderson joined04:59
→ dbr joined04:59
→ ajpiano joined04:59
→ reactormonk joined04:59
→ kpreid joined04:59
→ jstemmer joined04:59
→ Swathe joined04:59
→ bronsen joined04:59
→ piotrj joined05:00
→ SwK joined05:01
← mrwalker left05:02
← cjb left05:02
→ cjb joined05:02
← hobodave left05:02
→ etcetera joined05:02
← dfr|mac left05:03
← w09x left05:03
→ Tankado joined05:03
→ ovnicraft joined05:04
Tankado Is there any git client that shows the branches tree in a vertical way ? (similar to IBM telelogic synergy if anyone is familiar) like master-> commit -> commit , and then a down arrow for a branch05:05
→ zapsoda joined05:05
zapsoda Hi, I was wondering if there was a way to send a pull request from the Git Bash?05:05
← rue_XIV left05:06
→ rue joined05:07
→ carpii joined05:08
carpii i made some temp commits on a branch, and merged it into master. thats all fine, but now i want to revert those changes (it was for website downtime). Is there a way to revert 5 or 6 commits in one go?05:08
← Chappo left05:09
→ leeN joined05:09
→ Chappo joined05:10
← cwang left05:10
← dhruvasagar left05:12
→ cwang joined05:12
→ dhruvasagar joined05:12
→ jgoss joined05:14
→ arvind_khadri joined05:15
carpii seems its git revert --no-commit i needed05:16
← arvind_khadri left05:17
→ Fullmoon joined05:17
→ jargon- joined05:18
← dstorrs left05:19
← duckxx left05:19
← intripoon_ left05:19
← Lgb left05:19
← bizhanMona left05:19
← smarter_ left05:19
← SJS left05:19
← john_anderson left05:19
← dbr left05:19
← ajpiano left05:19
← reactormonk left05:19
← kpreid left05:19
← jstemmer left05:19
← Swathe left05:19
← bronsen left05:19
← _Vi left05:20
→ _Vi joined05:20
← Davey left05:20
→ BadDesign joined05:22
→ NyanChat joined05:22
← NyanChat left05:22
← BadDesign left05:22
→ BadDesign joined05:22
← TheLZA left05:23
→ arvind_khadri joined05:23
← carpii left05:23
→ leila joined05:24
→ unmanbearpig joined05:26
oojacoboo EugeneKay, that !lol didn't work for me, what's the sha, the latest commit for the branch, or the first?05:27
gitinfo EugeneKay: git config --global alias.lol "log --oneline --graph --decorate"05:27
EugeneKay Look at the log, don't ask me05:27
oojacoboo oh… it adds to the log05:27
doooop05:27
← arvind_khadri left05:28
← ovnicraft left05:29
→ jrondam joined05:29
← Fretta left05:29
oojacoboo how do I view the log… sigh05:29
→ arvind_khadri joined05:29
EugeneKay <EugeneKay> How you get the SHA is immaterial. `git log --oneline --graph --decorate` is just a quick, easily digestible view.05:29
Run that ^05:29
→ s0ber_ joined05:29
oojacoboo oh, the config was just to enable that, I assume05:30
I thought that was one command05:30
← gavin_huang left05:30
oojacoboo thanks05:30
EugeneKay the config creates an alias named lol which does that05:30
← dhruvasagar left05:30
← Turicas left05:32
← s0ber left05:32
← pigeon left05:32
oojacoboo EugeneKay, I'm still not sure where I'm suppose to find the hash, this appears as a series of latest commits05:32
s0ber_ → s0ber05:32
oojacoboo I feel like the biggest noob on the planet with this crap05:32
EugeneKay !treeish05:32
gitinfo A tree-ish is something that looks like a tree. Read 'man gitrevisions' and http://git-scm.com/book/en/Git-Tools-Revision-Selection05:32
EugeneKay "Short SHA"05:32
← jrondam left05:32
oojacoboo yes, right, but these are jsut commit hashes05:32
EugeneKay Yes.... that's what you want.05:33
oojacoboo just any one?05:33
doesn't matter?05:33
EugeneKay The one just before the merge commit....05:33
oojacoboo what if there are like 10 merge commits?05:33
and I want to keep all of the other commits, just not the merges05:33
→ beautifulmind joined05:33
→ pigeon joined05:33
oojacoboo I'm not trying to roll back one bad merge commit05:34
← iamjarvo left05:34
Tankado Is there any git client that shows the branches tree in a vertical way ? (similar to IBM telelogic synergy if anyone is familiar) like master-> commit -> commit , and then a down arrow for a branch05:34
oojacoboo this is an issue branch where the dev thought a continual merge was the best way to stay in sync05:34
→ snearch joined05:34
oojacoboo Tankado, that's the SVN way of thinking05:34
← Chappo left05:34
oojacoboo branches don't belong, they are linear05:35
→ Chappo joined05:35
oojacoboo dvcs is different05:35
and hard to understand05:35
→ ovnicraft joined05:35
Tankado i just want to see the branches vertically they can be linear05:35
oojacoboo Tankado, there are lots of apps out there05:36
← schemula left05:36
← wereHamster left05:36
→ pantsman joined05:36
← pantsman left05:36
→ pantsman joined05:36
oojacoboo EugeneKay, I'm assuming that's not going to work for my issue05:36
I'm thinking a git rebase is still needed05:36
EugeneKay Continual... merge... like, at every commit?05:36
oojacoboo EugeneKay, before coding05:36
EugeneKay makes a rude noise05:37
← rchavik left05:37
EugeneKay You're gonne have a buncha fun with rebase and/or cheryr-pick05:37
oojacoboo I know05:37
→ rchavik joined05:37
oojacoboo there has to be a way to trash this branch's history somehow05:37
flatten it05:37
EugeneKay reset --hard ;-)05:37
oojacoboo disassociate it, then try and fresh merge05:37
EugeneKay You have to recreate it commit-by-commit, sans the merging05:37
oojacoboo but that kills everything to that point, no?05:37
EugeneKay Yes, but if you leave a nonce branch there you can still get the SHAs via the log, and then cherry-pick into your pristine branch05:38
← AlbireoX left05:38
oojacoboo ah, so, you're saying just add a new branch and cherry pick?05:39
cherry-pick merge*05:39
if that's even a thing05:40
_ikke_ oojacoboo: dvcs is not really hard to understand, just different05:40
← Tankado left05:40
← marc_kdab_ left05:40
→ marc_kdab_ joined05:41
→ reds joined05:41
_ikke_ oojacoboo: cherry-picking is not merging05:41
oojacoboo _ikke_, right, I know05:42
reds → Guest8255905:42
oojacoboo but can I cherry-pick merge so to speak05:42
aka, grab a certain commit and merge into another branch05:42
_ikke_ You don't merge it in a different branch05:42
→ gavin_huang joined05:42
_ikke_ you copy it onto a different branch05:42
oojacoboo ok, sure, that too, I can do that?05:42
AAA_awright What would be the best way to search a list of local repositories for the existance of any number of commit Ids?05:42
_ikke_ AAA_awright: Write a script that does that05:43
AAA_awright How is the advice I am seeking05:43
_ikke_ oojacoboo: man git cherry-pick05:43
gitinfo oojacoboo: the git-cherry-pick manpage is available at http://git-scm.com/docs/git-cherry-pick.html05:43
→ Chris____ joined05:43
AAA_awright I could brute force search it but I'm thinking there's more efficent ways05:43
SwK damn it man... there needs to be a git cherry-pick --interactive like for rebase05:43
→ olaf_ joined05:44
→ irqq joined05:44
← juvenal left05:44
_ikke_ SwK: Why do you need that?05:44
SwK _ikke_: stable branching05:45
where as master continues along as dev branch and being able to interactively pull patches to say v1.x branches05:45
→ dvide joined05:46
← Guest82559 left05:46
→ floppyears joined05:48
← floppyears left05:48
→ floppyears joined05:48
← Kyzz left05:49
← Eridius left05:49
→ Eridius_ joined05:49
→ wereHamster joined05:51
→ tekbwainz joined05:52
← tekbwainz left05:52
→ tekbwainz joined05:52
→ j416 joined05:52
← gavin_huang left05:54
← Eridius_ left05:55
→ apok joined05:55
→ arkiver joined05:56
AAA_awright Well how about this: What's the fast way to check that a commit merely exists?05:56
← neurodrone_ left05:57
SwK fisheye05:57
AAA_awright actually I have a better idea05:57
→ Eridius_ joined05:58
_ikke_ AAA_awright: git cat-file -s <hash>05:58
Eridius_ → Eridius05:58
← cenuij left05:59
wereHamster rev-parse05:59
oojacoboo AAA_awright, http://stackoverflow.com/questions/2444100/how-can-i-determine-if-a-given-git-hash-exists-on-a-given-branch05:59
git branch -a --contains 4f08c85ad05:59
→ gavin_huang joined06:00
← yoklov left06:00
← OOPMan left06:01
← maestrojed left06:02
→ aeontech joined06:02
← codeFiend left06:02
aeontech → codeFiend06:02
_ikke_ That's on a branch06:03
he just want to check if it exists in the repository06:04
→ cenuij joined06:04
← cenuij left06:04
→ cenuij joined06:04
← codeFiend left06:08
→ mishok13 joined06:10
AAA_awright _ikke_: actually oojacoboo makes a good point, the commit could exist, but it's just not refereneced by any branch06:10
→ lirt joined06:10
→ afuentes joined06:11
_ikke_ AAA_awright: Let's drop back for a moment. What is the thing you actually want to achieve?06:12
AAA_awright Actually, I'm trying to see if two repositories are the same (i.e., compatible versions of each other)06:13
wereHamster AAA_awright: define 'compatible'06:13
→ lb joined06:14
← lb left06:14
→ lb joined06:14
AAA_awright Working on that heh, uh06:14
→ cbreak-work joined06:14
→ chriskon149 joined06:14
AAA_awright They may be from different remotes/mirrors, or be forks, but they both have commits in common that are marked as the commits that I'm compatible with06:14
oojacoboo diff ?06:15
→ jkremser joined06:15
wereHamster so your criterion is that both have a common commit?06:15
← lirt left06:16
AAA_awright Yeah.06:16
wereHamster (certain special commit which you know by hash)06:16
git branch --contains ...06:16
← osmosis left06:17
AAA_awright More specifically: I have a repository, and range of commits that I can check out, and another repository, and range of commits I can check out, I want to see if any of those commits are the same06:17
← ferdna left06:17
← j416 left06:17
oojacoboo AAA_awright, you could just run diff, ya know?06:17
→ arcanis joined06:18
→ papegaaij joined06:18
AAA_awright How would that help me06:18
← gavin_huang left06:18
oojacoboo if you want to know if they're different...06:18
wereHamster AAA_awright: 'I want to see if any of those commits are the same' - compare the commit hashes06:18
oojacoboo if you're looking to find missing commits, etc, that's a diff story06:18
→ sqlbyme joined06:18
AAA_awright That's not generally something you can run cross-repository06:19
← jh3 left06:19
→ sagan8972 joined06:19
← sqlbyme left06:19
→ jh3 joined06:20
→ marthinal joined06:20
← _Vi left06:20
← sagan8972 left06:20
→ sqlbyme joined06:20
← Tommy[D] left06:20
← jh3 left06:20
→ _Vi joined06:20
→ jh3 joined06:21
→ Tommy[D] joined06:21
→ gavin_huang joined06:21
→ codeFiend joined06:22
→ KTMBoID joined06:22
← KTMBoID left06:22
← arcanis left06:22
→ skorgon joined06:22
← skorgon left06:22
→ skorgon joined06:22
wereHamster AAA_awright: you can not run anything cross-repository. You always have to fetch first to get the objects into a single repo06:23
← chriskon149 left06:23
→ senny joined06:24
← sqlbyme left06:24
→ maletor joined06:25
← mishok13 left06:25
→ kleist joined06:27
← madewokherd left06:29
← darren left06:30
← maletor left06:30
← pantsman left06:31
→ pantsman joined06:32
← arkiver left06:33
← avpx left06:34
→ nor5trom joined06:35
→ toabctl joined06:37
→ dstorrs joined06:37
→ intripoon_ joined06:37
→ bizhanMona joined06:37
→ Lgb joined06:37
→ smarter_ joined06:37
→ SJS joined06:37
→ john_anderson joined06:37
→ dbr joined06:37
→ ajpiano joined06:37
→ reactormonk joined06:37
→ kpreid joined06:37
→ jstemmer joined06:37
→ Swathe joined06:37
→ bronsen joined06:37
← nor5trom left06:37
← reactormonk left06:37
← macIvy left06:38
→ rurufufuss joined06:39
← Chappo left06:42
tekbwainz → phreaky06:42
← chepre left06:42
→ Chappo joined06:42
← gavin_huang left06:43
← maleknet left06:44
← leila left06:44
→ giallu joined06:45
← marthinal left06:46
← molixiaoge left06:47
→ gavin_huang joined06:47
→ marthinal joined06:47
← Bosox20051 left06:48
← codeFiend left06:51
→ aeontech joined06:51
← noyb left06:52
→ molixiaoge joined06:52
→ molixiaoge_ joined06:53
← molixiaoge_ left06:53
← zapsoda left06:57
→ jceb joined06:57
→ jdav_gon` joined06:57
← aartist left06:58
← Chappo left06:59
→ Chappo joined06:59
← dharrigan left07:00
← jdav_gone left07:00
→ arkiver joined07:01
← etcetera left07:01
→ flijten joined07:02
← ovnicraft left07:02
← subbyyy left07:03
← Goplat left07:03
← igotnolegs left07:03
← apok left07:04
→ alexukf joined07:06
← sedavand_ left07:06
→ ph^_ joined07:06
→ mishok13 joined07:07
← gavin_huang left07:08
→ maleknet joined07:08
← boombatower left07:08
→ dhruvasagar joined07:09
← maleknet left07:09
← skorgon left07:10
→ gavin_huang joined07:10
→ maleknet joined07:10
→ nor5trom joined07:11
← nor5trom left07:12
→ psoo joined07:12
→ nor5trom joined07:13
→ arturaz joined07:14
→ thiago joined07:14
← Tommy[D] left07:15
→ fornext joined07:16
→ skorgon joined07:18
← skorgon left07:18
→ skorgon joined07:18
← palm_vue left07:20
← bizhanMona left07:20
← Lgb left07:20
← maleknet left07:21
→ maleknet joined07:21
← maleknet left07:21
→ Spabby joined07:22
gitinfo set mode: +v07:22
→ palm_vue joined07:22
→ Lgb joined07:22
→ bizhanMona joined07:23
→ etcetera joined07:23
Spabby good morning, I am trying to use git bisect to find out which commit has caused a bug, I have known good and bad revisions, but I can't seem to get it to work, should my local copy be at the good or bad revision before starting please?07:23
→ dsirijus joined07:23
→ Tommy[D] joined07:24
_ikke_ Spabby: That doesn't really matter. You just specify what what is07:24
Spabby: git bisect start07:24
If you're on a good revision: git bisect good; git bisect bad <sha-of-bad-hash>;07:24
After that, bisecting begins07:24
← nocturnal left07:25
← cdash004 left07:25
← rehat left07:25
_ikke_ Spabby: Note that you can also do git bisect start <bad> [<good>]07:25
← dhruvasagar left07:25
← toabctl left07:25
→ dhruvasagar joined07:26
→ cdash004 joined07:26
→ dc5ala joined07:27
→ soee joined07:28
← gavin_huang left07:29
→ OOPMan joined07:30
← lb left07:30
← aeontech left07:31
→ lb joined07:32
Spabby _ikke_: thanks07:32
← SyouJyoA left07:32
→ gavin_huang joined07:33
→ _Qman joined07:33
← flijten left07:34
magn0z_ if a branch has been pushed to other repos, and that branch needs to catch up with the master branch, what's the proper way of doing it? i find merging the master onto the branch is really ugly, since you end up with a merge commit that isn't really related to the branch. and i can't rebase, since the branch is pushed to others07:34
→ _oakdog8 joined07:34
← oakdog8 left07:34
→ bjensen joined07:35
← bjensen left07:35
→ bjensen joined07:35
wereHamster magn0z_: you can't rebase and you don't want to merge. The answer is pretty clear then.07:36
→ opus joined07:36
← etcetera left07:36
magn0z_ wereHamster, create a new branch?07:37
_ikke_ magn0z_: Are you the only one working on that branch, or are others also building on top of it?07:37
magn0z_ _ikke_, well, others review it and give me suggestions on improvements. im the one working on it, but others have it in their repos07:38
← dzonder left07:38
← Spabby left07:38
_ikke_ Do you really *need* to incorporate changes from master?07:39
← beautifulmind left07:39
→ reds joined07:42
→ rmr joined07:42
reds → Guest1857807:42
→ nhooey|mini__ joined07:42
← nhooey|mini_ left07:43
← Chappo left07:43
← ology left07:43
← SHAGGSTaRR left07:44
magn0z_ _ikke_, usually, no.. but if a branch has fallen very far behind, will it not cause merge problems if i attempt to merge it directly with master? instead i solve the merge conflict by merging master to my branch, and then merging the branch to master. perhaps thats what i'm doing wrong? is it possible that this will ease the merge?07:45
→ TheHaven joined07:45
→ shreya joined07:46
→ kkszysiu_work joined07:47
_ikke_ magn0z_: Well, merging master to branch or branch to master will not make a difference with merge conflicts07:47
← Guest18578 left07:47
→ Sebo joined07:48
→ Chappo joined07:49
→ polymar joined07:49
_ikke_ magn0z_: And if it's a topic branch where no one else builds on, rewriting it is also not a problem07:49
magn0z_: the git project also regularly rewrites published branches07:50
← gavin_huang left07:50
_ikke_ (pu)07:50
→ landypro_ joined07:50
← wolftankk left07:50
← shreya left07:51
→ shreya joined07:51
→ wolftankk joined07:51
← wolftankk left07:52
← rmr left07:52
→ SHAGGSTaRR joined07:52
→ timthelion joined07:53
→ flijten joined07:53
← arkiver left07:54
← harish left07:55
→ Element9 joined07:55
