IRCloggy #git 2013-05-31

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.

2013-05-31

← noxd left00:00
← kyledr left00:00
→ simesy joined00:00
← braisarias left00:00
← acharlieh left00:00
← Cykey left00:01
← amcsi left00:02
← flaguy48 left00:03
→ zedsix joined00:03
← Z6 left00:04
← sondr3 left00:05
eric_roque anyone know why pcre lib be in git?00:06
thiago in git's git?00:08
← hyperair left00:08
← davidchubbs left00:08
→ Milossh joined00:09
eric_roque git source code00:09
thiago I don't see it included in Git's source code00:10
Git uses pcre00:10
are you asking why git uses pcre?00:10
eric_roque thiago, yes00:11
thiago because Git requires a regular express engine00:11
→ boombatower joined00:12
← LuxuryMode left00:12
← refried_ left00:12
→ acharlieh joined00:12
← FernandoBasso left00:12
eric_roque thiago, but he is the interpreter git bash?00:13
thiago repeat the question, please00:14
rephrase, I mean00:15
← zequence left00:15
eric_roque thiago, exists include pcre.h in file grep.h00:15
thiago I know00:15
Git uses PCRE for its regular express engine00:15
→ zequence joined00:16
← boombatower left00:16
← DonRichie left00:18
→ DonRichie joined00:18
→ infynyxx joined00:19
← oriba_ left00:19
eric_roque thiago, revision.h within the file add.c includes grep.h, which connection with command add?00:20
← Thra11 left00:20
thiago I don't know00:20
→ LuxuryMode joined00:21
eric_roque thiago, then how you understand?00:21
→ Cykey joined00:21
thiago understand what?00:21
← LucasTizma left00:22
eric_roque thiago, the source code00:22
thiago I don't00:22
→ refried_ joined00:22
thiago I'm not a Git developer00:22
I am just a Git user00:22
eric_roque thiago, ok00:23
thiago, I want undestand the source code git00:24
→ ColoradoPods joined00:24
eric_roque thiago, I want to contribute00:24
← Chillance left00:24
EugeneKay We're a user support channel00:25
If you want to dev on it, I suggest just playing with it.00:25
Or looking at the mailing list00:25
→ Davey joined00:26
eric_roque EugeneKay, you may have some developer in channel00:27
EugeneKay, or no?00:27
EugeneKay If we do, they're not paying attention.00:27
!tryit !spoon00:27
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://gitolite.com/1-basic-usage/tias.html may help with git-specific TIAS.00:27
[!read] Don't expect everything to be spoon fed to you - the man pages and other documents do not bite, you need to spend some time and effort reading them.00:27
gtmanfred eric_roque: where are you seeing include grep.h in add.c ?00:28
EugeneKay Go forth and play ;-)00:28
← looper left00:28
← etcetera left00:29
eric_roque EugeneKay, tanks00:29
gtmanfred eric_roque: https://github.com/git/git/blob/master/builtin/add.c00:30
i don't see grep.h00:30
eric_roque gtmanfred, no no, add.c include revision.h which include grep00:30
← ryanisnan left00:30
→ Gooder` joined00:31
gtmanfred read the revision.h00:32
eric_roque gtmanfred, you know the reason00:32
gtmanfred they commented the grep_opt struct00:32
→ etcetera joined00:32
gtmanfred /* Filter by commit log message */00:32
struct grep_optgrep_filter;00:32
in revision.h00:32
→ guns joined00:32
→ Droolio joined00:32
← LuxuryMode left00:32
gtmanfred so it is for log.c00:33
← Milossh left00:33
→ roto joined00:33
gtmanfred add.c just also uses revision00:33
→ nbouscal joined00:33
← yshh left00:33
← infynyxx left00:33
← Gooder left00:35
eric_roque gtmanfred, the pcre.h00:36
gtmanfred you can go read teh source for that one00:36
→ parhs joined00:36
← parhs left00:36
→ parhs joined00:36
eric_roque gtmanfred, is used at what time?00:36
gtmanfred no idea00:37
go look00:37
eric_roque gtmanfred, tanks00:37
→ flaguy48 joined00:37
gtmanfred GREP_PATTERN_TYPE_PCRE00:37
just look for PCRE in there and match it with /usr/include/pcre.h00:37
→ Milossh joined00:38
→ davidchubbs joined00:39
eric_roque gtmanfred, i dont understand00:39
gtmanfred then you probably shouldn't be trying to develope00:39
→ rickruby joined00:39
gtmanfred go learn to read c first00:40
→ jexmex joined00:40
jexmex anybody have any tips when dealing with conflicts with submodules during a rebase?00:40
← tvw left00:40
eric_roque gtmanfred, english is no my native language00:41
← savid left00:41
thiago nor mine00:41
→ scottmey joined00:41
← scottmey left00:42
→ scottmey joined00:42
eric_roque thiago, where are you from?00:42
← etcetera left00:42
→ magwitch324 joined00:43
→ etcetera joined00:44
← aendrew left00:45
← dreiss left00:45
← SirCmpwn left00:45
← refried_ left00:46
← magwitch324 left00:47
→ trungqtran joined00:47
→ breck7 joined00:47
→ mag1 joined00:48
← davidchubbs left00:48
← Sigma left00:48
→ rakl joined00:49
→ LuxuryMode joined00:49
← Tecan left00:50
→ lresende joined00:51
← DonRichie left00:52
→ mckendricks joined00:53
→ SirCmpwn joined00:53
← mag1 left00:53
← andref_dias left00:53
→ infynyxx joined00:53
→ pastime joined00:53
← jaequery left00:53
SeoZ-work[AWAY] → SeoZ00:54
→ dorkmafia joined00:55
→ konundra joined00:55
← trungqtran left00:56
→ tmcmahon joined00:57
→ yshh joined00:58
→ stephCoue joined00:58
→ moop7 joined00:59
← breck7 left00:59
moop7 does git rebase --skip effectively delete the conflicting commit from the tree?00:59
→ japandroid joined01:01
← ColKurtz left01:01
← Chehri left01:02
← mixel left01:02
→ falood joined01:03
→ _BJfreeman joined01:04
→ k_sze[work] joined01:04
← stephCoue left01:05
← BSaboia left01:05
→ FreezingCold joined01:06
_BJfreeman → BJfreeman01:06
→ hyperair joined01:06
→ dhruvasagar joined01:07
← simesy left01:07
→ trungqtran joined01:08
← etcetera left01:09
← tmcmahon left01:10
→ plumpkin joined01:10
→ dsdeiz joined01:10
← chamunks left01:11
→ chamunks joined01:12
← trungqtran left01:12
→ allaire joined01:13
→ bmac joined01:16
→ savid joined01:16
← GarethAdams left01:17
← acharlieh left01:17
→ ryez joined01:17
← w0m_ left01:17
→ trungqtran joined01:18
← infynyxx left01:18
← modInfo3 left01:18
→ modInfo3 joined01:19
← trungqtran left01:19
← oso96_2000 left01:20
→ Goplat joined01:20
→ w0m joined01:21
→ jaequery joined01:22
← lresende left01:22
→ nathanr joined01:22
← squentin left01:22
→ squentin joined01:23
→ mmitchell joined01:24
← jdunck left01:25
→ jceb_ joined01:25
→ w09x joined01:26
→ mixel joined01:26
→ codeful joined01:27
← w0m left01:27
→ bandroid joined01:27
→ w0m joined01:27
→ oso96_2000 joined01:28
← bmac left01:29
← jceb left01:29
→ carlos-huang joined01:29
← trollwork left01:30
→ tjbiddle_ joined01:30
→ erkules_ joined01:30
← mikewintermute left01:31
→ thebetterbjasd joined01:31
← wazyl left01:31
→ Eremiell joined01:32
← codeful left01:33
← erkules left01:33
← tjbiddle left01:34
therealklanni OK, so to cherry-pick a merge you need to use -m and provide the mainline number. Where do you get that information from?01:34
tjbiddle_ → tjbiddle01:34
therealklanni Also, when cherry picking a merge that replays all the commits from that merge, is that right?01:35
← carlos-huang left01:35
← Vert left01:35
← pr0ton left01:36
→ carlos-huang joined01:36
bandroid → bandroidx01:37
→ acharlieh joined01:37
← rudak left01:37
← scottmey left01:37
→ JCii8_ joined01:37
→ dreiss joined01:38
← acharlieh left01:38
→ simesy joined01:39
← Cykey left01:39
← xxtjaxx_ left01:39
← ndim left01:40
→ intripoon joined01:40
→ acharlieh joined01:40
← hide_ left01:41
→ hide joined01:41
← donmichelangelo left01:41
← JCii8_ left01:41
→ daito joined01:42
→ donmichelangelo joined01:42
→ JCii joined01:42
← modInfo3 left01:42
JCii I have done something to my git repository that I don't know how to fix. I was working in branch A, made sub-branch B. Worked in B. merged back to A. When I git pull from the origin (which is behind me) to local A, the merge tries to undo my branch B work that I merged into A. I'm lost even researching the problem--there are so many git commands to learn about.01:42
← dreiss left01:43
→ modInfo3 joined01:43
JCii any suggestions for learning the repository structure would be appreciated01:43
← intripoon_ left01:43
→ ndim joined01:44
→ oojacoboo joined01:44
→ jgoss joined01:45
ojacobson JCii: !repro !lol if you could01:45
gitinfo JCii: Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session -- or, even better for complex issues, design a minimal case in which your problem can be reproduced, and share it with us. This will help immensely with troubleshooting.01:45
JCii: A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all01:45
→ trollwork joined01:45
← japandroid left01:46
JCii didn't know abotu the --decorate option will try01:46
→ zangxiangyu joined01:48
← DonVitoCorleone left01:48
← AWinterman left01:48
→ xxtjaxx joined01:49
← xxtjaxx left01:49
→ xxtjaxx joined01:49
→ Z6 joined01:50
← zedsix left01:50
← akurilin left01:50
→ btree joined01:51
← btree left01:51
← cilly left01:51
← carlos-huang left01:51
→ cilly joined01:51
shentino Can I have incoming git pushes authenticated by ssh keys, and if so, how?01:53
→ carlos-huang joined01:53
frogonwheels shentino: !gitolite01:54
gitinfo shentino: Gitolite is a tool to host git repos on a server. It features fine-grained access control, custom hooks, and can be installed without root. Download: https://github.com/sitaramc/gitolite Docs: http://gitolite.com/gitolite/01:54
shentino shakes, is asking about native support in git itself01:54
frogonwheels shentino: You can sign with keys ...01:54
(gpg keys)01:55
ojacobson git doesn't do authentication at all, at any point01:55
shentino oja: thanks01:55
ojacobson so adding that would be a pretty major undertaking01:55
→ dreiss joined01:55
frogonwheels shentino: There are many ways around that depending on what you want to do.01:55
ojacobson indeed01:56
frogonwheels shentino: ... well to varying degrees there are01:56
ojacobson git doesn't *interfere* with authentication much either01:56
!xy01:56
gitinfo 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?01:56
→ luc_ joined01:56
frogonwheels ojacobson: yeah, I nearly reached for that myself ;)01:56
← moop7 left01:56
← rickruby left01:56
shentino I'll be getting some z's01:56
frogonwheels ojacobson: btw, it helps if you prefix it with the person you're addressing it to.01:56
shentino: seriously, explain what you're trying to achieve, and it'll be easier to help01:57
shentino I know :P01:58
→ lresende joined01:58
shentino I have this happen, but just as often others make preemptive assumptions that I'm not being straightforward and flat out refuse to answer the question I actually pose01:58
and what I was trying to achieve was literally have git standalone handle it. bjacob's telling me that git doesn't do authentication like that was helpful in getting me to realize the futility01:59
only then was I prepared for answers such as gitolite01:59
← devoldmx3 left01:59
→ devoldmx joined02:00
ojacobson shentino: tab completion, man, tab completion02:00
shentino what is tab completion?02:00
ojacobson Type 'ojaco' and press tab02:00
90% chance you're using a client that completes my nick02:01
shentino it doesn't help when I make a typo :P02:01
← sekati left02:01
← griotspeak left02:02
← ojacobson left02:03
← carlos-huang left02:03
shentino but anyway, I really was trying to be simple and have git by itself handle this task02:04
→ carlos-huang joined02:04
shentino so thanks for the education ^.^02:05
→ bmac joined02:06
→ natechan joined02:06
← mfcovington left02:07
JCii ojacobson: here's the output of git log. http://pastebin.com/raw.php?i=hNY31Ehc It looks like I did something so that my checkins are in two different branches, like the branch I merged from. I'll work on trying to reproduce it. I don't have a terminal session; I was using tortoisegit.02:07
← savid left02:08
→ Velpoman joined02:09
Velpoman Good evening all.02:09
EugeneKay Git doesn't do checkins02:09
Just drop that terminology from your brain02:09
JCii ok. commits.02:09
EugeneKay A git commit creates a snapshot of the tree.02:09
This tree can be diff'ed against the tree owned by any other commit(typically, the one that came just before it in history), but each tree is a thing all by itself.02:10
Velpoman I am get clone failures and --verbose isn't showing any output of use other than clone of foo into submodule path bar failed. Is there a way to generate more detailed debug output other than --verbose?02:10
← Prazanna left02:10
JCii tree == a tree of sha1 hashes of each file in each directory?02:11
Velpoman clone via git://, http:// both work, but since it's a recursive clone the https locations for the submodules fail and I can't rewrite urls in flight during the clone.02:12
JCii EugeneKay: can i see timestamps in the git log --graph? is git log chronological? i see the same comment message in different parts of my git log output.02:12
→ Cykey joined02:13
← Cykey left02:13
EugeneKay The defautl log format includes a datestamp02:13
← jaequery left02:14
→ rickruby joined02:15
→ jaequery joined02:17
← jaequery left02:17
→ devth joined02:17
→ jaequery joined02:18
sitaram shentino: taken literally, "incoming git pushes authenticated by ssh keys" just means you disable password access to that user (via sshd config) and add people's pubkeys to ~/.ssh/authorized_keys on the account that is holding the repos. You don't need gitolite for that02:18
shentino: what gitolite helps with is *authorisation* not authentication; see http://gitolite.com/gitolite/why.html02:18
→ AWinterman joined02:19
shentino oh ok, so what I'd be doing is basically using git wrapped in sshd02:19
← tjbiddle left02:19
shentino ...duh, why didn't I think of that02:19
I thought you could start git in daemon mode or something for that, is why I asked02:20
sitaram git in daemon mode is "git-daemon" which is UN-authenticated. Unless you're running an open source project or such there's almost no reason to ever enable that02:21
← amcsi_work left02:21
→ curvedmark joined02:21
← nathanr left02:22
← madewokherd left02:23
← Silne30 left02:24
ChanServ set mode: +o02:25
EugeneKay changed the topic to: Welcome to #git, home of unofficial support for most things git | Current stable version: 1.8.3 | Start here: http://jk.gs/git/ | Getting "Cannot send to channel" or unable to change nick? /msg gitinfo .voice | Ring around the repo, pockets full of plumbing, objects, objects, we all commit.02:25
← Alz_ left02:25
Kyle EugeneKay: new jokes <302:26
← Velpoman left02:26
EugeneKay Does anybody here know Chinese?02:26
→ alvaro_o joined02:26
oojacoboo yi din dien02:26
EugeneKay I need a thing translated02:26
oojacoboo google.com/translate02:27
Kyle lol02:27
EugeneKay I'd like to it actually be valid/sensible in the other language :-p02:27
Or any sufficiently non-ASCII language, really.02:27
oojacoboo well, I'm not the one… yi din dien = a little bit02:28
EugeneKay And yes, this is for a git-joke.02:28
Kyle EugeneKay: It can't be any worse than what we get from them in English heh02:28
Norrin ::dead::02:28
← AWinterman left02:29
← knite left02:31
JCii so in "git log" I see 22 commits which are a series of 11 duplicates. when I use "git diff" to compare any pair, I see the same 1 file missing. it looks like when I was comitting, 11 commits were duplicated. the problem i am encountering is git push (to github) says I am behind. when I git pull, the merge undoes all of my work that I merged in.02:31
← dsdeiz left02:31
→ dsdeiz_ joined02:31
EugeneKay Did you rebase/rewrite history?02:33
JCii i tried to remove a file that had a password in it. i thought that had worked successfully.02:33
← rakl left02:33
EugeneKay You need to push -f to github02:33
JCii i dont think the password ever made it to github (if that's relevant)02:34
EugeneKay It sounds like it did02:34
What you've done is change history that was previously published; you need to force a push to override it.02:34
JCii ok. how can i backup the repo on github, in case it all goes bad? clone in a fresh directory? anything else?02:35
EugeneKay Merging what github has does exactly what you would expect it to - bring back in the change that you undid.02:35
A clone someplace oughta be plenty.02:35
← fbdystang left02:35
← mckendricks left02:35
→ m0ltar joined02:37
← alfplayer left02:37
→ thierryp joined02:38
← allaire left02:38