→ und3f joined07:55
← gxk left07:55
→ gavin_huang joined07:55
→ LimeBlast joined07:56
→ dpino joined07:59
→ Jay_Levitt joined07:59
← gavin_huang left07:59
→ tvw joined07:59
→ Speed joined08:00
→ eletuchy joined08:00
→ lukosanthropos joined08:00
← piotrj left08:02
← pantsman left08:02
← quazimodo left08:04
← ReekenX left08:04
→ gavin_huang joined08:04
→ quazimodo joined08:06
→ dr_lepper joined08:06
→ ReekenX joined08:06
magn0z_ _ikke_, ok, but won't people have to do git pull -f to get the new stuff? even worse, what will happen if I publish a branch for review, fix any comments, rebase the branch and then merge it to master, without publishing the new branch. in that case, the branch will be merged in my repo, and others will get the same stuff on their master, while still having this "dead" semi-complete published branch with the same name as the one i merged08:06
→ arkiver joined08:08
← hwrd|work left08:09
_ikke_ magn0z_: If you use git pull or git pull origin, you don't have to use -f08:09
And after that branch has been merged in, it can possibly be removed08:09
← arkiver left08:09
_ikke_ But if you fix anything, you should still push that so it can be reviewed again08:10
→ d3vic3 joined08:10
→ arkiver joined08:10
← Eridius left08:11
← jkremser left08:11
← arturaz left08:14
→ narfnarfnarf joined08:15
← arkiver left08:16
← shreya left08:16
→ shreya joined08:16
← bjensen left08:17
← _Vi left08:20
→ _Vi joined08:21
→ BiggFREE joined08:21
← d3vic3 left08:24
→ d3vic3 joined08:25
← corburn left08:26
← ph^_ left08:27
← d3vic3 left08:30
← asakura left08:31
→ ph^ joined08:33
→ d3vic3 joined08:34
← d3vic3 left08:35
→ torbjorn joined08:36
← opus left08:36
→ Nibato joined08:36
→ mabrand joined08:36
→ diegoviola joined08:39
→ asakura joined08:39
← hays left08:40
→ Vert joined08:40
→ d3vic3 joined08:40
← iori left08:42
→ martinjlowm joined08:42
← d3vic3 left08:44
→ tinti joined08:45
→ mmc1 joined08:46
→ d3vic3 joined08:46
← tinti left08:47
← d3vic3 left08:47
← mabrand left08:47
→ mabrand joined08:48
← flijten left08:49
← Araxia left08:49
← babilen left08:50
← dhruvasagar left08:51
← und3f left08:51
← javawizard left08:51
→ dhruvasagar joined08:51
→ und3f joined08:51
→ jkremser joined08:52
← thiago left08:53
← mabrand left08:53
← snearch left08:53
← tweedledumber left08:54
→ tatsuya_o joined08:55
← OOPMan left08:55
→ tweedledumber joined08:56
SeoZ → SeoZ-work[AWAY]08:56
← tqmz___ left08:58
→ tqmz joined08:58
← archdori left08:59
comps_ what is the branches/ directory inside GIT_DIR for? .. is it for some backwards-compatibility?09:00
→ archdori joined09:00
→ archdori_ joined09:00
← Fandekasp left09:00
← dhruvasagar left09:00
← asakura left09:00
→ dhruvasagar joined09:01
_ikke_ comps_: afaik, yes09:02
← oojacoboo left09:02
→ asakura joined09:02
→ ryez joined09:04
→ gusnan joined09:04
→ shruggar joined09:04
← john3 left09:05
→ oojacoboo joined09:05
→ gxk joined09:05
← ryez left09:05
→ flijten joined09:05
→ iori joined09:06
→ ashirus joined09:06
→ john3 joined09:06
archdori_ → Fandekasp09:07
→ OOPMan joined09:09
← alexukf left09:10
← pxdr left09:10
← BiggFREE left09:11
→ xMYTHICx joined09:11
← xMYTHICx left09:11
→ xMYTHICx joined09:12
← polymar left09:13
← timthelion left09:17
← borwar left09:18
→ fructose joined09:20
fructose Anyone know a way to deploy (push) a git project to a webserver when the host doesn't itself support git?09:20
I tried sshfs, but it was incredibly slow to run any git commands09:20
_ikke_ fructose: Don't use git itself to deploy09:20
→ LekeFly joined09:21
← _Vi left09:21
fructose _ikke_: Can you suggest an alternative that doesn't require server support?09:21
_ikke_ fructose: Just use something like git archive to extract the project, and use sftp, rsync or something else to deploy it09:21
→ borwar joined09:21
← e0je left09:21
← eletuchy left09:21
fructose _ikke_: The issue is that I do not want to copy over everything at once... don't think the host supports rsync either09:22
_ikke_ fructose: rsync uses ssh09:22
→ _Vi joined09:22
_ikke_ (can use)09:22
fructose: What do you mean that you don't want to copy over everything at once?09:22
→ babilen joined09:22
→ achadwick joined09:22
fructose _ikke_: Just want diffs, not the whole project every time09:23
_ikke_ rsync can do diffs09:23
← asakura left09:23
fructose _ikke_: Looking into rsync via ssh now, but the suggestion seems to be to set up public key encryption. Is that required to avoid typing a password for each file?09:23
_ikke_ fructose: Not sure it's required for each file, but I do recommend setting up ssh keys09:24
→ barq joined09:25
fructose _ikke_: I'll try it out, thanks09:26
← john3 left09:27
→ Xilent joined09:27
→ john3 joined09:27
→ alexukf joined09:29
→ Rajesh joined09:32
Rajesh → Guest8779209:32
→ thiago joined09:34
→ whitman joined09:34
→ TomyLobo joined09:35
barq How can I move a github repository to a local git repository?09:37
→ SyouJyoA joined09:37
_ikke_ barq: Just clone it09:37
barq So just git clone rep?09:38
_ikke_ yeah09:38
→ TheJH joined09:40
→ archdori_ joined09:40
← Fandekasp left09:41
← archdori left09:41
→ berserkr joined09:42
→ reds joined09:42
reds → Guest5755909:42
← SyouJyoA left09:43
→ udantee joined09:43
← Guest57559 left09:46
→ d3vic3 joined09:47
← LekeFly left09:48
→ Fandekasp joined09:50
→ tsdh joined09:51
barq _ikke_: I think I didn't make myself clear. I want to stop using github and use a local server as a repository instead. If I just do git clone rep I'm not sure I achieve that.09:52
← gavin_huang left09:52
← archdori_ left09:52
← jackh left09:52
_ikke_ barq: Yes, a git clone is a full clone of the repository09:52
cmn you can use --mirror09:52
but that will bring in the github-specific refs09:52
barq How do I pull/push to that repository instead of github now?09:53
_ikke_ barq: I would recommend looking at !gitolite09:53
gitinfo barq: Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html09:53
_ikke_ It's quite easy to set up09:53
You can push your repo to gitolite, and then everyone else can clone from that09:54
cmn you do the same as you do when the repo is on github, but you use the other url09:54
→ tjadc joined09:57
→ gavin_huang joined09:58
← Fandekasp left10:01
barq How can I change the url from git to my local repository which needs ssh access?10:02
← iori left10:03
→ timthelion joined10:05
← OOPMan left10:07
← d3vic3 left10:07
← gavin_huang left10:08
← hyperair left10:09
← jjang left10:10
cmn man git remote set-url10:11
gitinfo the git-remote manpage is available at http://git-scm.com/docs/git-remote.html10:11
→ beautifulmind joined10:11
← fructose left10:11
→ asakura joined10:11
← fireh left10:12
← shruggar left10:12
← dhruvasagar left10:14
→ dhruvasagar joined10:15
→ iori joined10:22
← xMYTHICx left10:23
→ fireh joined10:28
← comps_ left10:28
barq I added the remote and now I get the following error when attempting to push: http://pastie.org/private/z5orayhhvcorka3bgmiq10:29
← f0i left10:30
← unmanbearpig left10:30
→ LekeFly joined10:30
_ikke_ barq: read that message10:31
!bare10:31
gitinfo an explanation of bare and non-bare repositories (and why pushing to a non-bare one causes problems) can be found here: http://bare-vs-nonbare.gitrecipes.de/10:31
barq Shouldn't I use a bare rep instead?10:31
→ mritz_ joined10:32
→ eletuchy joined10:32
← mritz left10:33
_ikke_ barq: You *should* use a bare repo instead10:34
→ adityamenon joined10:34
→ jonnu joined10:35
barq That's what I was wondering before. So I should clone as bare10:35
← eletuchy left10:36
cmn create a bare repo and push whatever you need10:38
→ OOPMan joined10:39
→ f0i joined10:39
barq Thanks.10:40
← Guest87792 left10:44
→ api joined10:44
→ Fandekasp joined10:48
→ specialGuest joined10:48
← specialGuest left10:49
→ specialGuest joined10:49
barq Is it common for this remote to be named origin rather than master?10:50
← nor5trom left10:50
← arvind_khadri left10:52
_ikke_ barq: origin -> remote, master -> branch10:52
→ arietis joined10:52
← specialGuest left10:53
→ nor5trom joined10:53
← jceb left10:55
cmn it is very uncommon for a remote to be named master10:55
← alexukf left10:55
→ arvind_khadri joined10:55
sitaram and for a branch to be named origin I guess :)10:57
muep origin is a kind of default name for remotes10:57
_ikke_ muep: It's the name of the remote where the repository was cloned from10:57
← soee left10:58
→ cilly joined10:58
→ Superlokkus joined10:59
← eicto left11:00
← mbrochh left11:00
→ angelsl joined11:00
→ eicto joined11:01
← arvind_khadri left11:02
← flaguy48 left11:03
← cilly left11:04
→ d3vic3 joined11:04
→ arkiver joined11:10
→ aaronky- joined11:11
→ AAA_awright_ joined11:11
→ s0undt3ch_ joined11:11
← ReekenX left11:11
→ specialGuest joined11:12
→ aromasca_ joined11:13
← zdk left11:13
→ andres_f joined11:13
← eicto left11:13
→ pavelz_ joined11:13
→ kinlo_ joined11:13
→ raar joined11:13
→ wart____ joined11:13
→ tgummere1 joined11:13
→ alexukf joined11:13
→ svm_invi1tvs joined11:14
raar → Guest5349011:14
← adityamenon left11:14
→ saurik joined11:14
→ hendry_ joined11:14
→ malesca_ joined11:14
→ raek_ joined11:14
→ oehman joined11:14
andres_f → andres11:15
← gumnos left11:18
→ gry- joined11:18
← gry left11:18
← hendry left11:18
← kinlo left11:18
← AAA_awright left11:18
gry- → gry11:18
→ eighty4 joined11:19
← Xilent left11:19
← xiangfu left11:20
→ rwestphal joined11:21
← _Vi left11:22
← kinlo_ left11:22
→ _Vi joined11:23
→ fusion27 joined11:23
ChanServ set mode: +v11:23
→ Mannequin joined11:23
jast EugeneKay: no, but patches welcome11:23
→ Poapfel joined11:23
→ ReekenX joined11:24
→ sontek joined11:25
→ awx joined11:25
→ daethorian joined11:26
→ hyperair joined11:26
← arkiver left11:28
← kleist left11:28
← Superlokkus left11:29
← torbjorn left11:29
→ Guest2786 joined11:30
→ utdemir joined11:31
unreal_ → unreal11:33
← Guest2786 left11:33
→ torbjorn joined11:34
→ blami_ joined11:34
← aspotashev| left11:35
→ arkiver joined11:36
→ centipedefarmer joined11:38
← arkiver left11:40
→ beepbeep_ joined11:41
→ arkiver joined11:41
→ RaceCondition joined11:41
→ ipalaus joined11:41
Chris____ → chrisward11:42
← dr_lepper left11:42
→ reds joined11:43
reds → Guest5299511:43
← Psycho_pr1 left11:43
→ Psycho_pr2 joined11:43
→ cannonball joined11:43
← dhruvasagar left11:45
beepbeep_ Hey all, question, I don't get something here. I modify two files, add and commit. I try to push and git tells me that I have to pull first. Okay, I get that. I pull and get a ton of new files in. Still get this part. Now git tells me that auto merge failed on the two files I changed and that I need to merge them, when done I have to add them. So far so good. I edit and add them. Now I commit -m11:46
"Merge" and push. When I check the log, I see that it looks like I was the one who made all of those new files that I pulled in, I have 109000 additions lol.. What am I missing here -_-11:46
→ dr_lepper joined11:46
← dr_lepper left11:46
→ dr_lepper joined11:46
beepbeep_ Wall of text. sry.11:46
jast oh, that's fine. it's much better than people we have to interview for two hours to figure out what they want. ;)11:46
← narfnarfnarf left11:46
→ ChadStudmuffin joined11:46
charon also, finally someone who commits before pulling!11:47
jast anyway, to answer your question: you created a merge commit11:47
← Guest52995 left11:47
jast that is important information: it tells people that *you* combined the two lines of work11:47
beepbeep_ But only two files needed merging.11:47
not 89 files :(11:47
jast yes11:47
← pretty_function left11:47
jast the information you see tells you what the merge in total added to your current branch11:47
→ pretty_function joined11:47
jast but you can also see exactly what commits were merged in11:48
let's see...11:48
!lol11:48
gitinfo git config --global alias.lol "log --oneline --graph --decorate"11:48
jast !lg11:48
gitinfo git config --global alias.lg "log --graph --pretty=format:'%Cred%h%Creset -%C(yellow)%d%Creset %s %Cgreen(%cr) %C(bold blue)<%an>%Creset' --abbrev-commit --date=relative"11:48
lb jast: u stole !lol from schacon? ;)11:48
gitinfo jast: git config --global alias.lol "log --oneline --graph --decorate"11:48
jast one of those might help visualise it11:48
I don't know where it came from11:48
← arkiver left11:49
cbreak-work lol is old.11:49
FauxFaux lol'd, in fact.11:49
← denysonique left11:49
lb ^^11:49
beepbeep_ jast, I still think I didn't explain correctly and I did something wrong. Normally git automatically creates a commit msg for these merges, for me this wasn't the case, I wasn't allowed to commit without -m either.11:52
how do I reset the branch to a certain rev, undoing my stuff?11:53
cbreak-work git reset --hard to what ever you want11:53
you'll lose uncommitted changes as well11:53
charon beepbeep_: don't reset -- you stated that you pushed this already, so it would be !rewriting11:53
gitinfo beepbeep_: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)11:53
cbreak-work you'd have to reset in every repository11:53
charon beepbeep_: at least not simply because you changed the message from the autogenerated "Merge branch ..." to just "Merge" :-)11:53
beepbeep_ charon, I did, because when I used git commit without the -m option, it told me that I had to specify it. I thought this wasn't the case when in the "merging phase"?11:54
charon beepbeep_: as for why it didn't offer the autogenerated message: can you explain exactly what you did around the "commit" step? normally, saying 'git commit' after resolving conflicts pops up an editor that contains the pregenerated message11:54
(so that you can edit it)11:54
jast beepbeep_: what happened when you tried without -m?11:55
charon i suspect if git-commit refuses to pop up the editor, your $EDITOR is broken11:55
beepbeep_ jast, I can't scroll up that much :(11:55
jast ah, too bad11:55
→ jceb joined11:55
cbreak-work theory: people who can't scroll up far are using windows.11:55
← beautifulmind left11:55
charon 'screen' also defaults to some 1990's value for the scrollback... 100 i think11:56
cbreak-work :O11:56
that's less than a screen height11:56
jast anyway. in graph view of the history, does it show the merge commit as a merge commit? i.e. do two lines join together at that commit?11:56
cbreak-work ... well, maybe not less... bit it's little11:56
→ pretty_f_ joined11:57
← pretty_function left11:57
→ juvenal joined11:58
← OOPMan left11:58
beepbeep_ jast, the graph shows it like this:11:58
*11:58
|\11:58
next to asterix it says b2df1c7 (HEAD, origin/master, origin/HEAD, master) Merge11:59
jast okay, then it's a proper merge11:59
→ hostingh100202 joined11:59
← tgummere1 left12:00
beepbeep_ Still think this is weird for some reason. Darn it. At least in svn I was able to noob around and be fine, get the feeling you either have to be a git wizard or nothing.12:00
→ tgummerer joined12:01
← hostingh100202 left12:01
→ alexfu joined12:01
→ arkiver joined12:01
← arkiver left12:02
_ikke_ beepbeep_: svn has a very limited model. I's a bit easier to grasp, but it's also gives you less freedom12:03
beepbeep_: But you don't have to be a git wizzard to operate it12:03
→ SkAzZ joined12:03
→ arkiver joined12:04
→ OOPMan joined12:04
nyx hi12:04
how do I check whether there's anything to pull from github?12:04
git status does not tell me that12:04
nevyn git fetch12:04
charon !pull12:04
gitinfo pull=fetch+merge (or with flags/config also fetch+rebase). It is thus *not* the opposite of push in any sense. A good article that explains the difference between fetch and pull: http://longair.net/blog/2009/04/16/git-fetch-and-merge/12:04
charon so fetch.12:04
nyx git remote show origin does tell me, but is that the correct way12:05
← arkiver left12:05
alexfu +nyx: you could always do a git log and compare commits with the github repo.12:05
← asakura left12:06
alexfu is it possible to have Git rename a file if the file being pulled in is different from the local?12:06
→ arkiver joined12:06
FauxFaux alexfu: man git checkout --theirs kind of helps.12:07
gitinfo alexfu: the git-checkout manpage is available at http://git-scm.com/docs/git-checkout.html12:07
nyx is there something like git remote show origin to show if my local repo is up to date? git fetch downloads the changes, git log doesnt give me a summary12:08