← TomyLobo left02:39
← nedbat left02:40
→ w09x_ joined02:40
← thierryp left02:42
← ryez left02:43
← w09x left02:43
→ w09x joined02:44
← vultraz_laptop left02:45
→ Dougie187 joined02:45
→ relipse joined02:47
← w09x_ left02:47
← rnsanchez left02:48
→ mckendricks joined02:48
← Etheco left02:49
→ Etheco joined02:49
→ spb joined02:50
→ alfplayer joined02:51
→ rakl joined02:52
→ dsdeiz joined02:52
← devth left02:53
← dsdeiz_ left02:54
JCii EugeneKay: ok, push -f worked. thanks for the help on that. the duplicates are still there. they're visible on github (https://github.com/SavageLearning/Machete/commits/EmployerEmail) I wonder if that's easy or hard to clean up02:55
← m0ltar left02:55
→ AWinterman joined02:55
← flaguy48 left02:56
← mixel left02:56
→ mihamina joined03:00
← bbloom left03:00
Dougie187 If you do something like "git diff 6ab15c", is a the current working commit or is a the commit referenced on the line?03:00
→ bbloom joined03:00
← devoldmx left03:01
← guns left03:01
← Bass10 left03:02
← skorgon left03:02
← bitkiller left03:03
thiago Dougie187: git config diff.mnemonicprefix true03:03
it will show "c" for a commit and "w" for the working tree03:03
Dougie187 Ah, instead of a and b?03:03
thiago yes03:03
also "i" for the index03:03
← Noldorin left03:03
→ skorgon joined03:03
thiago it will show a and b when comparing two commits03:03
Dougie187 k, thanks03:03
→ theshadow joined03:04
Dougie187 so, c is the commit I reference, and w is the commit I'm on. Right?03:04
← alvaro_o left03:04
thiago w is your working tree03:05
← BJfreeman left03:05
thiago your working tree might differ from the commit you had checked out03:05
Dougie187 thiago: sure, but in a clean working tree it would be the commit I have in the directory.03:05
thiago right03:05
Dougie187 but it also includes any changes I've made since the commit that aren't committed.03:05
thiago if working tree == index == HEAD, then by the transitive property of comparisons, you're comparing one commit to the other03:06
Dougie187 :D03:06
← AWinterman left03:06
thiago btw, you may want git config --global03:06
Dougie187 yeah03:07
I used that when I set it.03:07
→ zxy joined03:08
← zangxiangyu left03:11
← lresende left03:11
← plumpkin left03:11
← SBoudrias left03:12
→ hoyang joined03:12
← surfdue left03:13
← pastime left03:13
→ Katafalkas joined03:14
← mihamina left03:14
← jacoluta left03:14
← LuxuryMode left03:15
→ beautifulmind joined03:15
→ surfdue joined03:17
← pitzips left03:18
→ devth joined03:18
→ flaguy48 joined03:18
← PanicG left03:19
→ tiny joined03:19
← devth left03:19
→ DolphinDream joined03:20
DolphinDream howdy03:20
tiny I have a project with submodules. Now I'd like to pull submodules on newly cloned project. Directories are there but no content. " git submodule foreach git update" doesn't do anything03:21
DolphinDream has anyone done a CVS to git migration ? i wonder what are the latest reliable tools to do that and what are the caveats to look for03:21
tiny Do I need to add submodules again even though I can see them with git submodule status?03:24
thiago tiny: git submodule init03:24
tiny no joy did it03:24
thiago: http://dpaste.com/1205780/03:25
thiago now git submodule update03:25
← Katafalkas left03:26
← natechan left03:26
tiny thiago: wow, your cmnd works03:28
thiago: ty03:28
→ Prazanna joined03:28
← Dougie187 left03:30
← nerdfiles left03:32
→ AWinterman joined03:32
← jaequery left03:33
← mmitchell left03:34
→ engkur joined03:34
→ darinmorrison joined03:37
← tinti left03:37
→ devoldmx joined03:38
→ b-wilson joined03:40
→ preflex_ joined03:42
← preflex left03:42
← AWinterman left03:43
preflex_ → preflex03:43
← pinpin404 left03:44
→ lifelike joined03:45
→ b-wilson_ joined03:45
← tiny left03:45
→ codeful joined03:45
← bmac left03:47
← b-wilson left03:48
← Davey left03:48
← lacrymology left03:49
← w09x left03:51
← chrob left03:53
→ subhojit777 joined03:53
→ fairuz joined03:56
← dougsland left03:56
← acharlieh left03:57
← howlymowly left04:01
← beautifulmind left04:01
← lifelike left04:01
← rakl left04:02
← pppaul left04:04
← thebetterbjasd left04:06
← Dreamer3 left04:06
← b-wilson_ left04:06
← kukks left04:06
→ b-wilson joined04:07
→ b-wilson_ joined04:07
→ skiernose joined04:08
→ AWinterman joined04:09
→ mihamina joined04:10
→ beautifulmind joined04:10
← codeful left04:10
← cilly left04:11
→ Katafalkas joined04:12
← bgerber left04:12
→ LuxuryMode joined04:12
→ cilly joined04:13
→ bgerber joined04:13
← AWinterman left04:20
← rostam left04:20
← imMute left04:22
→ rostam joined04:23
→ imMute joined04:24
→ saurabh joined04:25
← saurabh left04:25
← relipse left04:27
→ relipse joined04:27
← destiny left04:27
← eric_roque left04:28
← mencito left04:28
← daito left04:28
→ JosephSilber joined04:31
→ Davey joined04:31
→ tmcmahon joined04:32
← tmcmahon left04:32
→ lifelike joined04:32
→ cagedwisdom joined04:35
← cagedwisdom left04:35
← carlos-huang left04:36
← beautifulmind left04:36
← b-wilson left04:36
→ hobodave joined04:37
← the_cybersphinx left04:37
← JPeterson left04:37
→ JPeterson joined04:38
← raijin left04:39
← LuxuryMode left04:40
← nbouscal left04:43
→ ni291187 joined04:43
→ NaN joined04:43
→ LuxuryMode joined04:43
← bakingbread left04:44
→ mmitchell joined04:44
NaN if I checkout (to revert my changes to a previous commit) what means "Not currently on any branch." when I do $git status ?04:44
→ prashanth92 joined04:45
→ Tommy[D]_ joined04:45
→ nbouscal joined04:45
thiago you switched branches to a commit (a non-branch)04:45
if you want to check out the changes but not change branches, add a . at the end04:46
← _ikke_ left04:46
← b-wilson_ left04:46
→ AWinterman joined04:46
→ _ikke_ joined04:47
NaN thiago: so I "drop" my actual changes to get back to the past and now I'm not on any branch?04:47
offby1 !detached04:47
gitinfo A detached HEAD(aka "no branch") occurs when your HEAD does not point at a branch. New commits will NOT be added to any branch, and can easily be !lost. This can happen if you a) check out a tag, remote tracking branch, or SHA; or b) if you are in a submodule; or you are in the middle of a c) am or d) rebase that is stuck/conflicted. See !reattach04:47
thiago you checked out a commit04:47
git checkout does one of two things:04:47
check files out and switch branches04:47
you've done the latter04:47
← nbouscal left04:49
← mmitchell left04:49
← Tommy[D] left04:49
→ codepython777 joined04:50
← homunq left04:50
codepython777 how do i git clone a repo without any ".git" files.04:50
thiago hmm?04:50
→ MrRoot joined04:50
← prashanth92 left04:51
codepython777 thiago: I need to pull from one repo, then add the files to another repo. I would prefer a clean copy instead of having to deal with ".git" files.04:51
→ chimay joined04:51
NaN thiago: what? so a commit is a branch?04:52
→ rakl joined04:52
← fairuz left04:53
← skiernose left04:54
← AWinterman left04:57
→ hoschi joined04:57
offby1 codepython777: I think you want !git-archive04:57
← MrRoot left04:58
← luc_ left04:58
→ luc_ joined04:59
← LuxuryMode left04:59
→ LuxuryMode joined04:59
← LuxuryMode left05:00
← Rixius left05:00
← Davey left05:00
ryan_turner → ryan_turner_05:01
ryan_turner_ → ryan_turner05:01
→ japandroid joined05:02
← AlbireoX left05:02
EugeneKay codepython777 - man git-archive05:03
gitinfo codepython777: the git-archive manpage is available at http://jk.gs/git-archive.html05:03
EugeneKay Or read !deploy05:03
gitinfo Git is not a deployment tool, but you can build one around it(in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/the-list-and-irc/deploy.html05:03
→ pretty_function joined05:04
← Joshi3s` left05:04
← ni291187 left05:05
→ schua joined05:05
→ cbreak-work joined05:06
← chimay left05:07
→ tjbiddle joined05:07
→ jaequery joined05:09
← hobodave left05:10
codepython777 EugeneKay: or i could just delete .git from the root of the repo?05:13
EugeneKay Then you'll have to redownload all the git data when you want to do an update05:14
← tjbiddle left05:14
→ carlos-huang joined05:15
codepython777 EugeneKay: Thats the plan. I am adding it to a svn repo05:15
→ ffio joined05:15
EugeneKay You can svn-ignore the .git/ dir....05:15
And vice-versa05:15
← moondoggy left05:16
→ moondoggy joined05:16
← mribeirodantas left05:16
codepython777 true05:16
← agile left05:16
← _ffio_ left05:16
← a-l-e left05:17
→ daito joined05:18
→ peddamat joined05:20
← peddamat left05:21
← nicedice left05:21
← cbreak-work left05:22
← huoxito left05:22
← konundra left05:22
→ AWinterman joined05:23
→ irqq joined05:25
→ duko joined05:25
→ hwrd|work joined05:26
← duko left05:26
← avinashrbhat left05:27
← SeoZ left05:28
→ SeoZ joined05:28
← darinmorrison left05:29
→ bakingbread joined05:29
← relipse left05:29
← chamunks left05:30
← beneggett left05:31
→ chamunks joined05:31
← codepython777 left05:31
→ codepython7771 joined05:31
→ tjbiddle joined05:33
← nilg` left05:34
← AWinterman left05:34
→ Alz_ joined05:34
→ JohnBat26 joined05:34
→ fairuz joined05:34
EugeneKay yay, "Input/Output error"05:34
EugeneKay set mode: -o05:35
← solrize left05:35
← rakl left05:36
→ haijun joined05:37
← Goplat left05:37
→ tjbiddle_ joined05:37
→ rakl joined05:38
← tjbiddle left05:38
tjbiddle_ → tjbiddle05:38
EugeneKay runs fsck, cries a bit05:39
thiago NaN: yes, you can check out any commit05:39
NaN: but only "branches" show up in git st05:39
NaN thiago: got it, thanks :)05:39
thiago a branch is nothing but a label that points to a commit05:40
→ nilg joined05:41
→ jargon joined05:42
→ Cromulent joined05:45
← japandroid left05:46
→ scientes_ joined05:47
→ wazyl joined05:47
← haijun left05:48
wazyl when git tries to run my core editor being notepad -wl1 , i get the system cannot find the specified path .05:48
← threesome left05:50
→ saurabh joined05:50
← hoyang left05:50
→ breck7 joined05:51
→ epicowl joined05:51
← Alz_ left05:52
→ h0bbit joined05:53
← FreezingCold left05:53
→ AlbireoX joined05:54
← NaN left05:55
← SeoZ left05:56
← epicowl left05:57
→ epicowl joined06:00
→ suy joined06:00
← Gentlecat left06:00
→ AWinterman joined06:00
→ SeoZ joined06:01
← SeoZ left06:01
→ SeoZ joined06:01
→ ozette joined06:01
← stliu left06:02
→ Anders_J joined06:02
← Anders_J left06:02
→ avinashrbhat joined06:03
→ setmeaway joined06:03
← homer91 left06:05
← luc_ left06:05
→ r4ven joined06:05
→ luc_ joined06:06
← hwrd|work left06:09
← PaulCapestany left06:10
← AWinterman left06:11
→ Davey joined06:11
← zenwryly left06:11
→ danicampa901 joined06:12
→ pradi joined06:13
← breck7 left06:13
← mckendricks left06:14
→ ayaz joined06:14
← codepython7771 left06:14
→ ThomasLocke joined06:14
← Cromulent left06:16
← hakunin_ left06:18
→ hakunin joined06:19
Nickeeh When I git rebase, it puts my commits in the wrong order... I'd like it to take all the commits and apply them on top, not at the bottom. :/06:19
(it has to be the top because git-svn)06:19
→ eriik joined06:20
thiago sorry, that doesn't make sense06:20
→ thierryp joined06:20
thiago rebase applies the commits on top of the new base06:20
either you're mis-interpreting what Git is showing, or I misunderstand what you want.06:21
← thierryp left06:21
cmn Nickeeh: the instruction sheet reads top-to-bottom06:22
but you shouldn't be seeing the other commits anyhow06:23
Nickeeh Hmm... I'll take a screenshot, it shouldn't be too hard. :(06:23
Just early in the morning.06:23
← jaequery left06:23
← schua left06:23
Nickeeh http://i.imgur.com/kNzHfMk.png So I want to "cherry-pick" the commits of astrant/master onto master.06:25
oooh wait06:25
I should checkout a branch with astrant/master06:25
rebase on master06:25
derpderpderp.06:25
(right? :X)06:25
→ andref_dias joined06:25
Nickeeh hurray, cleanly applied06:26
For some reason cherry picking gave me a merge error. :/06:26
I probably messed that up as well.06:26
thanks for rubber-ducking guys. :D06:26
→ sebbi joined06:26
→ AdrienBrault joined06:28
← irqq left06:29
→ BiggFREE joined06:30
→ hinst joined06:30
← JosephSilber left06:31
Nickeeh (I did 'git branch temp; git reset --hard astrant/master; git rebase master; git checkout master; git merge --ff-only temp;)06:31
← Davey left06:31
← LekeFly left06:31
→ marthinal joined06:32
→ whowantstolivefo joined06:32
→ maslowbeer joined06:35
← zequence left06:35
→ rath joined06:35
→ zequence joined06:36
→ Anders_J joined06:37
→ AWinterman joined06:37
→ tsdh joined06:38
→ mishok13 joined06:39
→ Sigma joined06:44
Nevik Nickeeh: always glad to help by doing nothing \o/06:45
→ pantsman joined06:46
← pantsman left06:46
→ pantsman joined06:46
→ juanmabc joined06:46
→ chimay joined06:46
→ sekati joined06:47
← AWinterman left06:48
→ arvind_khadri joined06:49
→ d3pwnd joined06:49
← Adlai left06:49
← nilg left06:50
→ qiy joined06:50
gitinfo set mode: +v06:51
qiy what is a committer? should I change the committer or author?06:51
→ tjbiddle_ joined06:51
Nevik qiy: a committer is the person who creates the commit object06:52
the author is the person who is responsible for the changes contained therein06:52
qiy should I set the committer mail?06:52
→ Cassiel joined06:52
Nevik what do you mean?06:52
you set only your user.email and user.name config keys06:52
qiy i am both the committer and author06:52
Nevik yes06:52
git does that automatically06:52
← engkur left06:53
qiy i want to commit to my local repos and let other pull from it.06:53
the current committer mail is an internal one, should I set the committer mail properly or just set the author?06:53
→ thierryp joined06:54
qiy btw, where's the config keys set?06:54
Nevik you should set neither the committer nor the author env keys unless you have specific reason and know what you're doing06:54
qiy: http://git-scm.com/book/en/Getting-Started-First-Time-Git-Setup#Your-Identity06:54
qiy the default name/mail isn't the one i want to publish06:54
to the world06:55
→ nilg joined06:55
← tjbiddle left06:55
tjbiddle_ → tjbiddle06:55
Nevik you're not making any sense, man. git records your authorship identity in every commit06:55
you need to set it so git knows it06:55
if you could describe your whole problem, we can help you06:56
qiy yeah, but i want to set a proper name, instead of root@localhost06:56
Nevik well do what is described in that link06:56
set your config keys properly06:56
as you should have done when you first installed git06:56
→ LekeFly joined06:57
→ didlix joined06:58
← thierryp left06:58
→ ntus1017 joined06:59
← w0ng left07:00
← pantsman left07:00
← andref_dias left07:01
→ irqq joined07:02
→ thierryp joined07:02
→ miclen joined07:03
→ threesome joined07:03
→ booka3 joined07:04
→ norjor joined07:06
→ braisarias joined07:06
→ shentino_ joined07:08
← shentino_ left07:08
→ shentino_ joined07:08
← shentino left07:08
← AdrienBrault left07:09
shentino_ → shentino07:10
→ abrkn\ joined07:11
← albel727 left07:12
→ albel727 joined07:13
← hinst left07:13
→ daanavitch joined07:13
← daanavitch left07:14
→ ag4ve joined07:14
→ engkur joined07:14
→ AWinterman joined07:14
← abrkn left07:14
← epicowl left07:15
← jceb_ left07:15
EugeneKay qiy - just stop; start over. When you first started using git you should have set the options user.name and user.email in your ~/.gitconfig. man git-config and read more about these options.07:15
gitinfo qiy: the git-config manpage is available at http://jk.gs/git-config.html07:15
EugeneKay The difference betweeen author and committers is irrelavant to your needs, I guarantee it.07:16
→ jiku joined07:16
→ agile joined07:16
← shentino left07:16
EugeneKay You may want to go back and fix the user info in your existing repos, because you didn't do this to begin with. man git-filter-branch, see the example usage for --env-filter for how to do this.07:17
gitinfo the git-filter-branch manpage is available at http://jk.gs/git-filter-branch.html07:17
← norjor left07:18
→ pastime joined07:19
← mihamina left07:19
→ stamina1 joined07:20
→ _goTAN joined07:20
← perlygatekeeper left07:21
→ tvw joined07:21
← tjbiddle left07:21
← juanmabc left07:23
qiy EugeneKay: if some pull me. does he get my committer and author?07:23
← booka3 left07:23
← chichou left07:23
← irqq left07:24
→ chichou joined07:24
→ shentino joined07:24
← shentino left07:24
→ shentino joined07:24