← arkiver left12:08
→ asakura joined12:08
FauxFaux nyx: The only thing you can do to the remote except fetch is ls-remote; your repo is behind if it's not exactly what ls-remote says.12:08
→ arkiver joined12:08
C8H10N4O2 to create a server set up of localdev(localhost) - testing (live server) - production (live server) do you just create the repo at the product use git clone from the testing server then do the same locally? So i push to testing then log in there and pusht o development?12:08
→ iamjarvo joined12:09
FauxFaux nyx: (i.e. fetch is: ls-remote && download new stuff)12:09
C8H10N4O2: !deploy12:09
gitinfo C8H10N4O2: Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html12:09
nyx FauxFaux: Ok good to know! Do people ususllly just do a git pull to check? or git fetch12:09
I'm doing git pull both to check and to merge in any updates12:10
FauxFaux I would just use git pull and rollback if anything went wrong, but if you prefer to see what might happen instead of what's actually going to happen, you're welcome to use fetch + guess.12:10
(Many people do.)12:10
→ macIvy joined12:11
charon what guessing is there? you just 'gitk HEAD @{u}' on an alias12:11
er, gitk HEAD...@{u}12:11
nyx I see! How would I check for any new updates on all branches that I'm currently tracking?12:11
charon git branch -vv12:11
FauxFaux nyx: git fetch && git branch -avv12:11
charon: That's the commits that are going to merge, not the results of the merge (or rebase).12:11
charon FauxFaux: oh, you meant the conflicts. i see12:12
← OOPMan left12:12
← robbyoconnor left12:12
FauxFaux Well, even in the case where it doesn't conflict, the resulting tree is more interesting than the commits in most cases, imo.12:12
→ jacobat joined12:12
charon i'd state the exact opposite ;)12:13
but i guess it depends on how big the tree is12:13
→ skyf joined12:13
jacobat I think I have some objects in my packs that are no longer referenced from anywhere, but "git gc --aggressive --prune=now" seems to leave them anyways - can this be true?12:13
_ikke_ jacobat: Why do you think they're not referenced anymore?12:13
nyx FauxFaux: What is the 2nd v option flag in git branch -avv12:13
_ikke_ nyx: 2nd verbose level12:14
nyx oh cool12:14
jacobat _ikke_: Because the repository has been filtered with git filter-branch by my coworker and his repository is was smaller than mine now12:14
nyx thnx!12:14
charon jacobat: there's a checklist in man git-filter-branch...12:14
gitinfo jacobat: the git-filter-branch manpage is available at http://git-scm.com/docs/git-filter-branch.html12:14
_ikke_ jacobat: Well, you can have different references then your coworker, resulting in different objects still being referenced12:15
← macIvy left12:15
→ flaguy48 joined12:15
jacobat _ikke_: I've deleted all tags and branches except master though :/12:15
→ pretty_function joined12:17
← pretty_f_ left12:17
→ toothrot joined12:17
→ OOPMan joined12:18
← _oakdog8 left12:20
→ archdori joined12:21
→ Ryback_ joined12:21
← Ryback_ left12:21
→ Ryback_ joined12:21
← LekeFly left12:24
→ SyouJyoA joined12:24
→ shruggar joined12:24
← specialGuest left12:25
→ spaceonline joined12:28
← ketas left12:29
→ marhaban joined12:31
← dvide left12:32
← arkiver left12:32
← OOPMan left12:32
→ ketas joined12:33
→ revagomes joined12:33
→ tommyvyo joined12:39
→ OOPMan joined12:39
→ dvide joined12:42
← segher left12:44
→ arkiver joined12:44
→ Ded joined12:45
→ segher joined12:45
← dsirijus left12:47
← marhaban left12:47
→ solderion joined12:47
← pretty_function left12:47
→ pretty_f_ joined12:47
← segher left12:48
→ segher joined12:49
→ rohan32 joined12:49
→ nwest joined12:51
← C8H10N4O2 left12:52
→ freeed joined12:52
→ tinti joined12:53
→ hellopat joined12:55
← utdemir left12:56
→ subbyyy joined12:57
← subbyyy left12:57
→ LekeFly joined12:58
← SyouJyoA left13:00
← OOPMan left13:01
← arkiver left13:01
→ snearch joined13:02
← und3f left13:02
← segher left13:04
→ yoklov joined13:04
alexfu how would I handle files that only need to be tracked once to initialize the remote repository? then when clients clone, those same files are to be ignored.13:05
→ segher joined13:05
_ikke_ !config13:05
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.13:05
_ikke_ alexfu: What do you mean to initialize the remote repository?13:05
← spaceonline left13:06
← leeN left13:06
→ OOPMan joined13:06
← solderion left13:06
→ brentratliff joined13:07
← pretty_f_ left13:07
→ pinguinux joined13:07
← pinguinux left13:07
→ pinguinux joined13:07
→ pretty_function joined13:07
← ph^ left13:08
→ Octalot joined13:08
alexfu _ikke_: lets say I have a project and this project has a config file which is dependent on the persons computer and dependent on how they have their project setup. well, I dont want to track this config file but I do want to have a "skeleton" config file on the remote repo so that when clients clone, they can change it but not have it tracked.13:08
← cooloney left13:09
_ikke_ alexfu: Well, don't add the actual file, but add a template13:09
→ linoj joined13:09
_ikke_ alexfu: Let users copy the template to the actual file13:09
alexfu: You should not have environmental settings in git13:10
← afuentes left13:10
alexfu _ikke_: thats a good point. not sure why I haven't thought of it.13:10
→ Chibby joined13:11
→ Davey joined13:11
← pretty_function left13:12
← Ded left13:13
→ glennpratt joined13:13
→ Textmode joined13:14
← api left13:15
→ arvind_khadri joined13:15
→ mrwalker joined13:17
→ polymar joined13:18
← alexukf left13:19
→ dv310p3r joined13:19
← Lgb left13:20
← palm_vue left13:20
← bizhanMona left13:20
→ untrusted joined13:20
← arietis left13:21
→ C8H10N4O2 joined13:21
→ pppaul joined13:21
→ ZyX-I joined13:21
← _Vi left13:22
→ Vile joined13:23
gitinfo set mode: +v13:23
→ ddelrio1986 joined13:23
→ Lgb joined13:23
→ palm_vue joined13:23
untrusted sill a git noob but i find it hard to understand which actions might be bad for a public repo others might base work upon - 2 related questions: is there a convention naming branches which might cause trouble? is there a way to tell git to stop me if i want to do such an action?13:23
→ bizhanMona joined13:23
→ _Vi joined13:23
→ intransit joined13:23
← fusion271 left13:24
→ gefd joined13:24
→ fusion271 joined13:24
_ikke_ untrusted: 1: There is no convention, 2: Git does this by default, you'd have to add extra parameters to override it13:24
← fusion271 left13:24
← Vile left13:25
← hendry_ left13:25
→ kevwil joined13:25
→ Vile joined13:25
untrusted _ikke_: so 2 would be solved by not using which parameter?13:25
→ kurrataW joined13:26
_ikke_ untrusted: --hard, -f13:26
but those are not the only ones13:26
untrusted _ikke_: and git rebase doesn't need an option?13:26
← barq left13:27
→ ivan\ joined13:27
_ikke_ untrusted: You can rebase locally all you want. Git protests as soon as you want to push13:27
→ barq joined13:28
← hellopat left13:28
untrusted _ikke_: yes but i have to take care then13:28
→ Lord_DeathMatch joined13:28
untrusted _ikke_: i search for an automated way so i don't have to care13:28
_ikke_: less rope please ;-)13:28
← kevwil left13:28
_ikke_ untrusted: git is not a walled garden where you can't shoot yourself in foot13:29
untrusted: But git has safety nets to get you covered in most cases13:29
untrusted _ikke_: i just wonder wether there wouldn't be an automated solution to the problem13:29
_ikke_ untrusted: Because noone felt the need to build one13:30
untrusted mark some branch public or something like that13:30
maybe i just worry too much13:30
→ TeckniX joined13:30
→ gfontenot joined13:31
untrusted but regarding the search results on the topic there seems quite some confusion out there ;-)13:31
_ikke_ Well, I believe they were busy implementing some feature which would warn you when you rewrote published commits13:31
← nwest left13:31
_ikke_ But not sure about the status13:32
untrusted ah sounds to be exactly what i am looking for13:32
_ikke_ untrusted: Yes, but like I said, that's not inside git yet13:33
untrusted _ikke_: no problem, thanks13:34
← tommyvyo left13:34
_ikke_ untrusted: But git has the reflog, which allows you to find commits that might have become unrefenced13:34
→ marhaban joined13:35
→ xMYTHICx joined13:35
← yoklov left13:35
→ tommyvyo joined13:36
→ kevwil joined13:36
→ derekpioneer joined13:37
← Speed left13:38
← marhaban left13:39
→ marhaban joined13:40
→ leeN joined13:40
→ reds joined13:43
→ Destos joined13:43
→ tonyj joined13:43
reds → Guest4600713:43
→ schemula joined13:44
← marhaban left13:44
← OOPMan left13:45
← pppaul left13:45
rohan32 → rohan32|afk13:46
→ AaronMT joined13:46
→ ehsan joined13:46
← pinguinux left13:47
← Guest46007 left13:47
→ fly9 joined13:48
→ etcetera joined13:49
→ OOPMan joined13:49
← Destos left13:49
← ashirus left13:50
← etcetera left13:50
→ Dimitr joined13:51
← kurrataW left13:52
→ agile joined13:52
← martinjlowm left13:52
← kermit left13:52
→ kermit joined13:53
→ axl__ joined13:53
← nor5trom left13:53
→ jackh joined13:55
← jackh left13:55
← axl_ left13:55
axl__ → axl_13:55
→ dzonder joined13:55
→ clocKwize joined13:56
→ TheLZA joined13:56
clocKwize hi guys, i've put some stuff in a branch that I don't want there, how do I create a new branch, with those commits in it, and the remove them from the main branch?13:56
← Guest86474 left13:57
FauxFaux clocKwize: !fixup13:57
gitinfo clocKwize: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!13:57
→ oakdog8 joined13:57
→ nor5trom joined13:57
→ Dimitr|2 joined13:59
← sylr left13:59
→ orkz joined13:59
jh3 why would the git-shell-commands directory be missing after installing git? from what i've been reading it should be in /usr/share/doc/git/13:59
← OOPMan left14:00
jh3 all i have in /usr/share/doc/git are a bunch of .txt files14:00
← C8H10N4O2 left14:00
→ pppaul joined14:00
← nor5trom left14:00
→ ryez joined14:00
clocKwize FauxFaux: thanks..14:00
← Dimitr left14:02
→ specialGuest joined14:03
→ strk joined14:03
strk I created a tag named '-' by mistake, but can't delete it anymore, ideas ?14:03
tried: git tag -D \- ; git tag -D -- -14:03
→ luxurymode joined14:04
charon strk: git update-ref -d refs/tags/-14:04
strk would it be safe to rm .git/refs/tags/- ?14:04
→ C8H10N4O2 joined14:04
strk update-ref did it, thanks charon14:05
→ duckxx joined14:05
→ byte-smasher joined14:05
luxurymode my colleague created a new branch and pushed it to the remote. i can see (from our site that shows the repo) that the new branch exists on the remote, but i can't seem to pull it down and when i do 'git branch -r14:05
i don't see the repo14:05
*apologies for the accidential CR14:05
→ duckx joined14:05
→ Destos joined14:06
FauxFaux luxurymode: Did you run git remote update?14:06
luxurymode sorry git branch branch -r doesnt show me that branch14:06
→ leila joined14:06
luxurymode FauxFaux, that did it, thank you sir!14:07
→ nor5trom joined14:07
byte-smasher git noob here.... if I want to have two different folders (on the same machine) that contain different branches, am I forced to clone the entire repository and push/pull all the time? ... or can I use the same repository for both folders/branches?14:07
FauxFaux byte-smasher: Pretty much. It'll be fast and won't waste any disc space, etc. There's git-new-workdir and things like that, but they're worse than pullpush imo.14:07
← polymar left14:08
→ herdingcat joined14:08
← iamjarvo left14:08
luxurymode whats the right way to pull down a newly created branch from the remote?14:09
← duckxx left14:09
→ shruggar1 joined14:09
← kevwil left14:10
FauxFaux luxurymode: git remote update && git checkout nameofthebranch14:10
→ neurodrone joined14:10
luxurymode ah ok, i thought checkout would work then for some reason thought checkout was only for when you have the branch locally14:10
charon strk: it's safe, but it may also be in packed-refs already. update-refs is better because it also removes it from packed-refs14:10
_ikke_ FauxFaux: I don't think that works (the last command)14:10
← shruggar left14:10
grawity after `git remote update` or similar, I think you /will/ have the branch locally14:11
so should work14:11
FauxFaux _ikke_: Why not? The defaulting to git checkout -tb foo origin/foo # has been around for aaaaages.14:11
← herdingcat left14:11
byte-smasher FauxFaux: my only problem with pushing/pulling all the time is that my project is 700+mb (a news based website with archives going back to 2009... files are mostly images)... and it is kinda slow as a result.14:11
FauxFaux byte-smasher: Old history shouldn't make pull/push slow; only new history.14:12
byte-smasher k, I must be doing something wrong then14:12
clocKwize so, can I branch my branch, rollback several commits from the original branch and push it and the new branch still have those commits in it?14:12
FauxFaux git clone ../local/directory # should only be doing the same work that checkout (of an entirely new tree) does.14:12
← BadDesign left14:13
← leila left14:13
_ikke_ FauxFaux: Hmm, is it documented?14:14
m1sc clocKwize: y14:15
← jacobat left14:16
FauxFaux _ikke_: O_o I can't see it in git checkout. I swear it was.14:16
_ikke_ afaik, you have to pass in -t14:17
FauxFaux Well, it works, and has for ages.14:17
_ikke_ FauxFaux: Isn't that confusing?14:17
cmn it only works if it's not confusing14:18
FauxFaux It's what I expect. It only works in the unambigious case. I'd expect origin/foo to do a different thing, in fact.14:18
← snearch left14:19
→ macabre joined14:19
→ jjbohn joined14:19
→ geekbri joined14:20
FauxFaux _ikke_: https://github.com/git/git/blob/master/builtin/checkout.c#L823 second stanza.14:21
← TheLZA left14:21
FauxFaux <3 it's enabled by "dwim_ok".14:21
_ikke_ ok14:22
← kcj left14:22
_ikke_ what is dwim_ok?14:22
← dvide left14:23
← _Vi left14:23
grawity "do what I mean", I'd guess14:23
_ikke_ oh lol14:23
FauxFaux Yeah. i.e. if you're not already specifying branches, or in patch mode, or whatever, just do what the user means. <314:23
→ _Vi joined14:23
→ dbolser joined14:24
dbolser if a repo moves, how to 'switch'?14:24
← rchavik left14:24
_ikke_ dbolser: moves?14:24
grawity git remote set-url14:24
→ Superlokkus joined14:25
dbolser _ikke_: was a 'sandbox' project then became a real project with a new url14:25
ty14:25
grawity: how to see the current url?14:25
grawity git remote -v14:25
dbolser many thanks14:25
grawity or just check .git/config?14:25
← strk left14:25
→ zdk joined14:25
← Superlokkus left14:25
dbolser yups14:25
→ chrisgeorge joined14:26
← ReekenX left14:26
→ Dimitr joined14:26
← specialGuest left14:26
_ikke_ git git remote show <remote>14:26
← iori left14:27
← jonnu left14:28
← Dimitr|2 left14:29
← skyf left14:29
← Lord_DeathMatch left14:30
→ kevwil joined14:31
→ skyf joined14:31
← macabre left14:34
→ etcetera joined14:37
← Jay_Levitt left14:38
← etcetera left14:38
→ etcetera joined14:40
← papegaaij left14:42
→ herdingcat joined14:42
→ BadDesign joined14:44
← lb left14:44
← fission6 left14:45
← schemula left14:45
← untrusted left14:46
→ Remram joined14:46
→ jackh joined14:47
→ marhaban joined14:47
→ arturaz joined14:48
→ lacrymology joined14:49
← dbolser left14:49
→ msmithng joined14:50
→ proradii joined14:50
arturaz Hey :) Any quick way to determine last commit time for particular file in repo?14:50
← BadDesign left14:51
FauxFaux arturaz: git log --format='%ct %at' -1 -- path/to/file14:51
← marhaban left14:51
arturaz $ git log -1 --quiet options.yml14:51
pretty close :)14:51
thanks!14:52
→ _iron joined14:52
← etcetera left14:52
← tsdh left14:52
← flijten left14:53
→ tsdh joined14:53
→ ReekenX joined14:53
← tsdh left14:53
← d3vic3 left14:54
→ tsdh joined14:54
← john_anderson left14:54
→ brettnem joined14:54
brettnem hello all!14:54
gitinfo brettnem: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.14:54
brettnem I'm a little confused by the staging area.. I made a bunch of changes to a tracked file. But then used git add -i to just stage one hunk.. when I ran commit, it committed the entire file (all changes). Am I missing something? (and how do I undo it?) Thanks!14:55
→ johntron joined14:56
brettnem btw, I committed with git commit <filename> -m 'some log text'14:56