→ xrq` joined07:24
→ irqq joined07:25
← AWinterman left07:25
← thierryp left07:26
→ kewah joined07:26
→ kesshi joined07:26
→ Zeelot3k joined07:27
EugeneKay Yes.07:27
→ juanmabc joined07:27
→ causasui_ joined07:28
← pradi left07:28
→ Sophira_ joined07:28
→ OMGOMG_ joined07:28
→ kcj joined07:28
→ majuscul1 joined07:28
→ muep_ joined07:28
→ comps_ joined07:29
→ pasky_ joined07:29
→ wting_ joined07:29
→ Zwei_ joined07:29
→ Balliad joined07:29
→ rawtaz_ joined07:30
→ jceb joined07:31
← sekati left07:31
← irqq left07:31
← ayaz left07:31
→ milian_ joined07:32
→ s4muel_ joined07:32
→ nemesit|- joined07:32
→ rudak joined07:32
→ RichardBronosky_ joined07:32
→ esto- joined07:32
→ jnavila joined07:33
→ Brando753-o_O_o joined07:33
→ bvi joined07:33
→ psoo joined07:34
← Katafalkas left07:34
← redrocket left07:34
← nemesit|znc left07:34
← ping-pong left07:34
← causasui left07:34
← esto left07:34
← Nibato left07:34
← Khisanth left07:34
← BBonifield left07:34
← tirengarfio left07:34
← nutron|w left07:34
← TheMoonMaster left07:34
← ghoti left07:34
← rawtaz left07:34
← sgkim126t left07:34
← Brando753 left07:34
← jamesd left07:34
← xrq left07:34
← Sophira left07:35
← Exodist left07:35
← wting left07:35
← skorgon left07:35
← KillYourTV left07:35
← baneM left07:35
← asakura left07:35
Brando753-o_O_o → Brando75307:35
milian_ → milian07:35
← rchavik left07:35
← BiggFREE left07:35
→ ghoti joined07:35
→ albel727_ joined07:39
← dreiss left07:39
→ AlekxE joined07:40
→ redrocket joined07:40
→ nutron|w joined07:40
← ghoti left07:40
← stamina1 left07:41
→ tirengarfio joined07:41
→ Khisanth joined07:41
→ juril joined07:41
→ stamina joined07:42
→ ghoti joined07:42
← albel727 left07:42
albel727_ → albel72707:42
→ BiggFREE joined07:43
→ esc joined07:43
← esc left07:43
→ esc joined07:43
← thelamest left07:43
→ jnavila_ joined07:44
← jnavila left07:45
→ jnavila__ joined07:45
→ douglaslassance joined07:45
→ timj1 joined07:45
→ thelamest joined07:46
← Klowner left07:46
← ferdev left07:47
← ghoti left07:47
→ hinst joined07:47
→ ghoti joined07:47
← timj left07:48
→ Klowner joined07:48
← jnavila_ left07:49
→ Jackneill joined07:50
← rakl left07:50
adhawkins-away → adhawkins07:51
→ AWinterman joined07:51
← thiago left07:52
→ rchavik joined07:52
→ eijk_ joined07:52
← eijk_ left07:52
→ eijk_ joined07:52
← ghoti left07:52
← CapnKernel left07:53
→ ghoti joined07:53
← Prazanna left07:53
← braisarias left07:54
→ kerrick_ joined07:55
→ Macaveli joined07:56
qiy EugeneKay: ok, i've now set my user.name and user.email07:56
EugeneKay: are the committer and author the same thing now?07:56
→ snearch joined07:57
→ _Qman joined07:58
← _Qman left07:58
→ dreiss joined07:58
← ghoti left07:58
qiy Nevik:07:58
→ ghoti joined07:59
→ codeful joined07:59
← bvi left08:01
← Balliad left08:01
← AWinterman left08:02
→ _Qman joined08:03
← wazyl left08:03
→ Balliad joined08:03
← chamunks left08:03
→ howlymowly joined08:03
← ghoti left08:04
← cavel left08:04
→ chamunks joined08:04
→ Asenar joined08:05
← simesy left08:05
→ ghoti joined08:05
→ khushildep joined08:06
fairuz I try my luck here for gerrit specific question. :) for project access, how can I remove Rights inherits from? It seems there's no button to remove this.08:07
→ simesy joined08:07
← scientes_ left08:08
← ghoti left08:10
→ braisarias joined08:10
← LimeBlast left08:11
→ ghoti joined08:11
← chimay left08:11
→ ferdev_ joined08:12
← JohnBat26 left08:12
← dreiss left08:12
→ grizzle joined08:13
erkules_ → erkules08:13
→ stephCoue joined08:14
→ bvi joined08:15
→ gan joined08:15
→ gan_ joined08:15
← ghoti left08:16
← gan_ left08:16
← dhruvasagar left08:17
← xxtjaxx left08:17
→ ghoti joined08:17
→ dhruvasagar joined08:18
← carlos-huang left08:19
→ leeN joined08:19
→ mikecmpbll joined08:19
→ ping-pong joined08:20
→ Nibato joined08:20
→ BBonifield joined08:20
→ asakura joined08:20
→ TheMoonMaster joined08:20
→ sgkim126t joined08:20
→ jamesd joined08:20
→ Exodist joined08:20
→ andres joined08:20
→ ufk_ joined08:20
→ ickyb0d joined08:20
→ epitron joined08:20
→ hhuuggoo joined08:20
→ ajpiano joined08:20
→ MetaCosm joined08:20
→ RichardJ joined08:20
→ xoebus joined08:20
← noname001 left08:20
→ harish joined08:21
→ KillYourTV joined08:21
→ crash\ joined08:21
← cyphase left08:23
→ noname001 joined08:23
← ghoti left08:23
→ ghoti joined08:24
→ skorgon joined08:24
← Cassiel left08:24
← howlymowly left08:25
→ LimeBlast joined08:25
← daito left08:26
→ jnavila joined08:26
← crash\ left08:27
← alfplayer left08:27
← jnavila__ left08:27
→ AWinterman joined08:28
→ baneM joined08:29
← kcj left08:29
← harish left08:30
← dorkmafia left08:31
← danicampa901 left08:32
← simesy left08:32
→ danicampa901 joined08:32
→ Kwpolska joined08:33
→ Raging_Hog joined08:33
← OnlyMax left08:33
← _Qman left08:34
← skorgon left08:34
→ vedant joined08:35
→ cyphase joined08:36
← tharkun left08:37
→ irqq joined08:37
→ tharkun joined08:37
→ _Qman joined08:37
← dhruvasagar left08:37
← _Qman left08:37
← kerrick_ left08:38
→ sebrock joined08:38
← AWinterman left08:39
→ dhruvasagar joined08:39
← douglaslassance left08:40
→ thiago joined08:41
→ thierryp joined08:42
→ qiy_ joined08:42
gitinfo set mode: +v08:42
→ daito joined08:42
← daito left08:43
→ diegoviola joined08:43
→ zoresvit joined08:43
→ daito joined08:43
→ CapnKernel joined08:45
← qiy left08:45
qiy_ → qiy08:45
← vedant left08:45
→ alfplayer joined08:47
← daito left08:47
→ daito joined08:49
← thierryp left08:49
→ thierryp joined08:49
→ achadwick joined08:50
→ rob_w joined08:50
← ntus1017 left08:52
→ alorence joined08:52
→ m8 joined08:52
→ howlymowly joined08:52
← m8 left08:53
→ w0ng joined08:53
→ m8 joined08:54
→ cavel joined08:54
← Macaveli left08:56
→ Macaveli joined08:57
→ skorgon joined08:57
→ ayaz joined08:57
→ jkremser joined08:58
← jkremser left08:58
→ jkremser joined08:58
→ PigeonFriend joined08:58
← rickruby left08:59
→ pkrzywicki joined09:00
← ozette left09:01
← chamunks left09:02
→ solrize joined09:04
← miclen left09:04
→ AWinterman joined09:05
← sebrock left09:06
← tvw left09:07
← skorgon left09:07
← gan left09:08
← solrize left09:08
→ Iszak joined09:09
← atomx left09:09
← jgoss left09:12
→ abrkn joined09:12
← donmichelangelo left09:12
→ donmichelangelo joined09:13
← LekeFly left09:13
← babilen_ left09:13
← jiangenj left09:15
← dsirijus left09:15
← abrkn\ left09:16
← AWinterman left09:16
→ msekleta joined09:17
→ GarethAdams joined09:17
→ Thra11 joined09:17
→ spuz joined09:17
→ darnir_ joined09:18
← darnir_ left09:18
spuz hello, if I am detached from a branch, is it possible to reset the branch to my current HEAD?09:18
→ darnir_ joined09:19
→ solrize joined09:20
→ dsirijus joined09:20
← darnir_ left09:21
→ darnir_ joined09:21
← codeful left09:21
← darnir_ left09:21
→ darnir joined09:22
→ abrkn\ joined09:23
qiy what's commit object?09:23
← CapnKernel left09:24
← kalleth left09:24
→ nmend joined09:25
← m8 left09:25
← trollwork left09:25
→ m8 joined09:25
→ ozette joined09:26
spuz qiy: is that a question to me or to everyone?09:27
← abrkn left09:27
qiy spuz: every 109:28
include U09:28
← abrkn\ left09:28
spuz ok, a commit object is an object that represents a commit :p09:28
→ abrkn joined09:28
qiy git-commit-tree - Create a new commit object09:28
← BiggFREE left09:29
qiy i think git-commit does that too09:29
← ickyb0d left09:29
← RichardJ left09:29
→ RichardJ joined09:29
← TheMoonMaster left09:30
← asakura left09:30
rawtaz_ → rawtaz09:30
← andres left09:30
← albel727 left09:30
→ TheMoonMaster joined09:30
→ andres joined09:30
qiy ok, i see09:30
→ nizeguy joined09:30
qiy its parents09:30
← rudak left09:30
qiy bye09:30
→ albel727 joined09:30
← MetaCosm left09:31
← qiy left09:31
→ MetaCosm joined09:31
→ ickyb0d joined09:31
→ rudak joined09:32
→ netiul joined09:32
→ asakura joined09:32
→ CapnKernel joined09:33
charon not that someone who needs to ask that question should even consider using commit-tree09:33
← rudak left09:33
charon spuz: git checkout -B whatevername09:33
will also do the checkout, otherwise branch -f09:34
spuz: note that 'detached' is a global state, and does not apply to any branch (i.e., you are detached from *all* branches, or you are attached to exactly one branch)09:34
→ OOPMan joined09:34
charon then again i think we did hack up some help messages to show the last branch you were on, so meh.09:35
→ raging_swine joined09:35
→ knite joined09:35
← Raging_Hog left09:35
spuz charon: you are a git contributor?09:36
charon nowai09:36
(yes)09:36
spuz ok09:36
I won't tell anyone09:36
charon heh09:36
i think it's in my /whois09:36
(the name, but it's easy enough to go from there)09:37
→ kalleth joined09:37
spuz I think branch -f is what I wanted but I just ended up doing git checkout followed by git reset <hash>09:38
→ skorgon joined09:39
charon i hope you used --keep or --hard there, as otherwise the index will be off09:40
but anyway, other than the fact that all your mtimes changed and you will probably have to rebuild more than necessary, they are equivalent09:41
← CapnKernel left09:41
→ trollwork joined09:41
→ pradi joined09:41
→ AWinterman joined09:43
→ simesy joined09:43
← Macaveli left09:43
→ Macaveli joined09:43
→ babilen_ joined09:44
babilen_ → babilen09:45
→ NET||abuse joined09:45
← solrize left09:46
→ crash\ joined09:46
← h0bbit left09:47
→ tbosua joined09:48
→ Rangarao joined09:48
← m8 left09:48
← simesy left09:48
→ freestyl3r joined09:49
← zxy left09:51
→ gasbakid joined09:51
→ beautifulmind joined09:51
← freestyl3r left09:53
← AWinterman left09:53
→ GenTAR joined09:55
→ gusnan joined09:55
→ m8 joined09:55
← fairuz left09:56
← GenTAR left09:56
→ GenTAR joined09:56
→ jnewt joined09:57
← GenTAR left09:57
← nmend left09:58
→ semferon joined09:58
← raging_swine left09:58
→ VinceThePrince joined10:00
← yessir left10:00
→ gan joined10:02
→ BBonifield_ joined10:03
→ jnavila_ joined10:03
→ jamesd_ joined10:03
← jnavila left10:03
→ cb_ joined10:03
→ hhuuggoo_ joined10:03
→ sgkim1261 joined10:03
→ jnavila__ joined10:03
← Macaveli left10:04
→ exodist_ joined10:04
→ epitron_ joined10:04
→ Vile joined10:04
← BBonifield left10:05
← skorgon left10:05
← jamesd left10:05
← hhuuggoo left10:05
← xoebus left10:05
← Vile left10:05
← asakura left10:05
← baneM left10:05
← KillYourTV left10:05
← ping-pong left10:05
← Nibato left10:05
← sgkim126t left10:05
← ufk_ left10:05
← epitron left10:05
← ajpiano left10:05
← Exodist left10:05
→ Nibato joined10:05
← semferon left10:05
→ ping-pong joined10:05
YoungFrog Yesterday I asked about this, but I don't think an answer was found, perhaps I have more luck today : where is the use of "*" in a refspec for "git fetch" documented ? I was naively looking at "git help fetch" but under <refspec> I see no mention of it.10:05
→ ufk_ joined10:05
→ ajpiano joined10:05
charon YoungFrog: well, it's sort of implicit, it says "The remote ref that matches <src> is fetched". the src * will just match "more components here" and the * on the dst side will be replaced with the matching part.10:06
but you're right, it's not documented very well10:06
→ tvw joined10:07
← bvi left10:07
← codertux left10:07
← jnavila_ left10:07
YoungFrog charon: thanks for your answer. As someone pointed out yesterday, it's a bit more generic than usual shell gobbling too since it can match subdirs, right ? Should I report a bug to have it documented ?10:09
→ codertux joined10:09
→ RaceCondition joined10:09
→ rickruby joined10:10
→ xxtjaxx joined10:12
← xxtjaxx left10:12
→ xxtjaxx joined10:12
→ Katafalkas joined10:13
← rickruby left10:14
← darnir left10:14
charon sure, sounds like an idea. but yes, it does match subdirs. if you also check whether it can match parts of components (i don't think so) you have all the information and can in fact suggest a reworded explanation ;)10:17
← avinashrbhat left10:18
← falood left10:19
← mikecmpbll left10:19
→ KillYourTV joined10:19
→ CapnKernel joined10:19
→ CapnKernel1 joined10:20
→ AWinterman joined10:20
→ FernandoBasso joined10:21
→ bvi joined10:22
Nevik osse: around?10:22
← msekleta left10:22
← yacks left10:22
→ Marenz joined10:23
→ mklappstuhl joined10:23
→ raging_swine joined10:23
→ falood joined10:23
← CapnKernel left10:24
Marenz Hello10:24
→ baneM joined10:24
→ asakura joined10:24
Marenz I have two files in one repo who are pretty similar. I wanna merge them to one file (at a new location). Can I keep the history of both linked to that new file?10:24
charon Marenz: !track_rename, also applies to copying10:25
gitinfo Marenz: git does not track renames of files in history, git only records tree snapshots. It can try to guess whether a change was a rename, or just unrelated removal+addition, when asked. You can help this by making the rename in a different commit from a modification. See man git-log -C and -M10:25
← stamina left10:25
charon Marenz: note that 'blame -C -C' will probably figure it out10:25
→ skorgon joined10:25
Marenz so git would probably figure out that it has two parents.. kind of10:27
charon not really, git-blame would figure out that the file contents in parts are so suspiciously similar to what you removed in another file that they can be "explained" as having been copied/moved from there10:27
Marenz hmm thanks10:28
→ marhaban joined10:29
→ albel727_ joined10:29
← AWinterman left10:30
← k_sze[work] left10:30
← albel727 left10:31
→ nedbat joined10:32
→ DonRichie joined10:33
→ the_cybersphinx joined10:33
→ _Qman joined10:36
→ Noldorin joined10:37
← flavius left10:37
← gan left10:37
albel727_ → albel72710:38
→ flavius joined10:38
→ aendrew joined10:39
osse Nevik: perhaps10:40
Nevik osse: >:O10:41
← luc_ left10:41
← raging_swine left10:42
← jiku left10:44
→ jiku joined10:44
← flavius left10:44
→ quetz joined10:44
→ marhaban1 joined10:44
← mklappstuhl left10:44
Nevik osse: do you feel like testing a new script of mine? ;)10:44
← aendrew left10:45
YoungFrog charon: update on the previous question: it seems also that multiple fetch = entries are allowed in .git/config -- that's not documented either, is it ?10:45
← marhaban left10:47
theahindle Hi - getting this error: CONFLICT (modify/delete): inc/settings.php deleted in c00736c3a1ffcb94f5c38436b49f781af9c2bbb8 and modified in HEAD. Version HEAD of inc/settings.php left in tree.10:48
← bvi left10:48
theahindle I removed inc/settings.php on my local machine, added it to .gitignore as 'settings.php' (does that have to be 'inc/settings.php'?) and then tried to pull on a remote server (after pushing), and that happened :(10:48
→ aendrew joined10:48
moritz and the upstream modified it10:49
← stephCoue left10:49
moritz and now git tells you it doesn't know what to do10:49
because it has two different modifications (one changed, one deleted)10:49
!config_files10:49
gitinfo [!configfiles] It is recommended to store local configuration data in a file which is not tracked by git, but certain deployment scenarios(such as Heroku) may require otherwise. See https://gist.github.com/1423106 for some ideas10:49
→ natechan joined10:50
ChanServ set mode: -v10:50
djh_ → djh10:50
→ polymar joined10:50
→ mklappstuhl joined10:52
← FernandoBasso left10:53
→ Raging_Hog joined10:53
← marhaban1 left10:54
theahindle I removed inc/settings.php from the remote server and repulled and it still happened - unless it's considering the removal on that server as a modification even though it's the same as the one I made locally10:54
← mklappstuhl left10:55
→ AWinterman joined10:57
→ marhaban joined10:57
→ albel727_ joined10:58
esc whats the sausage thing again?10:58
!sausage10:58
gitinfo [!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 !perfect10:58
→ mz|`_ joined10:58
→ noxd joined10:58
← mz|` left10:59
← mz|`_ left10:59
→ mz|`_ joined10:59
← hyperair left11:01
← rob_w left11:01
← albel727 left11:01
moritz theahindle: how did you remove it from the remote server?11:02
← Kadmos left11:02
→ chamunks joined11:03
→ daniel_- joined11:03
← parhs left11:03
← d3pwnd left11:03
← liquid-silence left11:04
moritz theahindle: anyway, you'll likely want to do this: git rm inc/settings.php; git commit11:04
→ mikecmpbll joined11:04
→ liquid-silence joined11:05
← cilly left11:05
→ Kadmos joined11:05
→ parhs joined11:06