→ corburn joined14:56
_ikke_ brettnem: By specifying <filename> you say git should ignore the staging area14:57
brettnem *oh*14:57
← Guest49313 left14:57
_ikke_ brettnem: you can undo this with git reset HEAD^14:57
← johntron left14:57
→ adiabatic joined14:57
pierre__ → nowhere_man14:58
brettnem so if I use the staging area, I don't need to specify explicit paths.. I suppose that makes good sense14:58
_ikke_: thanks!14:58
_ikke_ brettnem: yw14:58
← babilen left14:58
brettnem does that reset give me my staging area back?14:58
I suspect no?14:58
_ikke_ brettnem: No, you have to readd that hunk14:58
brettnem ok, thanks14:59
_ikke_ brettnem: note you can do git add -p to add hunks immediatel14:59
immediately14:59
brettnem like right after I make the hunk?14:59
← oehman left15:00
_ikke_ instead of git add -i; <select option to add hunks>, you can do: git add -p15:00
C8H10N4O2 when creating git branches do you do it from within the git repo or from outside of it. Like I want to create a branch for a whole folder name interested but the folder itself is the git repo. So would i do "git branch interested-issue1" from www directory or just do "git branch issue1" from within interested directory15:00
brettnem I'm using git add -i for this.. work work work work.. oh changes 2-3 and 5 are all one logical commit.. stage those, commit together..15:00
→ leila joined15:00
_ikke_ C8H10N4O2: Are you comming from SVN? git branches are global to the repository. You cannot branch seperate folders15:00
C8H10N4O2 i come from svn yes lol15:01
← JPT_ left15:01
_ikke_ brettnem: Yeah, git add -p is a shortcut for git add -i; 515:01
→ JPT joined15:01
brettnem cool.. yeah the thing is, and maybe it's a problem with my coding style, sometimes I end up fixing 2 or 3 different things before i commit.. so this is a way to not having to commit things that have nothing to do with each other together..15:02
→ Nirvanko joined15:03
C8H10N4O2 I guess my other confusion come from how does my IDE know which branch I am using? Sorry for the noob questions. SVN is much more simple, though now I am wishing I was a git person from the start15:03
→ ashirus joined15:03
C8H10N4O2 I am reading pro git just not fast enough15:03
→ Alvo joined15:04
Alvo hi, how can i backup a stash?15:04
Nirvanko I've added a submodule to the project source tree. Now I would like to push those submodules so other people will see it. Is it possible?15:04
→ nwest joined15:05
→ elux joined15:05
Nirvanko I don't see them in commit.15:05
← xMYTHICx left15:05
→ maestrojed joined15:05
→ babilen joined15:06
m1sc Nirvanko: "them"? when you add a submodule, only the metadata is added in the parent repository15:07
brettnem hah.. finally cleared up months of confusion of why other hunks were automatically committed. geez I feel silly15:08
← asedeno_work left15:08
← beepbeep_ left15:09
cbreak-work git doesn't automatically commit anything15:09
→ xMYTHICx joined15:10
← jackh left15:10
← cbreak-work left15:11
← ryez left15:12
← phreaky left15:13
← dr_lepper left15:13
← corburn left15:15
→ private joined15:15
← Zeeshan_M left15:15
← ashirus left15:15
private → Guest1293515:15
← Guest12935 left15:15
→ Zeeshan_M joined15:16
← cagedwisdom left15:16
→ Xilent joined15:16
→ fission6 joined15:16
→ specialGuest joined15:17
← specialGuest left15:17
→ specialGuest joined15:17
→ asedeno_work joined15:21
→ dvide joined15:22
← neurodrone left15:22
→ cilly joined15:22
→ fisted joined15:23
← _Vi left15:23
→ apok joined15:24
→ _Vi joined15:24
→ reds joined15:24
→ ashirus joined15:24
reds → Guest8148815:25
← geekbri left15:26
← tsdh left15:26
→ SirCmpwn_ joined15:27
SirCmpwn_ → SirCmpwn15:28
→ Synthead joined15:29
← lacrymology left15:29
← C8H10N4O2 left15:30
→ corburn joined15:30
→ C8H10N4O2 joined15:32
→ Speed joined15:33
← Speed left15:33
→ Speed joined15:33
→ AlbireoX joined15:34
→ lacrymology joined15:34
← duckx left15:36
← xMYTHICx left15:39
← msmithng left15:39
← ddelrio1986 left15:40
→ thelazydeveloper joined15:40
← blami_ left15:41
→ dhruvasagar joined15:41
← AaronMT left15:42
← skorgon left15:42
← segher left15:44
→ ovnicraft joined15:44
diminoten "your branch is ahead of 'origin/develop' by 8 commits.15:45
what does that mean?15:45
git merge has never gone well for me :(15:45
← cilly left15:45
cjb it just means that you have 8 commits that you haven't pushed out yet15:45
diminoten pushed out to where15:46
→ robbyoconnor joined15:46
cmn take a look at what gitk shows you15:46
← hyperair left15:46
diminoten oh shit gitk15:46
→ segher joined15:46
diminoten eff it15:47
← ChadStudmuffin left15:47
cmn eff what_15:47
diminoten I'm deleting the branch15:47
muep git log --graph --all --decorate15:47
diminoten it's worthless15:47
cmn what branch?15:48
diminoten develop15:48
I didn't actually develop in there15:48
cmn and yet you've made 8 commits?15:48
← leila left15:48
diminoten I haven't though15:48
that was a merge15:48
→ gigamonkey joined15:48
diminoten I tried to merge master into develop15:48
cmn not much difference15:48
→ alexisabril joined15:48
cmn you brought those eight commits into the branch15:48
diminoten to catch it up15:48
yeah they're 8 commits I won't miss15:49
← C8H10N4O2 left15:49
← arvind_khadri left15:49
← robbyoconnor left15:50
cmn fine, but it has little to do with the message15:50
muep if they also are part of history of some other branch, deleting the develop branch will not actually get you rid of those eight commits15:50
diminoten cmn you're right15:50
→ robbyoconnor joined15:51
← Alvo left15:52
→ artnez joined15:53
→ neurodrone joined15:53
→ C8H10N4O2 joined15:53
← artnez left15:54
→ leila joined15:55
→ dr_lepper joined15:56
→ rmr joined15:58
← mritz_ left15:59
← leila left15:59
→ OOPMan joined16:00
← Guest81488 left16:01
← clocKwize left16:01
← OOPMan left16:01
→ yoklov joined16:02
← Vert left16:02
← mmc1 left16:03
→ talexb joined16:03
← corburn left16:04
← whitman left16:04
→ corburn joined16:05
FauxFaux sighs; using maven so version numbers are hard-coded into the build files. merge, guess what, only conflicts in the version field of the build files.16:05
→ BadDesign joined16:06
← dhruvasagar left16:06
← irqq left16:07
← centipedefarmer left16:07
→ leila joined16:07
← giallu left16:08
→ imachuchu joined16:08
← mishok13 left16:08
→ Hasbro joined16:10
← Hasbro left16:10
→ Hasbro joined16:10
← fisted left16:10
→ kukks joined16:11
→ boombatower joined16:11
← yoklov left16:12
→ mritz joined16:13
→ nglorioso joined16:13
→ ShadesofGraylin| joined16:14
← fornext left16:14
→ goshawk joined16:16
→ Fretta joined16:16
← tjadc left16:17
AAA_awright_ → AAA_awright16:17
← LekeFly left16:18
← segher left16:18
→ rehat joined16:19
→ duckxx joined16:19
← psoo left16:19
→ maletor joined16:19
→ segher joined16:20
← asedeno_work left16:20
← segher left16:20
→ asedeno_work joined16:21
→ Shusshu joined16:21
→ etcetera joined16:22
← Shusshu left16:22
→ segher joined16:22
← pppaul left16:22
← herdingcat left16:22
← apok left16:23
→ reds joined16:23
reds → Guest6298216:24
← tnzr left16:24
← corburn left16:25
→ corburn joined16:26
→ abstrusenick joined16:26
← LimeBlast left16:27
← nor5trom left16:30
← Chappo left16:30
← senny left16:32
→ TheLZA joined16:33
→ Chappo joined16:34
→ jonshea joined16:34
← Guest62982 left16:34
→ msmithng joined16:35
← AlbireoX left16:37
← ShadesofGraylin| left16:37
→ hyperair joined16:39
→ spaceonline joined16:39
← graphitemaster left16:40
← f0i left16:40
← svm_invi1tvs left16:40
→ erynofwales joined16:40
→ nveid joined16:40
← specialGuest left16:40
← etcetera left16:41
→ hwrd|work joined16:41
← msmithng left16:42
→ Sigma joined16:42
→ mishok13 joined16:42
→ btanaka joined16:43
← Chappo left16:44
→ Chappo joined16:44
→ msmithng joined16:44
→ Koolguydude joined16:44
→ codeFiend joined16:44
→ barneybook joined16:44
→ Gentlecat_ joined16:45
← ovnicraft left16:45
← Koolguydude left16:45
← Gentlecat_ left16:45
→ beneggett joined16:46
→ setmeaway joined16:46
→ Gentlecat_ joined16:46
→ dreiss joined16:47
← Gentlecat left16:47
→ wolftankk joined16:48
← dstorrs left16:48
→ Sym3tri joined16:49
← duckxx left16:49
← kermit left16:50
→ skorgon joined16:51
→ apok joined16:51
← _Qman left16:52
→ duckxx joined16:54
→ etcetera joined16:54
← abstrusenick left16:54
← angelsl left16:55
→ hndrk joined16:56
← dreiss left16:57
← ehsan left16:58
← _sundar_ left16:58
→ ehsan joined16:58
→ piotrj joined16:59
C8H10N4O2 if anyone is familiar with netbeans using git in the repo browser is the one that is bold the active one? I have two branches locally, develop and master and develop is in bold.16:59
→ Ecesis joined16:59
← Ecesis left17:00
→ iamjarvo joined17:00
← piotrj left17:00
→ centipedefarmer joined17:00
→ cmullican joined17:00
← thiago left17:01
→ btree joined17:01
← Chappo left17:02
← babilen left17:02
← RaceCondition left17:03
← erynofwales left17:05
→ msmithng_ joined17:05
← msmithng left17:05
msmithng_ → msmithng17:05
← hndrk left17:05
← tvw left17:07
← apok left17:07
← beneggett left17:08
→ apok joined17:08
← shruggar1 left17:08
→ unmanbearpig joined17:08
← Vile left17:09
→ mtytel joined17:10
→ Crocell joined17:11
Crocell hi17:11
gitinfo Crocell: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.17:11
→ _sundar_ joined17:12
→ fisted joined17:13
← robbyoconnor left17:13
Crocell i'm trying to use an svn repo as a remote source for git. i did this with git svn clone -r1234:HEAD http://some.svn.server/someproj/trunk/subproj17:13
→ mcluskydodallas joined17:13
→ robbyoconnor joined17:13
→ tydeas joined17:13
Crocell works fine so far and i can commit my changes back to svn with git svn dcommit, except for this problem:17:13
→ PetrMotejlek joined17:13
→ Teo_ joined17:13
← Teo_ left17:13
tydeas Hello ppl I have a "flow" question. I did a commit, reverted and no I want to reintroduce what is the best way to do it ? revert the revert ?17:14
nglorioso tydeas: Yes, revert the revert17:14
← skyf left17:14
→ Teo_ joined17:14
tydeas Thanks nglorioso17:15
Crocell if i modify a file (e.g. myproj.pro) and don't commit it to my local git, then the git svn dcommit fails with error myproj.pro: needs update17:15
thorbjornDX are hooks only per-clone? I would like to share a few of them on a per-project basis17:15
Crocell update-index --refresh: command returned error: 117:15
← Dimitr left17:15
imachuchu tydeas: have you pushed it to anyone else?17:16
Crocell is there no way to commit to svn only the files committed to the local git without committing the other modified files?17:16
← gusnan left17:16
imachuchu tydeas: or have they pulled from you (does anyone else know of your change yet)?17:17
cbreak Crocell: just commit all of them17:17
→ mwrosen joined17:17
cbreak Crocell: since svn sucks, you can't partially commit things17:17
there's no staging area like with git17:17
Crocell ouch17:17
cbreak commit them to a different branch17:17
→ artnez joined17:17
cbreak or use git stash for a temporary commit17:17
→ Chappo joined17:17
→ mattalexx joined17:17
cbreak afterwards, you can use git rebase to plant them on the new dcommitted commit17:18
← spaceonline left17:18
→ beepbeep_ joined17:18
Crocell those are temporary test changes that i can't commit to the global svn.17:18
→ kermit joined17:18
Crocell what's that stash thing?17:18
cbreak doesn't matter17:18
← leila left17:18
cbreak commit them in git17:18
git commits are local17:18
→ hndrk joined17:19
Crocell well sure but then how do i prevent the dcommit from committing them upwards to svn?17:19
→ dinesh6654 joined17:19
→ skyf joined17:19
cbreak if the local commits are not in the branch you dcommit, then they won't get dcommitted17:19
Bilge >what's stash17:19
>doesn't matter17:19
← robbyoconnor left17:19
Bilge Sounds legit17:19
← Chibby_ left17:20
→ dwmw2_gone joined17:20
nglorioso !stash17:20
cbreak Bilge: man git-stash, read it or stay a noob17:20
gitinfo Bilge: the git-stash manpage is available at http://git-scm.com/docs/git-stash.html17:20
→ LekeFly joined17:20
Bilge Enjoy getting preached at by a commit-everything faggot17:20
← byte-smasher left17:20
← nglorioso left17:20
ChanServ set mode: +o17:21
Bilge Oh he's going to be a big man now17:21
← Sigma left17:21
cbreak kicked Bilge (no calling other people faggots please)17:21
→ Bilge joined17:21
→ Merwin joined17:21
Bilge So only ops can call names then17:21
Pretty good channel17:21
cbreak don't act like a moron :)17:22
imachuchu Bilge: a stash is like a stack made of degenerate commits. You can push/pop/apply them and they stick around like commits, but aren't really on any branch nor have things like commit messages17:22
Bilge I know what the stash is17:22
cbreak set mode: -o17:22
← corburn left17:23
→ spaceonline joined17:23
← diegoviola left17:24
← _Vi left17:24
← neurodrone left17:25
Vampire0_ → Vampire017:25
→ _Vi joined17:25
→ AaronMT joined17:25
← Chibby left17:25
→ ddelrio1986 joined17:26
← achadwick left17:26
thorbjornDX !hooks17:26
→ Fretta_ joined17:27
← Element9 left17:27
→ Element9 joined17:27
→ tplaner_ joined17:27
← centipedefarmer left17:28
← Fretta_ left17:28
→ dstorrs joined17:29
← rurufufuss left17:29
← fisted left17:29
← Fretta left17:29
→ rvl joined17:29
→ diegoviola joined17:29
→ irqq joined17:29
← mrwalker left17:31
→ yoklov joined17:31
← yoklov left17:31
→ opus joined17:33
← tydeas left17:33
← rvl left17:34
→ jceb__ joined17:34
→ Anubisss_ joined17:34
→ efreese joined17:34
→ OOPMan joined17:34
→ pdtpatrick joined17:35
← OOPMan left17:35
→ SegFaultAX|work2 joined17:37
← tureba left17:37
→ j416 joined17:37
← jceb left17:37
← maestrojed left17:37
→ himanshu_linux joined17:38
← Anubisss left17:38
→ dinesh6654_ joined17:39
← rmr left17:39
← lacrymology left17:39
himanshu_linux hi i am new . when every i am try to push . it shows me "did you run git update-server-info on the server? " . what does it mean ?17:39
← linoj left17:39
→ thiago joined17:39
→ Sky[x] joined17:40
→ rmr joined17:40
→ corburn joined17:40
himanshu_linux its whenever.. sorry for typing mistake .17:40
→ dunpeal joined17:40
← dinesh6654 left17:40
→ OOPMan joined17:40
← codeFiend left17:40
cbreak don't push to an http remote unless you know it is valid17:41
→ bitkiller joined17:41
dunpeal Hi. I'm on Git 1.7.11, and wanted to see the datetimes for each entry in `reflog`. Why does `git reflog --verbose` return `fatal: unrecognized argument: --verbose` despite the `--verbose` flag being listed in the help message?17:41
← LekeFly left17:42
→ whitman joined17:42
← himanshu_linux left17:42
← PetrMotejlek left17:43
→ cybersphinx joined17:43
cbreak dunpeal: --verbose is for expire17:44
not for show17:44
→ lacrymology joined17:44
→ john_anderson joined17:44
dunpeal cbreak: oh, I see. is there a way to get human-readable datestimes for `show`? I can see that the log file itself does preserve them as "seconds since epoch"17:45
cbreak dunpeal: try git log -g17:45
→ LekeFly joined17:45
→ nelk joined17:45
cbreak although that seems to show the commit date and not the ref log update date17:45
← rmr left17:46
dunpeal yeah, exactly. it's a bit strange that such valuable information, while preserved, can not be readily shown.17:46
→ robbyoconnor joined17:46
→ rmr_ joined17:46
→ darren joined17:47
→ raykyri joined17:47
← rehat left17:47
→ Vile joined17:48
← alexfu left17:49
← dr_lepper left17:49
dunpeal I use the reflog to debug problems caused by faulty git usage, and datetimes are helpful in assembling a post-mortem crash report.17:49
→ codeFiend joined17:50
→ nor5trom joined17:50
→ tureba joined17:50
→ tsdh joined17:51
→ oliver_ joined17:51
→ beneggett joined17:51
← irqq left17:51
→ irqq joined17:53
→ mmc1 joined17:54
→ clocKwize joined17:54
cbreak git can crash?17:54
now that's something I've yet to see myself17:54
→ beepbeepbeep_ joined17:54
→ HyperNerdV2 joined17:54
← gigamonkey left17:54
← oliver_ left17:54
→ nifty_ joined17:55
→ oliverc joined17:56
→ dhruvasagar joined17:56
sente i have a repo with no files 'git ls-files' shows nothing, git log shows no commits, but there's a 2.1MB file: ./.git/objects/pack/tmp_pack_3Q513a17:56
← oliverc left17:56
← opus left17:56
sente this was emailed to me (a .tar.gz of the .git dir), can i unpack that file somehow?17:57
→ cakehero joined17:57