← parhs left11:06
→ parhs joined11:06
→ ryez joined11:06
← RaceCondition left11:06
← daito left11:06
→ brocken_ joined11:07
← danicampa901 left11:07
← AWinterman left11:08
← brocken left11:10
← marhaban left11:10
→ marhaban1 joined11:11
→ RaceCondition joined11:13
→ Andre_work joined11:14
← kewah left11:14
← arnihr left11:14
← PaulePanter left11:14
→ Moult_ joined11:14
← BtbN left11:14
← kemyd left11:14
← _rgn left11:14
→ txwikinger2 joined11:14
← txwikinger2 left11:14
→ txwikinger2 joined11:14
→ ferai joined11:14
← rndbit left11:14
← jefferai left11:14
← cmn left11:14
→ KRF_ joined11:14
← wereHamster left11:14
→ PeterBooker joined11:15
← pradi left11:15
→ wereHamster joined11:16
→ marhaban joined11:16
← marhaban1 left11:17
→ PaulePanter joined11:17
→ rndbit joined11:19
← rndbit left11:19
→ cmn joined11:19
→ marhaban1 joined11:19
→ manitu joined11:19
← marhaban1 left11:19
→ BtbN joined11:19
← DolphinDream left11:20
→ rndbit joined11:20
→ Sonderblade joined11:20
→ d3pwnd joined11:20
← d3pwnd left11:20
→ knittl joined11:20
→ d3pwnd joined11:21
← d3pwnd left11:21
← w0ng left11:22
→ d3pwnd joined11:22
← d3pwnd left11:22
→ d3pwnd joined11:22
← d3pwnd left11:22
→ w0ng joined11:23
→ d3pwnd joined11:23
← d3pwnd left11:24
→ pdelvo joined11:25
→ pradi joined11:26
← w0ng left11:27
← jnavila__ left11:27
→ jnavila__ joined11:27
→ kewah joined11:27
→ angelsl joined11:29
→ bmac joined11:29
→ raijin joined11:30
← yshh left11:30
→ arnihr joined11:33
← nedbat left11:33
→ AWinterman joined11:34
← CapnKernel1 left11:35
← braisarias left11:35
→ mrwonko joined11:39
→ sharpshooter joined11:39
→ nerocrux joined11:40
→ vultraz_laptop joined11:40
→ acharlieh joined11:40
mrwonko svn does not store the history locally so if the remote repo no longer exists the best I can get is converting the locally checked out revision to git, right?11:41
← brocken_ left11:41
jast mrwonko: yes11:41
mrwonko any way to automate that or do I just manually remove all the .svn directories, then git init . && git add * ?11:42
→ yacks joined11:43
→ MarkTomm joined11:43
moritz find -type -d -name .svn -delete11:43
← knite left11:44
← AWinterman left11:45
gitinfo set mode: +v11:45
→ brocken joined11:46
← mikecmpbll left11:46
← MarkTomm left11:48
→ barneybook|2 joined11:48
→ relipse joined11:49
osse Nevik: bring it on11:50
→ style joined11:50
← Fandekasp left11:50
← barneybook left11:50
style Hi, I merged my dev branch into master. it failed and I made an commit (no conflicts). But now it seems like it didn't merge everyting. Just some parts of my dev branch. And now if I try to remerge it to the master it says "Already up to date". So is there some "force merge" or can I rever merge somehow (remember I made commit after merge) and then merge it again11:51
→ AdrienBrault joined11:51
← AdrienBrault left11:52
→ mmitchell joined11:52
→ zulu-gula joined11:52
← zulu-gula left11:53
← bmac left11:54
charon YoungFrog: hmm, i'd have to go looking. i suspect it falls in the category of "so immediately obvious to anyone who has been working with it for years that it's not worth documenting".11:54
YoungFrog: so yes, by all means point these things out so we can fix the docs11:54
← OOPMan left11:55
→ looper joined11:55
→ fridim_ joined11:56
← G0re left11:56
→ albel727 joined11:56
← mmitchell left11:56
→ albel727__ joined11:58
→ daube joined11:58
← engkur left11:58
→ sgambino joined11:59
← albel727_ left11:59
→ julien76 joined11:59
→ stephCoue joined11:59
← Immutable left12:00
→ G0re joined12:00
← albel727 left12:01
→ Immutable joined12:01
→ BW^- joined12:02
BW^- to get a repo file of a certain commit printed out to the console,12:02
is it git show commitid:filename ???12:02
so like, cat filename-of-commitid , this is what i want :)12:03
how do?12:03
charon BW^-: 'git show commit:path' is correct12:04
BW^- charon: weird! that just gives me the diff!12:04
charon then you most likely misspelled it, as 'git show commit -- path'12:04
BW^- charon: or gives me,12:04
charon which is something entirely different12:04
BW^- fatal: ambiguous argument '0b634.....filename': unknown revision or path not in the working tree.12:05
Use '--' to separate paths from revisions12:05
charon !repro12:05
gitinfo Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session -- or, even better for complex issues, design a minimal case in which your problem can be reproduced, and share it with us. This will help immensely with troubleshooting.12:05
← mrwonko left12:05
BW^- hmm12:05
← stephCoue left12:05
→ codepython777 joined12:05
→ brocken_ joined12:05
→ jimi_ joined12:06
BW^- git-show <option>...12:06
DESCRIPTION12:06
Shows commit log and textual diff for a single commit. The command internally invokes git-rev-list piped to git-diff-tree, and takes command line options for12:06
both of these commands. It also presents the merge commit in a special format as produced by git-diff-tree --cc.12:06
this is not what I want, i want a copy of the file printed to the console!12:06
← Raging_Hog left12:06
jimi_ I have a github repo. I also have a local directory w/ the same files as my github repo, but isn't initid as a git repo locally. I want to use my local directory and push it to that repo, using my local copy as the most recent. How can I do this12:06
charon BW^-: well, i already told you that 'git show commit:path' is correct. if that does not work, please paste(!) what you tried and what it said, and preferable something that also proves that commit and path actually exist, e.g. by saying 'git show commit' and 'git ls-tree -r commit'12:07
← devoldmx left12:08
BW^- charon: i don't want git show; i want to get a copy of the entire file content of a file in a particular commit12:08
how do i do that?12:08
i used to know how to do this but just can't recall, and am in man git :))12:08
gitinfo the git manpage is available at http://jk.gs/git.html12:08
charon BW^-: git show commit:path >filename12:08
BW^- git tap or something12:08
charon using, you know, the shell.12:08
BW^- yeah12:08
weird12:08
charon: but then why does man git-show say that "Shows commit log and textual diff for a single commit"?12:09
gitinfo charon: the git-show manpage is available at http://jk.gs/git-show.html12:09
charon BW^-: because that's what it will show when you give it a commit, and the description never got extended to several commits and blobs12:09
BW^- aha, "For plain blobs, it shows the plain contents."12:09
← thiago left12:09
← brocken left12:09
charon jimi_: sounds like you're going to do this for all the wrong reasons, but meh. while there are shorter ways if you feel comfortable with git, the safest route is to clone the repo from github, copy over your new content, 'git add -A' and 'git commit' and 'git push'12:09
→ w0ng joined12:09
jimi_ charon, before i do that, whats the wrong reasons?12:10
charon, I did some dev work offline on a new laptop, and now once I got everything set up, i want to use the new files instead of the old ones in git12:10
→ PanicG joined12:11
charon jimi_: ah, well then. just make sure you actually commit on top of the previous state12:11
i.e., if you did not work on what is currently github's 'master', do something like 'git checkout -b topic <commit you started working at>; <commit stuff as above>; git checkout master; git merge topic'12:11
jimi_ charon, instead of clone, can I just git init, then git add remote... and then push?12:11
ah12:11
charon if you want history to remain connected, it'll be more of a hassle than just cloning12:12
→ AWinterman joined12:12
← daniel_- left12:12
charon it's entirely possible but you'd need some lowlevel git commands, and for what gain? you have to fetch the history regardless12:12
jimi_ oh ok12:12
→ braisarias joined12:13
BW^- charon: got it to work :D12:13
← GarethAdams left12:13
BW^- charon: weird. so i was in a subdir and did git show commitid:filename , did not work, got the above error12:13
← brocken_ left12:13
← parhs left12:13
BW^- so i did cd .. to get to the root dir of the git repo12:14
and did git show commitid:dirname/filename , and that worked12:14