cbreak why bother?17:57
grawity try `git unpack-objects < .git/objects/pack/tmp_pack_3Q513a`17:57
though it's probably half-downloaded or something17:57
cbreak git will use pack files directly17:57
→ piotrj joined17:58
cbreak and without refs, object files are rather pointless17:58
← beepbeep_ left17:58
grawity at least you can poke around separate objects then17:58
aep_ → aep17:58
grawity (I thought git ignores tmp_pack_* normally?)17:58
cbreak he should just throw the thing away17:58
imachuchu shouldn't they show up as dangling heads to a git fsck?17:58
cbreak and get himself a working copy17:58
dangling heads?17:59
← jceb__ left17:59
cbreak what's that?17:59
← LekeFly left17:59
← Speed left17:59
imachuchu sorry, dangling things (trees/blobs/commits)18:00
→ pppaul joined18:00
cbreak maybe. but still, more trouble than it's worth18:00
easiest way is to get a working copy18:00
← dwmw2_gone left18:01
→ gigamonkey joined18:01
imachuchu cbreak: "easiest" yes, but if he can't get a clean copy18:01
← Teo_ left18:01
cbreak then he's in for lots of puzzling :)18:01
imachuchu sente: also tell your friend about the git-bundle command if they're trying to send repos around via e-mail18:02
← spyro left18:02
← Yuuhi` left18:02
→ dr_lepper joined18:02
→ dwmw2_gone joined18:02
→ codetroll joined18:03
→ cirenyc joined18:03
← btree left18:03
← BadDesign left18:04
→ gusnan joined18:04
sente hrm, i did the unpacked it but still lack what i want18:04
← Tabmow left18:05
sente he uploaded it to me with curl, the post appeared to work18:05
→ fayimora joined18:05
sente i think he just messed up making the tarball or doing something else18:05
dunpeal cbreak: it's a "crash report" in the broader sense: a user used Git incorrectly, and ended up with a branching / merge / rebase / stash snafu, usually involving his inability to recover some of his changes.18:05
← HyperNerdV2 left18:06
← lonewulf` left18:07
→ shruggar joined18:07
→ HyperNerdV2 joined18:07
→ opus joined18:07
→ Yko joined18:07
← jkremser left18:07
→ leila joined18:08
cbreak sente: if you don't have any refs (git branch -a), then you don't have any proper history18:08
gitinfo set mode: +v18:08
Yko Hey guys18:08
gitinfo Yko: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.18:08
dunpeal !backup18:08
gitinfo Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups18:08
← OOPMan left18:09
Yko I downloaded and install the package for git 1.7.11.3 but after installing, git --version still shows up as 1.7.5.4 on Mac OSX Lion. Any idea why?18:09
→ reds joined18:09
cbreak because you have xcode18:10
xcode has its own git in /usr/bin18:10
reds → Guest6948518:10
→ lonewulf` joined18:10
← clocKwize left18:10
cbreak you can change $PATH to have /usr/local/bin first18:10
← arturaz left18:10
Yko Thanks much cbreak18:10
← ReekenX left18:10
← dunpeal left18:10
→ rehat joined18:11
← AaronMT left18:11
← dwmw2_gone left18:11
sente cbreak: thanks for info, imachuchu also and grawity18:12
turns out the shit just isn't legit18:12
which i more or less suspected from the get go18:12
← skorgon left18:13
→ dwmw2_gone joined18:13
→ abletony84 joined18:15
abletony84 Hey18:15
I accidentally did a "git stash" and then "git stash clear" and notice all my work is gone. Is there any way to restore this?18:15
← corburn left18:15
→ spyro joined18:15
← spyro left18:15
→ spyro joined18:15
cbreak abletony84: :/18:16
not easily18:16
abletony84: read man git-stash18:16
gitinfo abletony84: the git-stash manpage is available at http://git-scm.com/docs/git-stash.html18:16
→ nickpresta joined18:16
cbreak there's an example at th eend18:16
"Recovering stashes that were cleared/dropped erroneously"18:17
← diegoviola left18:17
← cirenyc left18:18
← The_Ball left18:18
imachuchu abletony84: The lesson here, I think at least, is that git stash should not really be relied upon. Yes they are essentually lite-commits, but commits are so cheap anyway just commit them and rebase them into whatever you want later18:18
→ ovnicraft joined18:19
→ sq-one joined18:19
→ corburn joined18:19
← Dreadnaught left18:19
cbreak yeah. I always recommend that you just commit on a temp branch18:19
less likely to lose track of also :)18:19
← tatsuya_o left18:20
→ drizzd joined18:20
Yko @cbreak. Actually, my PATH is set to /usr/local/bin first. For some reason, my git is inside /usr/local/git/bin/git instead of /usr/local/bin.18:20
Should it be inside /usr/local/bin?18:20
cbreak :/18:20
weird18:20
grawity although git makes it quite easy to recover from `stash clear` and such stuff18:20
jast yes, it should be18:21
cbreak maybe the package is broken18:21
jast the prefix was probably set to /usr/local/git18:21
cbreak make sure you also have a /usr/local/libexec/git-core18:21
rohan32|afk → rohan3218:21
cbreak grawity: compared to normal git recovery it's hard18:21
← jjbohn left18:21
← dwmw2_gone left18:21
→ specialGuest joined18:22
Yko cbreak, I don't have /usr/local/libexec/git-core18:22
→ jjbohn joined18:22
→ dwmw2_gone joined18:22
Yko jast, is there a way that I can change the prefix?18:22
← Sebo left18:22
cbreak check this /usr/local/git thing18:22
you'll have to move everything inside it out one level18:22
jast Yko: no... it's defined at build time18:22
cbreak don't replace stuff though18:22
jast it *probably* won't work if you just move everything up one level18:23
cbreak reason?18:23
Yko Ya...hmmm...let me just uninstall git18:23
cbreak does git have hard coded paths?18:23
grawity cbreak: not really – `git fsck`, `git diff ..foo | git apply -R`, and such18:23
jast well, where do you think git gets the libexec path from?18:23
Yko and then retry it agian which is recommended but nothing seems to happen when I tried that18:23
cbreak grawity: see? harder :)18:23
grawity cbreak: well, not /very/ easy, but at least I have all the tools to do it18:23
← mcluskydodallas left18:23
cbreak normal git snafu fix is just a git checkout or git reset away18:23
→ mcluskydodallas joined18:24
cbreak thanks to the reflog18:24
grawity hmm, seems like I /can/ `git checkout foo`18:24
jast anyway, I'm going to get some food. I'll be back and stuff. :)18:24
Yko aight18:24
thanks jast18:24
→ maestrojed joined18:24
← _Vi left18:24
cbreak jast: relative path18:24
→ _Vi joined18:25
→ The_Ball joined18:25
→ andrew__ joined18:26
← leila left18:26
andrew__ → andrewhl18:26
← skyf left18:27
Yko This is soo confusing. I ran the uninstall.sh script and it removes git from /usr/local. Then, I ran the .pkg file for 1.7.11.3 and now it creates the folder again in /usr/local. The git-core is under /usr/local/git/libexec/git-core18:28
← specialGuest left18:28
→ specialGuest joined18:28
→ littlebearz joined18:28
← dpino left18:29
→ BadDesign joined18:29
← hndrk left18:29
← Nirvanko left18:30
cbreak Yko: sounds like the package is broken.18:30
Yko Hmmm18:31
Maybe I should installing it from homebrew18:31
I'll uninstall and try it from there18:31
← dwmw2_gone left18:31
cbreak I just install it from source :)18:32
→ dwmw2_gone joined18:32
← shruggar left18:33
← nveid left18:33
→ osmosis joined18:34
→ nveid joined18:34
← boombatower left18:34
← HyperNerdV2 left18:34
→ hndrk joined18:35
← The_Ball left18:36
Yko From source...good idea LOL18:36
Let me try that18:36
→ skyf joined18:36
Yko Hehe18:36
← TheLZA left18:36
→ towski joined18:36
→ fisted joined18:36
→ Zenopus_ joined18:36
Yko So, in the package README, Step 1 - Install Package ------------------------ Double-click the package in this disk image to install. This installs git to /usr/local/git. Root access is required.18:36
So, I guess it does install it to /usr/local/git18:37
Instead of /usr/local/bin/git18:37
jjbohn → jjbohn|afk18:37
→ tjholowaychuk joined18:37
jast that's a weird package you got there, then :)18:37
... and I still got to go and get food18:37
Yko LOL. It's from http://git-scm.com/downloads18:37
Sigh18:38
← Zenopus left18:38
Yko Oh well, so it's better for it to be in usr/local/bin?18:38
Zenopus_ → Zenopus18:38
← littlebearz left18:38
jast git-scm.com is the 'official' website for git, but the content is pretty much a one-man show18:39
← ddelrio1986 left18:39
← Guest69485 left18:39
→ ovnicraft_ joined18:39
cmn and the packages are done by whoever has time18:39
jast well, it doesn't matter in the end where the files are. you just need to have the right location(s) in your path.18:39
imachuchu Yko: assume they know what they are doing?18:39
→ littlebear joined18:40
jast anyway, my point about git-scm.com is that the links there are pretty much arbitrary and not officially sanctioned by 'the git team' or anything like that18:40
← dwmw2_gone left18:41
← ovnicraft left18:41
ovnicraft_ → ovnicraft18:41
← beepbeepbeep_ left18:42
cmn the only official thing in git are Junio's tarballs18:42
Yko OK18:42
→ psoo joined18:42
→ dwmw2_gone joined18:42
Yko But, where is the "best" place to put git?18:42
→ praba joined18:43
Yko Now that I kno where it is, I can put my $PATH /usr/local ahead of /usr/bin but18:43
is /usr/local/git or /usr/local/bin/git better?18:43
→ LekeFly joined18:43
← ivan\ left18:43
← ovnicraft left18:43
→ ovnicraft joined18:44
cmn put it wherever your system puts them, typically it would be /usr/local/ for user-installed software18:44
← nveid left18:45
→ Eridius joined18:45
← maestrojed left18:45
→ nveid joined18:45
→ Jon-- joined18:45
→ maestrojed joined18:46
Jon-- is there any way to lock down a user so it can only do git fetching? a restricted shell that only allows git perhaps?18:46
→ Hamled joined18:46
m1sc Jon--: !gitolite18:46
gitinfo Jon--: Want to host as many git repos (and users!) as you like, on your own server, with fine-grained access control? You want gitolite: https://github.com/sitaramc/gitolite - Documentation: http://sitaramc.github.com/gitolite/master-toc.html18:46
cbreak Jon--: why?18:47
Jon--: why bother giving that user a shell?18:47
just give him a program that fetches every time he presses return18:47
Jon-- cbreak, repository updating over SSH18:47
I want to lock it down as much as possible18:47
cbreak ?18:47
why even give a shell then?18:48
→ linguini joined18:48
cbreak you can make users via .authorized_keys that can only execute one single command18:48
Jon-- oh?18:48
I don't suppose you can allow anything with git, only one specific command on connect right?18:49
command="string" in .authorized_keys18:50
← troyt left18:50
imachuchu Jon--: have you looked at git-shell? It's a restricted shell that only allows the person to use git18:50
Jon-- I know this is more of a unix question, my apologies.18:50
imachuchu, perfect, thnx18:50
cbreak Jon--: only one command, yes.18:50
imachuchu Jon--: read it first before you say it's perfect ;)18:51
cbreak but that's enough if you want to only allow fetch18:51
← ShooterMG left18:51
← dwmw2_gone left18:51
→ ReekenX joined18:51
← tsdh left18:51
Jon-- cbreak, file permissions are only r, should block any nasty repo changes right?18:52
← changedNicks left18:52
→ changedNicks joined18:52
changedNicks → sweet|kid18:52
cbreak uhm... it's pointless to fetch if he has only read rights18:52
→ pantsman joined18:52
cbreak the whole point of fetch is to change the repository18:52
← fayimora left18:52
→ dwmw2_gone joined18:53
← gtrak left18:53
Jon-- cbreak, unset GIT DIR && git fetch origin && git reset --hard origin/master18:53
← rmr_ left18:53
Jon-- this needs +w?18:53
cbreak looks pointless18:54
and of course18:54
fetching requires write access18:54
how else would you fetch?18:54
Jon-- I'm looking to replace whatever code I have in that repo with the latest from origin/master18:54
cbreak where do you think the fetched history ends up?18:54
← dwmw2_gone left18:54
cbreak then do what I wrote above18:55
Jon-- I meant +w remotely on the repo, not locally.18:55
Of course I have +w on that user locally18:55
cbreak make a script that fetches, and only let him execute that script via ssh18:55
then all he can do is fetch18:55
← Merwin left18:56
cbreak (No clue why you would want something like that though)18:56
→ Merwin joined18:56
fission6 if i rm a file from tracking then i checkout it out, will it automaically be tracked again?18:56
if i include it in .gitignore too18:56
← adiabatic left18:57
← msmithng left18:57
cbreak fission6: depends.18:57
← etcetera left18:57
cbreak did you commit the removal?18:57
← LekeFly left18:57
Jon-- cbreak, I think I am not communicating properly. I'm on box1, box1 needs the latest production code from server1. I want to grab it over ssh using a fetch. git shouldn't be run on server1 at all, it's just hosting the repo that box1 is updating.18:57
→ diegoviola joined18:58
→ dwmw2_gone joined18:58
cbreak git has to run on a server18:58
you need git upload-pack there18:58
Jon-- It is running18:58
→ msmithng joined18:58
cbreak it gets started if you clone/fetch via ssh18:58
if you want to clone/fetch from the server, then it's easy18:58
gitolite and done.18:59
imachuchu fission6: are you trying to have a file present on the remote branch that isn't in your local checkout?18:59
cbreak or use git-shell as mentioned before18:59
Jon-- There are networking restrictions I have to work around.18:59
All I have is an SSH tunnel.18:59
cbreak no need for a tunnel18:59
Jon-- I don't believe gitolite supports ssh addressing, etc.18:59
cbreak you need an ssh connection18:59
what?18:59
gitolite REQUIRES ssh18:59
more or less :)18:59
← iamjarvo left19:00
cbreak remember the authorized_keys thing I told you before about?19:00
that's how gitolite works.19:00
→ the_cybersphinx joined19:00
abletony84 cool19:01
thanks guys19:01
Jon-- Will it allow me to specify ssh://user@host:port designation? This machine only has one connection (guaranteed anyway), and it's an SSH tunnel back to server1 in this case.19:01
← opus left19:01
cbreak sure.19:01
as long as user is the user gitolite is set up for19:02
abletony84 was mostly html and css work, luckily i had the page loaded in firebug and could restore the work from there19:02
→ etcetera joined19:02
→ ZaherDirkey joined19:02
← ZaherDirkey left19:02
jast I set up a gitolite/HTTP setup for work recently19:02
Jon-- Just seems like a pain in the ass to me19:02
Overkill for what I need19:02
cbreak gitolite? 5 mins.19:02
otherwise, read what I wrote above19:02
you should read what I wrote above either way :D19:02
← psoo left19:03
Jon-- which? You've said many things ;)19:03
cbreak either gitolite19:03
← cybersphinx left19:03
cbreak or git-shell19:03
← darren left19:03
cbreak combine git-shell with read only access if you want19:03
jast or just you a standard shell... works just as well if all the involved machines are trusted19:04
s/you/use/ jeez, how did that happen19:04
Jon-- jast, That defeats the whole purpose of hardening things, which is why I'm here. Standard user is easy :P19:04
← etcetera left19:04
jast right. git-shell then.19:04
Jon-- cbreak, chsh to git-shell, then I can do git clone/fetch calls?19:04
jast yes19:04
Jon-- I'll give it a go, thnx.19:05
jast depending on filesystem permissions19:05
← Destos left19:05
→ opus joined19:05
jjbohn|afk → jjbohn19:06
→ etcetera joined19:06
Jon-- I have 664, I'm the 4 in this case19:06
← opus left19:06
cbreak Jon--: chsh what?19:06
Jon-- (on files)19:06
cbreak you just have to set it as shell for that user19:06
Jon-- cbreak, chsh to git-shell, is that not how you restrict a user to using git-shell? [not sure what you're asking here]19:06
cbreak you can't git clone/fetch19:07
but you can do git upload-pack19:07
which is what you need for fetching19:07
Jon-- will that break if there are potential local changes that I want to discard?19:07
→ leila joined19:07
cbreak irrelevant19:07
fetching doesn't care about any of that19:08
Jon-- so I can't use fetch?19:08
cbreak it just fetches history into remote tracking branches19:08
Jon-- with git-shell?19:08
cbreak why not?19:08
of course not19:08
← maestrojed left19:08
→ shruggar joined19:08
cbreak didn't you say above that you don't want to do git fetch?19:08
← leila left19:09
Jon-- No, I want to restrict user compromise from causing damage, either commit, delete, or other arbitrary unix commands that could do some damage. I'm referring to using git-fetch on the box, ssh for auth, over to an account on the server with a git repo to grab changes from19:09
→ fayimora joined19:10
Jon-- I want to restrict the user on that server side to being able to fetch that git code, but nothing else, if possible19:10
cbreak so... answer:19:10
do you want him to run git fetch server side?19:10
or not?19:10
you've said both so far19:10
Jon-- server side, no19:10
cbreak which is of course nonsensical19:10
← corburn left19:11
cbreak alright, then don't complain!19:11
→ ddelrio1986 joined19:11
Jon-- I was unclear thought you meant I couldn't fetch on the remote repo19:11
→ oddover joined19:11
EugeneKay Jon-- - !xy19:11
gitinfo Jon--: This sounds like an "XY Problem" http://mywiki.wooledge.org/XyProblem So let's step back for a minute. What are you actually trying to achieve? Why are you doing it this way?19:11
cbreak you can't19:11
Jon-- throws hands up19:11
cbreak you can not do anything on the remote repo appart from using the pack commands19:11
so don't complain19:12
Jon-- I've been very clear what I need here.19:12
I am not complaining.19:12
cbreak no. you're contradiction yourself all the time19:12
← maletor left19:12
Jon-- < Jon--> cbreak, I think I am not communicating properly. I'm on box1, box1 needs the latest production code from server1. I want to grab it over ssh using a fetch. git shouldn't be run on server1 at all, it's just hosting the repo that box1 is updating.19:12
← ovnicraft left19:12
→ StefanM joined19:13
cbreak then you will not be able to do anthing with git...19:13
Jon-- That last bit about "git not running" means "I don't need to be able to run git commands on server1 as that user, only be able to fetch from it".19:13
cbreak as I said above: you need git on the server19:13
Jon-- cbreak, for the second time. ^^^19:13
cbreak you should read what I write :/19:13
EugeneKay "only be able to fetch from it" tells me that you're XY-ing this. What's the point?19:13
cbreak you can not do anything on the remote repo appart from using the pack commands19:13
EugeneKay In english, not in git-commands.19:14
← specialGuest left19:14
Jon-- EugeneKay, I want to be able to fetch the latest production code from that server, and restrict it so it can only fetch code and not make modifications.19:14
EugeneKay Ok, I just said "not in git-commands".19:14
!deploy19:14
gitinfo Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html19:14
EugeneKay Read that ^19:14
cbreak Jon--: so, what's the problem?19:15
why are you still unhappy?19:15
I gave you two potential solutions ages ago19:15
both will work19:15
EugeneKay It sounds like you want to be able to let $USER put code on a server. Your concept of the user logging in and then pulling it down is flawed. Read deploy.19:15
cbreak just pick one and implement it!19:15
Jon-- cbreak, besides you being an ass about it?19:15
cbreak you're the one assing around19:15
Jon-- mostly that you seem to be contradicting yourself19:15
cbreak first you say "no git command", then "fetch on the server" then "not fetch on the server"19:16
decide!19:16
Jon-- Or at least, I am not clear what you mean, and feel I'm being ridiculed for it.19:16
cbreak what I say is simple:19:16
use gitolite.19:16
→ specialGuest joined19:16
cbreak use git-shell19:16
it's either one or the other of course19:16
they don't work well together.19:16
bitkiller is there any way to replicate some repo confs (those that may make sense in replicating) from bare to clones?19:16
Jon-- EugeneKay, Close, I want to use $USER to authenticate, and grab code from there.19:16
cbreak bitkiller: not with git.19:16
jast git doesn't have authentication19:16
all it does is respect filesystem permissions19:17
← fisted left19:17
Jon-- It does when you use SSH to access the repo... that is all I mean.19:17
cbreak Jon--: use ssh for authentication.19:17
jast well, okay19:17
so, git-shell19:17
when did that stop being a viable solution?19:17
I haven't been following the discussion19:17
bitkiller cbreak, is gitolite an option?19:17
Jon-- jast, when cbreak told me I couldn't use git fetch if I restrict $USER to git-shell19:17
cbreak he keeps complaining about wanting to fetch on the server...19:18
→ arietis joined19:18
jast sure you can19:18
cbreak Jon--: I told you three times that there's NO POINT IN GIT FETCH19:18
EugeneKay cbreak - have a !beer !19:18
gitinfo cbreak: Beer! It's what's for dinner!19:18
Jon-- cbreak, for the tenth fucking time, NOT ON THE SERVER.19:18
cbreak you don't need it19:18
EugeneKay Jon-- - !fsck19:18
cbreak Jon--: get some manners19:18
gitinfo Jon--: [!profanity] Hey! This is a family-safe channel, so keep it frakking clean you fierfekker!19:18
cbreak Jon--: and get some reading glasses19:18
jast cbreak: I'm not following19:18
blathijs Is it me, or does Jon-- just wants to fetch _from_ the server, not _on_ the server?19:18
jast blathijs: it's not just you19:18
cbreak Jon--: I told you that git shell only restricts what happens on the server19:18
jast that's precisely my interpretation, too19:18
cbreak blathijs: that's why I told him many times that he doesn't have to do that19:19
blathijs and perhaps Jon-- doesn't realize that fetching _from_ the server involves running git commands _on_ the server (underwater)19:19
Jon-- blathijs, Correct. I came in here for help locking it down for security.19:19
jast all the locking down pretty much happens on the ssh side of things19:19
→ iamjarvo joined19:19
jast (and making sure the ssh user doesn't have filesystem access to repositories he shouldn't19:19
)19:19
Jon-- I want to minimize the permissions on $USER so if that machine is compromised there is minimal damage to the other side.19:20
← palm_vue left19:20
← bizhanMona left19:20
← Lgb left19:20
joshc Jon--: what you want is git-shell.19:20
Jon-- joshc, Alright19:20
jast so, you have to make sure that the user can read but not write19:20
and for that you need to figure out appropriate file permissions19:20
cbreak as I told you many times already Jon-- use git-shell or better gitolite19:20
Jon-- jast, I have that already, was looking for more, eg a shell that can only run git commands... ala git-shell19:20
jast (or use gitolite which does all the work for you)19:20
Jon-- Then cbreak made me think that wasn't a viable solution19:20
cbreak no19:20
jast git-shell is fine19:20
cbreak when did I do that?19:21
jast I'm pretty sure that was just a misunderstanding19:21
cbreak I told you many times it is the solution19:21
Jon-- cbreak, You're either a troll or are terrible at communication Sir.19:21
Thanks for the help guys19:21
← kevwil left19:21
cbreak what?19:21
← marthinal left19:21
cbreak you keep repeating the same things over and over and call me a troll?19:21
→ jceb__ joined19:21
jast well, to be fair, someone who accuses people of being trolls doesn't qualify as "excellent communicator" in my book, either ;)19:21
cbreak people get ruder every day :(19:21
→ psoo joined19:22
→ hellopat joined19:22
jast ah, I read the backscroll now and I understand the disconnect19:22
blathijs fwiw, here cbreak implied git-shell wouldn't work:19:22
21:07:04 < cbreak> you can't git clone/fetch19:22
→ Sigma joined19:22
cbreak blathijs: no...19:22
blathijs but there's really just a lot of confusion betweeon _on_ the server and _from_ the server19:23
cbreak that's the whole point of git-shell19:23
jast it's because cbreak was explaining things from an internals perspective whereas Jon-- was interpreting them from a user's perspective19:23
→ Lgb joined19:23
← shreya left19:23
→ palm_vue joined19:23
blathijs cbreak: I said "implied" for a reason :-)19:23
→ oddover_ joined19:23
cbreak blathijs: it's quite unambiguous19:23
→ bizhanMona joined19:23
← freeed left19:23
cbreak blathijs: you can not do anything on the server19:23
apart from upload-pack and a few other pack things19:23
→ afuentes joined19:23
jast the *user* typically doesn't know/care what happens on the server when you fetch/push from/to it19:23
cbreak you can NOT fetch on the server19:23
→ kevwil joined19:23
cbreak you can NOT clone on the server19:23
→ private joined19:23
Jon-- blathijs, That was what I was referring to, yes.19:23
cbreak and so on19:23
jast: Jon-- seemed to care about it19:24
gitinfo set mode: +v19:24
← oddover_ left19:24
oddover Hi all. is there some way to do a `git ls-remote` and only get a list of tags back? (no sha1)19:24
cbreak he kept asking about fetching on the server despite me telling him that it's pointless19:24
jast but of course cbreak is right in that git-shell will not allow users to run git fetch and friends remotely on the server19:24
private → Guest6814619:24
blathijs cbreak: Yes, I understand how it works, but "git clone [email@hidden.address] could be described by running a clone on [email@hidden.address]19:24
jast no, I don't think he said "on the server"19:24
← kevwil left19:25
← _Vi left19:25
cbreak blathijs: no...19:25
jast oddover: no, but the format is pretty easy to parse/process, isn't it?19:25
cbreak blathijs: clones are local19:25
you can not clone on a remote19:25
→ _Vi joined19:25
oddover +jast: yes, but I'm having trouble using it inside of another language. I figured I'd ask to make sure19:26
jast clones are an operation with two end points. it makes little sense to pin them to one of the sides exclusively19:26
→ maletor joined19:26
blathijs cbreak: Yes, I understand how it works, but "git clone [email@hidden.address] could be described as running a clone on [email@hidden.address] by users who don't fully understand how things work19:26
cmn oddover: you can restrict it by passing refs/tags as a parameter19:26
jast the 'clone' git builtin itself runs on only one side, but the overall clone operation is distributed across (typically) two machines19:26
cmn but parsing which side is which is up to the user19:26
cbreak blathijs: no, it can't19:26
jast and this is where user perspective disconnects from developer perspective19:27
blathijs cbreak: I'm not saying they would be correct, of course19:27
jast the user simply doesn't care what it's called on the server side19:27
cbreak runnong a clone on foo#bar.com would be ssh [email@hidden.address] git clone19:27
cmn blathijs: if you don't know how things work, you can describe anything as anything, it doesn't make it right19:27
blathijs cbreak: I'm just saying there's a fair chance they'll doing it anyway19:27
→ freeed joined19:27
→ scwizard joined19:27
cbreak blathijs: if you don't assume correctness, then you could describe it as recalibrating the dilitium matrix.19:27
_ikke_ :D19:28
scwizard what's the most sensible way to undo the most recent commit non interactively?19:28
→ Corey joined19:28
← luxurymode left19:28
cbreak scwizard: git reset --soft HEAD~119:28
_ikke_ scwizard: describe non interactively? Through a script?19:28
jast scwizard: define 'undo'19:28
(see !revert)19:28
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [git checkout -- .]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?19:28
← rohan32 left19:28
→ leila joined19:29
jast does anyone know why it suggests the 'checkout .' instead of 'reset --hard'?19:29
← oddover left19:30
cbreak I don't think checkout even works19:30
also, --hard is bad in this context19:30
jast sure does, but doesn't necessarily do it completely19:30
cbreak as I said above, --soft should do19:30
jast I was talking about (b)19:30
--soft is useless for (b)19:30
cbreak discard uncommitted changes...19:30
scwizard a) in this case. If you git-revert through a script it won't prompt the user?19:30
cbreak I use checkout -p for that19:31
imachuchu scwizard: why do you want to do this in a script?19:31
jast scwizard: if it's the most recent commit, nope. in some other cases it may cause conflicts.19:31
cbreak scwizard: revert needs no additional infos unless it fails19:31
since it applies a patch, it can fail19:31
jast cbreak: yeah, -p is nice, but sometimes reset --hard is simply much easier/faster :)19:31
blathijs jast: Perhaps because git checkout can be applied to individual files as well, and reset --hard is always for the whole tree?19:31
← msmithng left19:31
→ msmithng joined19:32
← tweedledumber left19:32
jast blathijs: yes, but on the other hand checkout -- . doesn't clear away newly tracked files, whereas reset --hard does19:32
→ tweedledumber joined19:32
jast (afaik)19:32
→ tellone joined19:33
← ska-fan left19:33
blathijs jast: I think none of them removes them from your WC19:33
→ ska-fan joined19:33
← ska-fan left19:33
cbreak the main problem with checkout . is that it doesn't affect the whole repository19:33
→ ivan\ joined19:33
blathijs jast: And checkout reverts changes back to the state in the index, not HEAD19:33
cbreak people always think that git add ., git checkout . and so on magically affect the whole repo19:34
jast unless you happen to be in the right place :)19:34
cbreak but . means current directory19:34
jast blathijs: yes. but (b) is about removing all *uncommitted* changes, which includes staged ones.19:34
I guess that means that technically the answer is wrong19:34
← etcetera left19:35
scwizard when I say " git revert HEAD~1" it gives an error, but still changes my working copy19:35
jast I've changed it19:35
← dstorrs left19:36
blathijs !revert19:36
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [no turning back: git reset --hard]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history] ?19:36
jast scwizard: that's nice. are we supposed to guess the error? :P19:36
cbreak scwizard: resolve the conflict an commit19:36
scwizard jast: it says "error: could not reverse $revisionnumber $revisiondescription"19:36
blathijs jast: Might make sense to include an option d that shows git checkout, since git checkout is actually very useful for "reverting" changes (especially for people coming from SVN)19:37
jast let's see if we can fit in another one19:37
← diegoviola left19:37
→ dstorrs joined19:37
jast !revert19:38
gitinfo "Revert" is a heavily overloaded term. Do you mean: a) make a commit that "undoes" the effects of an earlier commit [man git-revert]; b) discard the uncommitted changes in the working tree [no turning back: git reset --hard]; c) undo committing [git reset --soft HEAD^, but type !rewriting_public_history]; d) restore staged versions of files [git checkout -p]?19:38
→ AaronMT_ joined19:38
→ ska-fan joined19:38
→ maestrojed joined19:38
← maestrojed left19:38
scwizard I don't understand what kind of conflicts "git revert head~1" could encounter19:38
→ corburn joined19:39
blathijs jast: Looks good19:39
← abletony84 left19:39
← ReekenX left19:39
← tellone left19:39
blathijs scwizard: if HEAD removes a file that is changed in HEAD~1?19:39
cbreak scwizard: it has to undo the second last commit19:39
→ AdrienBrault joined19:39
blathijs or changes it in a conflicting way19:39
jast nah19:39
cbreak and if you changed something in the last commit19:39
that touched the second last commit's changes19:40
then you will get a conflict19:40
← specialGuest left19:40
→ specialGuest joined19:40
← BadDesign left19:40
jast did you actually mean HEAD~1? you said "most recent commit" earlier, but HEAD~1 is actually the commit before that19:40
scwizard jast: oh, what's the most recent commit?19:40
blathijs HEAD :-)19:40
_ikke_ HEAD19:40
jast man gitrevisions if you want to understand all this weird syntax19:41
gitinfo the gitrevisions manpage is available at http://git-scm.com/docs/gitrevisions.html19:41
scwizard wait head is the current commit. If I want to "undo" in the fashion of a) then I want to go back to the commit before HEAD right?19:41
jast yes, but 'git revert' does it differently19:41
'git revert foo' means: undo the effects of foo19:41
blathijs Oh nice, didn't know man gitrevisions existed :-)19:41
gitinfo the gitrevisions manpage is available at http://git-scm.com/docs/gitrevisions.html19:41
scwizard jast: ahh, I see19:41
→ rohan32 joined19:41
→ duckx joined19:42
jast that's the only way it makes sense for 'git revert'... otherwise you couldn't do surgical removal of a single change way back in the history or stuff like that19:42
→ Adys_ joined19:42
→ The_Ball joined19:43
→ neurodrone joined19:43
← AdrienBrault left19:43
jast (of course 'removal' technically isn't the right word)19:43
← Adys left19:43
← duckxx left19:44
Adys_ → Adys19:45
→ LekeFly joined19:45
→ etcetera joined19:48
← Sky[x] left19:50
← piotrj left19:53
→ piotrj joined19:53
← piotrj left19:55
→ piotrj joined19:55
← duckx left19:57
→ smarter joined19:58
→ duckx joined19:58
→ Araxia joined19:59
EugeneKay `git revert` is really `git cherry-pick` in reverse19:59
← smarter_ left19:59
→ rcu joined20:01
scwizard is there a way a utility for quickly checking if the code of a remote folder matches up with the code in a git repository?20:03
← irqq left20:04
scwizard *local git repo20:04
FauxFaux A remote folder? As in, over ssh?20:04
CareBear\ git ?20:04
scwizard FauxFaux: yeah over ssh20:04
CareBear\ just use git20:04
← Sigma left20:04
scwizard CareBear\: how do I do such a thing in git?20:04
CareBear\ copy the files or use sshfs20:05
scwizard CareBear\: wouldn't git only work if the remote folder is a git repository. Copying the files is destructive, I don't want to do this destructively20:05
CareBear\ what destructive?20:05
cmn copying files is only destructive if you make it so20:05
scwizard well maybe destructive is the wrong word20:05
cmn copy them somewhere empty20:05
scwizard I don't want to write to the remote server20:05
CareBear\ right20:05
you don't have to20:05
scwizard in fact lets say I don't even have access to20:05
CareBear\ if you don't have access to the files how can you compare them?20:06
cmn if you don't have access to the files, you can't compare them20:06
scwizard I can read certain files from the remote server20:06
over sftp or ssh20:06
CareBear\ then you can compare those files20:06
← hellopat left20:06
scwizard how?20:06
copying them from the remote server to the local machine is slow20:06
cmn scp remoteserver:files/ .20:06
CareBear\ either copy them to your workdir with uncommitted changes20:07
*without* uncommitted changes sorry20:07
cmn and how do you suggest comparing them if you don't want to read them?20:07
scwizard cmn: you can read files without copying them over right?20:07
cmn no20:07
CareBear\ ?!20:07
wtf20:07
scwizard : how would you do that?20:07
scwizard hmm20:08
cmn you are able not to write down the data on your disc, but you need to transfer the information20:08
use rsync to speed up the process20:08
scwizard good point20:08
← nveid left20:09
→ eletuchy joined20:10
→ nveid joined20:10
← jceb__ left20:10
→ giallu joined20:11
← tjholowaychuk left20:12
→ yoklov joined20:12
← etcetera left20:13
→ jay joined20:13
← duckx left20:13
jay → Guest9975120:13
← corburn left20:14
→ luxurymode joined20:15
→ etcetera joined20:15
← Crocell left20:15
← raykyri left20:15
→ duckx joined20:15
← gfontenot left20:16
← nifty_ left20:17
→ raykyri joined20:17
← Guest99751 left20:18
→ corburn joined20:19
→ gdogg joined20:19
← dc5ala left20:19
→ gfontenot joined20:19
→ Speed joined20:20
← Speed left20:20
→ Speed joined20:20
gitinfo set mode: +v20:20
gdogg Git Q. I have an android repo (a collection of a bunch of git repos) and I want to bisect, but I can't use git bisect. I saw a snippet to do something similar, just need to keep track of the date myself: git checkout `git rev-list -n 1 --before="2009-07-27 13:37" master` Problem is that with repo we aren't on a branch at all, and rev-list doesn't work without a branch specified20:20
→ d3vic3 joined20:21
d3vic3 stat20:21
gdogg I could manually create a master branch for each repo, but I am hoping there's a more elegant solution20:21
← ruskie left20:22
→ avpx joined20:23
→ tjholowaychuk joined20:23
← tjholowaychuk left20:23
← unmanbearpig left20:24
gdogg ah I guess I can just specify HEAD instead of master20:24
→ dinesh6654 joined20:25
→ iliketurtles joined20:25
← Fullmoon left20:27
← skyf left20:28
CareBear\ gdogg : how is your repo a bunch of git repos without branches and why can't you use plain bisect?20:28
← dinesh6654_ left20:29
→ skyjumper joined20:29
← duckx left20:29
→ ruskie joined20:29
→ ferdna joined20:29
← Yko left20:30
scwizard Is there a way to have local contain all history. But remote only include history after a certain revision? You want to keep track of changes to local and remote and eventually merge them. However you don't want remote to have all that history.20:30
← revagomes left20:30
cmn no20:30
scwizard darn, that would be nice to have :(20:31
gdogg CareBear\: They have branches but when you use repo sync it just gets the TOT/head of each git repo, but deletes all branch names locally. You can't use plain bisect because between each git repo changes aren't distributed evenly so you'20:31
cmn a limited version may or may not work using shallow clones20:31
→ duckx joined20:31
gdogg ..you'd have to check each repo manually and find the halfway point of all of the changes aggregated together20:31
cmn but they weren't meant to be used for sharing anything20:31
scwizard shallow clones are almost exactly what I'm looking for20:32
except I want to be able to push to the shallow clone.20:32
CareBear\ scwizard : you're doinitrong20:33
scwizard CareBear\: how so?20:33
← spaceonline left20:33
CareBear\ scwizard : change the workflow - use the tool better20:33
→ tylersanders joined20:34
← specialGuest left20:34
← tylersanders left20:35
imachuchu scwizard: why don't you want remote to have all of that history (an extension of the question CareBear\ asked)?20:36
scwizard CareBear\: it's like this. I want to track changes to code in production, but I can't have all that history stored in production because it could potentially take up space20:36
← whitman left20:37
CareBear\ code changes do not take much space20:37
if your repo is all binaries it's a different story, but you said code20:37
your chief problem is that code changes are happening in production20:37
that's the doinitrong20:38
→ opus joined20:38
← msmithng left20:38
scwizard yes I know, yes it's dumb, but I can just tell its going to bug people who are not me20:38
← fusion27 left20:38
scwizard as for code changes in production, how to do a hotfix?20:39
CareBear\ clone fix commit push20:39
→ msmithng joined20:39
cmn there is no need to have changes in production20:39
scwizard cmn: how to hotfix production?20:39
CareBear\ I JUST SAID20:40
clone fix commit push20:40
cmn that depends completely on what you're saying20:40
scwizard CareBear\: yes, that's fine and well, but then devs need to be trained to do that20:40
cmn what you're doing, I mean20:40
CareBear\ um20:40
yes if you give monkeys sharp tools they will bleed20:40
cmn if you can't train your devs to use a VCS in the most basic way, you need to fire them20:40
CareBear\ what cmn said20:40
← oakdog8 left20:40
scwizard I don't make staffing decisions, also we're working on it20:41
CareBear\ if you take pity, you could let them bleed once, give them a band-aid, and see if they learned something20:41
→ specialGuest joined20:41
CareBear\ git is a great tool20:41
but it can not compensate for incompetent staff20:42
sorry about that20:42
→ revagomes joined20:43
→ adamm joined20:43
SaCu CareBear\, your 'if you take pity...' is worth to be put into an email signature :)20:43
CareBear\ scwizard : http://joemaller.com/990/ would probably be a step up from what you do now20:44
SaCu : feel free to use it20:44
SaCu hehe20:44
← fly9 left20:44
scwizard CareBear\: neato20:44
(anything would be a step up from anything now :( )20:44
← flaguy48 left20:45
scwizard anyways I can't believe I was worrying about the space thing, that was a silly thing to think about >_>20:45
CareBear\ you want to avoid me doing a recursive download of your .git/ though20:45
because who knows what kind of passwords the monkeys commited two years ago, and you can bet that they haven't been changed20:46
scwizard well of course :|20:47
→ darren joined20:47
→ fly9 joined20:48
scwizard I'm just glad, that monkeys aren't just a thing I deal with, or else you wouldn't have known to say that >_>20:48
thorbjornDX is there a way to filter through my entire python repo history and convert leading tabs to 4-spaces? or should I make a single commit that does that20:48
← gdogg left20:48
← d3vic3 left20:48
thorbjornDX I can pretty safely edit published history20:48
← mtytel left20:49
← lacrymology left20:49
→ hellopat joined20:49
joshc thorbjornDX: see man git-filter-branch20:49
gitinfo thorbjornDX: the git-filter-branch manpage is available at http://git-scm.com/docs/git-filter-branch.html20:49
CareBear\ if you have many commits it will be SLOW20:50
but it is mighty! :)20:50
thorbjornDX CareBear\: it's about 100 commits total? something like that20:50
← mwrosen left20:50
CareBear\ oh, no problem20:50
thorbjornDX CareBear\: though, I have to be pretty damn careful with my regex, since it's python...20:50
CareBear\ well, never a problem, just takes time20:50
← specialGuest left20:50
CareBear\ thorbjornDX : oh that's no problem, because you also have a test suite for your code - right?20:51
mohaha20:51
← The_Ball left20:51
thorbjornDX CareBear\: wishful thinking with this repo20:51
CareBear\ never has tests20:51
← raykyri left20:51
CareBear\ maybe you can make one20:51
thorbjornDX I try to use TDD when I can. But usually I just need to bang something out20:51
← etcetera left20:51
CareBear\ nothing comprehensive, but at least import everything so that there are no exceptions20:52
thorbjornDX or I inherit ~5000 lines of gui code and don't feel like testing it20:52
→ RaceCondition joined20:52
thorbjornDX CareBear\: will this solve my problem?: http://stackoverflow.com/questions/2316677/can-git-automatically-switch-between-spaces-and-tabs20:53
→ raykyri joined20:53
← AaronMT_ left20:53
thorbjornDX CareBear\: I'd rather the source be clean too, so I'd need a filter-branch too20:53
imachuchu thorbjornDX: single commit. Yes it's possible, but what really do you save? The worst that happens is that git blame doesn't help to much20:53
thorbjornDX: ack, disregard that, I was looking through history (that was for the "go back in time to change line endings question)20:54
→ Destos joined20:54
thorbjornDX imachuchu: ah, ok. For leading tabs I should still filter-branch you think?20:54
CareBear\ thorbjornDX : that will solve your problem in the future20:55
← freeed left20:55
→ gtrak joined20:55
← beneggett left20:55
CareBear\ thorbjornDX : question is how important it is for you to make it look like you knew about the trick before the repo was created20:55
imachuchu thorbjornDX: the past is the past, what do you hope to get out of it? A slightly more readable reference?20:55
thorbjornDX imachuchu: I'd prefer if I diff HEAD to something in the past that it not highlight every single line20:56
→ etcetera joined20:56
→ dinesh6654_ joined20:57
← fly9 left20:57
→ fly9 joined20:57
→ velikasha joined20:57
→ The_Ball joined20:57
→ spaceonline joined20:57
← robbyoconnor left20:58
← dinesh6654 left20:58
← nveid left20:59
← sq-one left20:59
← velikasha left20:59
→ nveid joined20:59
→ orkz_ joined20:59
→ NomNomNom joined21:00
CareBear\ then you need to filter-branch21:00
→ palithius joined21:00
palithius Is there a formal way to backup git repositories? Or do I just copy the folder and maybe zip it and store that somewhere?21:01
← axl_ left21:01
imachuchu palithius: !backup I prefer git-bundle myself (doesn't backup unreference commits though)21:01
gitinfo palithius: Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups21:01
← orkz left21:01
orkz_ → orkz21:01
CareBear\ palithius : git push --mirror into a bare rpo does a good job21:01
← ddelrio1986 left21:01
→ axl_ joined21:01
← axl_ left21:02
← nveid left21:02
← Gentlecat_ left21:02
palithius thanks!21:02
← pppaul left21:02
→ Gentlecat_ joined21:02
← tommyvyo left21:02
← efreese left21:03
→ nifty_ joined21:03
→ AaronMT joined21:03
← RaceCondition left21:04
→ dinesh6654 joined21:05
→ axl_ joined21:05
← cannonball left21:05
→ pppaul joined21:05
→ eletuchy_ joined21:06
← dinesh6654_ left21:07
→ linoj joined21:07
← fission6 left21:07
thorbjornDX alright, I'm doing it in one commit, damn the torpedos, thanks for the input CareBear\, imachuchu21:07
← kermit left21:07
→ GameGamer43 joined21:08
→ specialGuest joined21:08
← specialGuest left21:09
→ specialGuest joined21:09
← eletuchy left21:09
← codeFiend left21:10
← rwestphal left21:10
← hwrd|work left21:10
← jjbohn left21:11
← oojacoboo left21:11
← pppaul left21:12
← SkAzZ left21:12
C8H10N4O2 when working on a local environment and you want to see how the changes of a branch effect site how do you get your local website to reflect the currently selected branch? I made changes to a branch and it is still showing the master view? I have the branch active21:13
← ashirus left21:13
← _iron left21:14
← nifty_ left21:14
cmn that depends completely on how you create the site and how you view it21:14
→ queso joined21:14
← NomNomNom left21:14
→ oojacoboo joined21:14
queso I want to see what commits exist on one branch that do not exist on another. What's the best way to do this?21:14
← rohan32 left21:15
scwizard can you pull from a shallow clone?21:17
← codeshah left21:17
→ tjadc joined21:17
← mmc1 left21:17
cmn often21:17
→ thomas_b joined21:18
→ pppaul joined21:18
← tplaner_ left21:18
cmn you mean pull from as in fetch from that repo or run the pull command in that repo?21:18
← iamjarvo left21:19
→ hwrd|work joined21:19
→ api joined21:20
cmn actually, nvm, the answer is always going to be "it depends"21:20
→ fission6 joined21:21
← skyjumper left21:21
← praba left21:21
scwizard well better question would be. Lets say you do a shallow clone of repo A. You commit a few changes to your clone. Then the maintainers of A want to merge your commits into A.21:21
← The_Ball left21:22
← AaronMT left21:22
scwizard how would the maintainers of A do that?21:22
← s0ber left21:22
avpx queso: See the manpage for rev-list21:22
← todd_dsm left21:22
→ bubbles10101 joined21:23
queso avpx: Thank you.21:23
→ s0ber joined21:23
← pppaul left21:24
→ revele joined21:24
← nickpresta left21:25
avpx queso: You're welcome. You can specify revisions that way with a lot of other utils as well.21:25
→ kcj joined21:25
← _Vi left21:25
→ _Vi joined21:26
→ efreese joined21:26
← dwmw2_gone left21:26
← pantsman left21:26
← olaf_ left21:26
queso avpx: Nice, I'll check it out.21:27
← psoo left21:28
→ ijuma joined21:28
← elux left21:29
→ The_Ball joined21:29
← bgerber left21:29
← fission6 left21:29
← mritz left21:30
← j416 left21:30
→ kihap joined21:30
← queso left21:31
→ dwmw2_gone joined21:31
← msmithng left21:31
kihap I put "sites/default/settings.php" in my .gitignore file but git still identifies the file as modified. I also tried putting the file name in the excludes file in the .git folder but git still identifies it as modified. How can I get git to ignore this file?21:31
→ ddelrio1986 joined21:31
→ HyperNerdV2 joined21:32
← spaceonline left21:32
← rehat left21:32
cbreak Kiall: just git rm it21:32
kihap: you can't ignore and track a file at the same time obviously21:33
kihap: also read !config21:33
gitinfo kihap: [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.21:33
→ msmithng joined21:34
→ lemonjelly joined21:35
← revele left21:35
→ ShadeTornado joined21:36
→ iamjarvo joined21:36
← bitkiller left21:36
→ chrisgeorge_ joined21:36
→ bgerber joined21:38
→ Superlokkus joined21:38
← dwmw2_gone left21:38
→ ChadStudmuffin joined21:38
← chrisgeorge left21:39
← hndrk left21:40
scwizard How do you convert a fork to a branch? Lets say Alice the developer clones the git repo at revision X and does his own work on things. Bob the maintainer wants to download Alice's changes and make them into a branch off of A named "Alice's branch"21:40
EugeneKay I think that could use a rewording21:40
← luxurymode left21:40
scwizard How do you convert a fork to a branch? Lets say Alice the developer clones the git repo at revision X and does her own work on things. Bob the maintainer wants to download Alice's changes and make them into a branch off of X named "Alice's branch"21:41
← lemonjelly left21:41
→ dwmw2_gone joined21:41
← axl_ left21:41
cmn those are completely different things21:41
scwizard (mangled the first try at that question)21:41
← corburn left21:41
EugeneKay scwizard - is it a github Fork?21:41
cmn a repo contains branches21:41
scwizard EugeneKay: no github isn't involved at all21:41
→ rehat joined21:42
→ lustucru joined21:42
cmn you can't convert the one into another, just as you can't transform a bucket into grains of salt21:42
→ axl_ joined21:42
→ WolWid joined21:42
EugeneKay scwizard - Bob would add Alice's repo, fetch it, then create a local branch from Alice's branch.21:42
(add as a remote)21:42
→ skorgon joined21:42
scwizard yes that is what I want, how do you do that EugeneKay?21:43
→ w41_ joined21:43
EugeneKay scwizard - man git-remote; man git-fetch ;-)21:43
gitinfo scwizard: the git-remote manpage is available at http://git-scm.com/docs/git-remote.html21:43
scwizard: the git-fetch manpage is available at http://git-scm.com/docs/git-fetch.html21:43
EugeneKay See also the !book chapter on remotes21:43
gitinfo There are several good books available about git; 'Pro Git' is probably the best: http://git-scm.com/book but also look at !bottomup !cs !gcs !designers !gitt !vcbe and !parable21:43
→ HG` joined21:43
← dstorrs left21:43
→ aspotashev joined21:43
← bartman left21:44
← chrisgeorge_ left21:44
scwizard EugeneKay: remote and fetch explain the first two steps. What about the third "create a local branch from from Alice's branch"21:44
after you do a fetch, you have this "FETCH HEAD" thing or something21:44
how to turn that into a branch off a particular revision?21:44
← w41 left21:44
EugeneKay Use the three-word fetch, not !fetch421:44
gitinfo [!fetchfour] Never use the four-word version of git-fetch or git-pull (e.g. git fetch remote refspec). It always ends in tears. Yes, if you understand the implications of FETCH_HEAD it can technically be done, but really it is easier to just fetch the whole remote (or perhaps edit the fetchspec if you never want other bits). If you must, see !fetch4why21:44
← raykyri left21:44
EugeneKay Use the third syntax in man git-checkout to create a new branch tracking Alice/Foo21:45
gitinfo the git-checkout manpage is available at http://git-scm.com/docs/git-checkout.html21:45
→ raykyri joined21:45
→ esc joined21:45
← esc left21:45
→ esc joined21:45
→ dstorrs joined21:45
scwizard so git checkout "Alice's branch" X creates a new branch named Alice's branch with X as its starting point21:46
← rcu left21:46
← juvenal left21:47
scwizard so basically you'd checkout that branch, then git pull alice's remote21:47
EugeneKay `git checkout -b AliceBranch Alice/Branch`21:47
This sets the default push/pull location for AliceBranch as Alice. You will probably want to change this to origin.21:47
.trigger_edit configfiles Best practice is to store an example config file and then copy+modify it when you install your app. If your config file must be tracked by git to deploy(eg, Heroku), see https://gist.github.com/1423106 for some workarounds21:48
gitinfo EugeneKay: Okay.21:48
scwizard EugeneKay: I'm confused. In this case the head when Alice cloned was X. So you want the starting point for alice's branch to be X right?21:48
← dwmw2_gone left21:48
EugeneKay scwizard - no; Bob is creating a local branch based upon Alice's remote branch.21:49
→ chrisgeorge joined21:49
scwizard yes, and the starting point of that remote branch is X right?21:49
EugeneKay Yes, but Bob's repo doesn't care about that.21:49
scwizard wait erm21:49
bleh I typoed21:50
well didn't21:50
but21:50
← gfontenot left21:50
scwizard the starting point of the remote branch is X. So Bob wants to make the start point of the local branch X also21:50
EugeneKay git doesn't care where a branch forks from, only where it is.21:50
scwizard otherwise he'd have to do a merge right?21:50
← raykyri left21:50
→ TheLZA joined21:50
EugeneKay He will, when he merges AliceBranch back into X21:50
← dv310p3r left21:51
EugeneKay !tryit ;-)21:51
gitinfo [!tias] Try it and see™. You learn much more by experimentation than by asking without having even tried. If in doubt, make backups before you experiment (see !backup). http://sitaramc.github.com/1-basic-usage/tias.html may help with git-specific TIAS.21:51
scwizard I don't understand. If you don't change anything locally, then you do "git pull" then you don't have to merge right?21:51
→ dwmw2_gone joined21:51
EugeneKay git pull is fetch && merge.21:51
scwizard ok so what I mean when I say "you don't have to merge" is "you don't have to merge manually"21:52
EugeneKay If you haven't created any new commits then a fast-forward merge happens. There's no new commit created, but it's still a "merge"21:52
scwizard I see21:52
← Element9 left21:52
→ raykyri joined21:53
← hellopat left21:55
← TheLZA left21:57
← HG` left21:58
HyperNerdV2 → conversationkill21:58
← Chappo left21:59
conversationkill → HyperNerdV221:59
→ Chappo joined21:59
← dwmw2_gone left21:59
→ graphitemaster joined22:00
→ thursday joined22:00
→ oakdog8 joined22:01
→ alanp_ joined22:01
→ dwmw2_gone joined22:01
← thursday left22:01
← Jon-- left22:01
← opus left22:02
→ opus joined22:03
→ juvenal joined22:03
← bubbles10101 left22:04
→ resmo_ joined22:04
→ jaclinuxhelp joined22:04
← Destos left22:04
esc what does '+' mean in a gitolite config22:04
jaclinuxhelp i'm trying to add a submodule with22:04
← alanp left22:04
jaclinuxhelp git submodule add gitosis@myserver:library.git library22:04
but ... it doesnt works22:04
but, when i do22:05
git submodule add gitosis@myserver:library.git PRUEBA22:05
it works!22:05
← eletuchy_ left22:05
jaclinuxhelp any advice? is "library" a "reserved word" in git ?22:05
→ eletuchy joined22:06
scwizard bleh the phone lights just went up to 2 :(22:06
I hope they don'te ver go up to three22:06
→ mutex joined22:06
imachuchu scwizard: I have no idea what you are talking about22:07
scwizard wrong channel :|22:07
imachuchu scwizard: do they go all the way to eleven?22:07
m1sc esc: + == allow non-fast-forward updates22:07
← resmo__ left22:08
→ kermit joined22:08
← dwmw2_gone left22:08
← Hasbro left22:08
→ ShooterMG joined22:09
← leeN left22:09
oakdog8 → oakdog8_22:10
→ nglorioso joined22:10
microcolonel http://magnifi.ca/photos/2012-08-04_12-13-23_36.jpg22:10
esc m1sc: so like in refspec22:10
→ oakdog8 joined22:10
esc m1sc: thanks22:10
← TeckniX left22:12
← revagomes left22:12
→ bubbles10101 joined22:13
→ dwmw2_gone joined22:13
→ codeFiend joined22:13
→ tinti_ joined22:13
← raykyri left22:14
→ cagedwisdom joined22:15
esc whats the deal again with debugging the ssh used by clone?22:15
→ schemula joined22:16
scwizard how can you merge changes made in a remote that was shallowly cloned if you can't fetch a shallowly cloned repo?22:16
→ cybersphinx joined22:16
← chrisgeorge left22:17
→ mcluskyd_ joined22:17
→ raykyri joined22:17
← WolWid left22:18
← mcluskydodallas left22:18
← vibhav left22:18
→ chrisgeorge joined22:18
← the_cybersphinx left22:18
→ nveid joined22:18
→ Shuro joined22:19
← dwmw2_gone left22:19
→ guide joined22:19
← netj left22:20
→ basepi joined22:20
→ _br_ joined22:20
→ dwmw2_gone joined22:20
← iamjarvo left22:21
← Guest68146 left22:21
← rehat left22:22
← shruggar left22:22
berndj scwizard, are you looking for a solution based on a particular flow, or would something based on git bundle work for you?22:22
imachuchu scwizard: the general concensous on that is "don't make changes on a shallow clone". Either convert it to a full clone, or pull out the patches and apply them on a full repo22:22
scwizard imachuchu: how do you pull out the patches? Patches are generally pulled out with a fetch and merge22:23
← Sym3tri left22:23
← nveid left22:24
berndj scwizard: or with format-patch. or bundles.22:24
→ nveid joined22:24
→ Destos joined22:24
berndj depends on what workflow constraints you need to observe22:25
← ShadeTornado left22:25
imachuchu berndj: bundle won't work (I know, I know, I'm normally the first person to suggest it) unless you want to create it from a full repo and then use that as a repo replacement (but then just make a full repo)22:25
scwizard what are the technical reasons a fetch and merge doesn't work with shallow clones?22:25
EugeneKay Because shallow clones are the children of Lucifer22:26
→ ShadeTornado joined22:26
berndj imachuchu, yeah, i wondered after i wrote that if the bundle would try to include one mega-commit to represent the entire repo to the point where the shallow clone begins22:26
→ aselder joined22:26
← _Vi left22:26
← Destos left22:26
berndj or maybe something else than that22:26
← dwmw2_gone left22:26
→ _Vi joined22:27
imachuchu scwizard: diff and patch (actually don't even touch git except to get the diff from the previous commit)22:27
← ppppaul left22:28
→ bogwonch joined22:28
imachuchu scwizard: http://www.linuxtutorialblog.com/post/introduction-using-diff-and-patch-tutorial22:28
scwizard EugeneKay: thinking about it, it should be theoretically possible for git to merge commits from a remote shallow clone22:28
← opus left22:28
→ Contexual joined22:28
→ btree joined22:28
scwizard of course there's that gap between theoretically possible and implemented22:28
→ lemonjelly joined22:29
berndj it's theoretically possible for CVS to have atomic commits :-/22:29
→ AaronMT joined22:29
← raykyri left22:29
← towski left22:29
→ opus joined22:30
scwizard hehe22:30
→ maestrojed joined22:30
→ raykyri joined22:30
→ btree_ joined22:30
→ dwmw2_gone joined22:30
← btree_ left22:30
← Contexual left22:30
← ku left22:30
SwK anyone have a good example of a post update hook script to push a specific branch to another repo?22:31
→ ku joined22:31
imachuchu SwK: just looking at one actually, http://joemaller.com/990/22:31
← raykyri left22:31
SwK sweet thanks imachuchu22:32
← dzonder left22:32
imachuchu SwK: well read it first, it seems really small (the page is about a specific workflow that just has a hook in it)22:32
→ Hunner joined22:32
Hunner Hi. Is it possible to use ansi escape codes to add color to git hooks?22:33
canton7 SwK, !deploy has one22:33
gitinfo SwK: Git is not a deployment tool. You can build one around it for simple environments. http://sitaramc.github.com/the-list-and-irc/deploy.html22:33
← btree left22:33
← nglorioso left22:33
Hunner When I add the \033[36m etc. to my post-receive, it works when run by hand but shows the literal characters when run remotely22:33
SwK this is not for deployment its to make users lifes easier22:33
→ vibhav joined22:33
vibhav → Guest6945522:34
← specialGuest left22:35
SwK I'm the stable maintainer for FreeSWITCH, we have stable rev branches that only get bug/security fixes so it diverges from origin/master ... instead of making them have to switch to a branch and get everything they dont want, I just want to push the current stable maintained branch to a sep repo thats read only22:35
imachuchu SwK: (I think he meant it as an example only, not to say you were doing it wrong)22:35
→ ashirus joined22:35
SwK imachuchu: yeah some people try to do some stupid things w/ SCMs22:35
← opus left22:36
← dwmw2_gone left22:38
canton7 yeah, I was using that as a link to the page which contains the hook I'm referring to, only22:38
SwK ahh ok22:38
← timthelion left22:38
→ madewokherd joined22:38
← gerard0 left22:39
→ dwmw2_gone joined22:40
→ lindenle joined22:40
gitinfo set mode: +v22:40
lindenle What does ^{} mean in a git tag from ls-remote?22:41
cmn it's the target22:41
← aselder left22:41
grawity the commit object which the tag is pointing at, see man git rev-parse22:41
gitinfo the git-rev-parse manpage is available at http://git-scm.com/docs/git-rev-parse.html22:42
→ gerard0 joined22:42
← maestrojed left22:42
→ maestrojed joined22:42
← fayimora left22:42
← AaronMT left22:43
→ chriskon149 joined22:44
→ flaguy48 joined22:45
→ Sym3tri joined22:45
→ netj joined22:45
→ rehat joined22:45
← wolftankk left22:46
→ eigenburak joined22:47
← dwmw2_gone left22:48
← eigenburak left22:48
← skorgon left22:48
← bogwonch left22:48
→ bogwonch joined22:49
→ dwmw2_gone joined22:50
→ Jay_Levitt joined22:50
← nveid left22:51
← bogwonch left22:53
→ TheLZA joined22:53
→ chrisf|walrus joined22:54
→ nveid joined22:55
← andrewhl left22:56
← api left22:56
← lustucru left22:57
→ Lord_DeathMatch joined22:57
→ ashirus_ joined22:57
← dwmw2_gone left22:57
← john_anderson left22:58
← ashirus left22:58
→ specialGuest joined22:58
← gusnan left22:59
← Milossh left22:59
← maletor left22:59
→ dwmw2_gone joined22:59
→ unmanbearpig joined23:01
→ jez9999 joined23:01
→ maletor joined23:01
jez9999 ok for some reason, "git tag --contains" isn't working as I'd expect23:01
← unmanbearpig left23:02
→ simesy joined23:02
jez9999 I downloaded the Mono repo and I'm trying out "git tag --contains" for some commits that should DEFINITELY be contained in some tags, but i am getting no results23:02
any idea what i could be doing wrong?23:02
s/downloaded/cloned/23:02
→ opus joined23:03
→ lustucru joined23:03
← lustucru left23:03
→ e0je joined23:04
gitinfo set mode: +v23:04
← leila left23:04
→ iamjarvo joined23:04
jez9999 ok for some reason, "git tag --contains" isn't working as I'd expect23:05
I downloaded the Mono repo and I'm trying out "git tag --contains" for some commits that should DEFINITELY be contained in some tags, but i am getting no results23:05
any idea what i could be doing wrong?23:05
s/downloaded/cloned/23:05
← chrisgeorge left23:05
→ Xizor joined23:06
→ Destos joined23:06
← Synthead left23:06
← e0je left23:07
→ e0je joined23:07
← chriskon149 left23:08
← dwmw2_gone left23:09
← e0je left23:09
← Gizmokid2005 left23:09
← maestrojed left23:09
← axl_ left23:09
→ e0je joined23:09
← adamm left23:10
← cybersphinx left23:10
← linoj left23:11
← ShadeTornado left23:12
→ dwmw2_gone joined23:12
→ raykyri joined23:12
← raykyri left23:13
→ ShadeTornado joined23:13
← ashirus_ left23:14
← Araxia left23:16
← berserkr left23:16
→ maestrojed joined23:17
← specialGuest left23:17
→ specialGuest joined23:17
← dwmw2_gone left23:18
← alanp_ left23:18
← etcetera left23:19
→ shruggar joined23:19
→ dwmw2_gone joined23:20
← iamjarvo left23:21
→ chriskon149 joined23:22
→ joules joined23:22
→ Esa joined23:22
← russt left23:23
joules hi, I have question. How are backups delt with? Just another repo?23:24
← eletuchy left23:24
→ john_anderson joined23:24
joules gitinfo23:24
sorry23:24
→ ashirus joined23:24
← opus left23:25
← chriskon149 left23:25
joules so a git clone is just another repo? can one start developing with the clone?23:25
← SirCmpwn left23:26
← Jay_Levitt left23:26
→ opus joined23:26
→ Jay_Levitt joined23:26
← _Vi left23:26
← Anders_J left23:26
→ etcetera joined23:26
→ _Vi joined23:27
← shruggar left23:27
imachuchu joules: !backup and yes a clone is just another repo except with some extra remote things set up to easily push/pull from the original repo23:27
gitinfo joules: Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups23:27
→ praba joined23:28
imachuchu joules: as for backups I like bundles, but they aren't for everything23:28
→ russt joined23:28
joules ok cool thanks.23:28
→ highway900 joined23:28
← dwmw2_gone left23:29
→ dwmw2_gone joined23:29
← joules left23:30
→ eletuchy joined23:31
→ codeshah joined23:31
→ _W_ joined23:31
← nveid left23:33
→ nveid joined23:33
→ alanp joined23:33
← specialGuest left23:34
← zackiv31 left23:35
← lemonjelly left23:35
← drizzd left23:35
← scwizard left23:36
← orkz left23:36
← s0undt3ch_ left23:36
← mattalexx left23:37
→ RaceCondition joined23:37
→ leila joined23:38
→ bartman joined23:38
← dwmw2_gone left23:38
← nwest left23:38
→ dwmw2_gone joined23:40
← ehsan left23:40
→ fisted joined23:42
← leila left23:42
→ Psycho_pr joined23:43
← Psycho_pr2 left23:43
← maletor left23:44
← eletuchy left23:44
→ maletor joined23:45
← dr_lepper left23:45
→ ISF joined23:46
← alanp left23:47
← etcetera left23:47
← dwmw2_gone left23:47
→ s0undt3ch joined23:48
→ dwmw2_gone joined23:50
→ barrbrain joined23:51
→ etcetera joined23:51
→ lustucru joined23:52
← alexisabril left23:52
← ChadStudmuffin left23:53
← ShadeTornado left23:54
→ ShadeTornado joined23:55
← codetroll left23:56
← palithius left23:56
→ Sky[x] joined23:56
← giallu left23:57
← efreese left23:57
← Lord_DeathMatch left23:58
← arietis left23:58
← thiago left23:58
← dwmw2_gone left23:58
→ dwmw2_gone joined23:59
← imachuchu left23:59
→ whomp_ joined23:59

Logs Search ←Prev date Next date→ Channels Documentation