← pastime left12:14
→ brocken joined12:14
← rurufufuss left12:14
charon BW^-: this is how that : "operator" works. you should probably upgrade your git, we've had a heuristic error message for a while that would tell you 'path does not exist in commit, did you perhaps mean commit:./path?'12:14
→ destiny_ joined12:15
BW^- charon: aha got you12:15
← Kadmos left12:15
→ kemyd joined12:16
← daube left12:16
→ tinti joined12:18
→ gan joined12:21
→ Raging_Hog joined12:21
→ andref_dias joined12:21
→ mmitchell joined12:22
← AWinterman left12:22
← chamunks left12:22
→ mikecmpbll joined12:23
→ CapnKernel joined12:23
→ chamunks joined12:23
→ DoniRS joined12:24
→ m0ltar joined12:24
← VinceThePrince left12:24
← m0ltar left12:24
← nerocrux left12:25
→ m0ltar joined12:25
→ Macaveli joined12:25
→ mklappstuhl joined12:25
← CapnKernel left12:25
← jnavila__ left12:25
→ pmdarrow joined12:26
→ tmcmahon joined12:27
→ bodom joined12:27
→ albel727 joined12:27
← barneybook|2 left12:27
→ barneybook|2 joined12:27
← albel727__ left12:30
→ _rgn joined12:30
→ sampiero joined12:30
→ cortes joined12:32
→ hyperair joined12:33
ferai → jefferai12:34
← mtk left12:35
← dr_lepper left12:35
← pradi left12:35
→ konundra joined12:36
→ TomyLobo joined12:36
→ yshh joined12:37
→ dr_lepper joined12:37
→ geekbri joined12:37
→ mattp_ joined12:38
→ mtk joined12:39
→ stephCoue joined12:39
← tmcmahon left12:41
→ trungqtran joined12:42
← arand left12:43
→ pdelvo_ joined12:43
← noxd left12:43
→ pitzips joined12:44
→ danicampa901 joined12:44
← pdelvo left12:45
→ arand joined12:45
→ rnsanchez joined12:46
→ bmac joined12:46
→ allaire joined12:48
→ breakingthings joined12:49
→ daniel_z joined12:49
→ AWinterman joined12:49
← arvind_khadri left12:50
→ norberhuis joined12:51
→ decotii joined12:52
← norberhuis left12:52
→ fitztrev joined12:52
← fitztrev left12:52
→ fitztrev joined12:52
← CodeShark left12:53
jimi_ how can i create a patch from my last 2 commits?12:54
→ _BJfreeman joined12:54
← acharlieh left12:55
_BJfreeman → BJfreeman12:55
← fridim_ left12:56
→ hellopat joined12:56
→ homunq joined12:57
← saurabh left12:58
jokajak jimi_: a single patch or multiple patches? You could do something like git diff HEAD^^..HEAD12:58
← jimi_ left12:59
→ jimi_ joined12:59
← kaictl left12:59
jimi_ whooops12:59
I was thinking of something like git format-patch -212:59
→ kukks joined12:59
← kukks left12:59
→ kukks joined12:59
→ nerocrux joined13:00
← AWinterman left13:00
← nerocrux left13:00
→ kaictl joined13:01
→ mcWolf2 joined13:01
→ harish joined13:01
← mcWolf2 left13:02
← Raging_Hog left13:04
→ lamarus joined13:05
← lamarus left13:05
← dsdeiz left13:05
→ lamarus joined13:05
→ AaronMT joined13:05
→ d3pwnd joined13:06
← harish left13:06
← threesome left13:06
→ shadok joined13:06
← pretty_function left13:09
← codepython777 left13:09
← Macaveli left13:09
← allaire left13:10
→ codeful joined13:10
→ Macaveli joined13:10
→ jonatasnona joined13:10
→ codepython777 joined13:11
→ rickruby joined13:12
← codepython777 left13:12
→ MrRoot joined13:12
← cirenyc left13:13
← kewah left13:14
← gusnan left13:15
→ shong joined13:15
→ ffio_ joined13:15
→ barneybook joined13:16
← codeful left13:16
← ffio left13:16
← shong left13:17
← falood left13:17
→ devoldmx joined13:18
→ kewah joined13:19
← barneybook|2 left13:19
← gasbakid left13:20
← rickruby left13:20
→ fooey joined13:20
→ albel727_ joined13:21
→ sebrock joined13:21
→ v0n joined13:22
← devoldmx left13:22
→ devoldmx joined13:23
← devoldmx left13:23
→ devoldmx joined13:23
→ scampbell joined13:23
← albel727 left13:23
← fitztrev left13:23
→ AWinterman joined13:27
→ eijk joined13:28
← Immutable left13:29
→ Immutable joined13:29
osse jimi_: git format-patch HEAD~313:29
→ vpopov joined13:30
← beautifulmind left13:30
lov I'm curious about git's revision selection. Is there a way to say HEAD...@{u} but ALSO include the commit from which both HEAD an @{u} diverged, but NOT any commits before?13:31
my usecase is that I want a nice alias for git --graph --decorate HEAD...@{u} so I can see where my branch and its upstream have diverged, and having the parent makes for a much nicer presentation.13:31
← eijk_ left13:31
SeoZ → SeoZ-work[AWAY]13:32
osse lov: you can do HEAD~1..@{u}13:32
lov that would only show commits that are in @{u} but NOT HEAD, which doesn't help when I've diverged.13:33
Take a look at http://git-scm.com/book/en/Git-Tools-Revision-Selection for the example nomenclature; I want to see B,C,D,E, and F in my output13:33
triple dot does ALMOST what I want but not quite.13:33
osse oh, I meant HEAD~1...@{u}13:34
lov that still doesn't help, since it doesn't show @{u} if there's nothing new in @{u}.13:34
→ narcos joined13:34
osse :/13:34
← devoldmx left13:34
lov as I mentioned, I want HEAD, @{u} AND the commit where they split.13:34
(but not anything further!)13:34
→ bvi_ joined13:35
→ devoldmx joined13:35
→ R1KRDO joined13:35
narcos Hi all. 'git diff foobar.py' tells me: "old mode 100644 new mode 100755", without listing any line changes. What does this mean?13:35
charon lov: --boundary13:35
lov is this possible or am I asking for something that requires a shell script? :(13:35
charon with triple-dots as before13:35
osse lov: git log --graph --decorate ^HEAD~1 HEAD @{u}13:35
perhaps13:35
canton7 narcos, it became executable13:36
lov charon: nice, thanks13:36
exactly what I wanted13:36
→ eurippa joined13:37
charon it's funny however that if i apply this to origin/next..origin/pu in git.git, the presentation turns from very nice to absolutely horrible :)13:37
i guess for a more straightforward divergence it makes more sense13:37
lov yeah, I have a pretty simple use case (I've gotten used to doing fetch/merge instead of pull) so it's not a big deal for me :)13:38
← AWinterman left13:38
adhawkins → adhawkins-away13:39
→ john2496 joined13:40
→ nedbat joined13:42
→ DolphinDream joined13:42
gitinfo set mode: +v13:43
DolphinDream howdy gitters13:43
→ chazcafe joined13:43
→ nicknisi joined13:43
→ charliesome joined13:44
← chazcafe left13:45
← devoldmx left13:45
→ devoldmx joined13:45
narcos canton7: ahhh! Thanks13:46
→ cannonball joined13:47
← Milossh left13:47
→ TeckniX joined13:47
→ Gentlecat joined13:48
→ codeful joined13:48
← dhruvasagar left13:48
→ chazcafe joined13:50
← BW^- left13:50
→ dhruvasagar joined13:50
→ Milossh joined13:53
→ nerdfiles joined13:54
→ cirenyc joined13:55
→ rpg joined13:56
→ Fandekasp joined13:56
← eriik left13:57
→ JosephSilber joined13:58
← stephCoue left13:58
→ daanavitch joined13:59
← barneybook left13:59
→ pdelvo joined13:59
← devoldmx left14:00
→ devoldmx joined14:00
→ epicowl joined14:00
→ bitkiller joined14:01
→ fitztrev joined14:01
→ pradi joined14:02
→ BSaboia joined14:02
← w0ng left14:02
→ threesome joined14:03
← pdelvo_ left14:03
← netiul left14:04
→ AWinterman joined14:04
← epicowl left14:04
→ epicowl joined14:05
← eurippa left14:05
← devoldmx left14:05
→ devoldmx joined14:05
← ayaz left14:06
← thierryp left14:07
→ jaequery joined14:07
→ huoxito joined14:08
← nicknisi left14:08
→ davidchubbs joined14:08
→ dsdeiz joined14:08
→ Stavros joined14:08
← jaequery left14:08
Stavros hello14:09
is there any set of scripts that make the git porcelain saner? something close to bzr maybe?14:09
→ jaequery joined14:09
Stavros i currently use legit, but i'm looking for something that also has revert, branching, etc14:09
→ VinceThePrince joined14:09
← devoldmx left14:10
← davidchubbs left14:10
→ devoldmx joined14:10
→ davidchubbs joined14:10
← RaceCondition left14:10
→ OOPMan joined14:12
← cirenyc left14:12
← jaequery left14:12
← subhojit777 left14:12
← pradi left14:12
→ pdelvo_ joined14:12
cmn git isn't bzr, and trying to pretend it is won't make anything saner14:13
← Macaveli left14:13
Stavros that's one way to interpret my request14:13
cmn tools like legit and hub an git-flow provide some abstraction for particular workflows14:13
if that works for you, use them; otherwise create some that fit yours14:14
← angelsl left14:14
← raijin left14:14
Stavros i already use legit, i'm looking for something that reworks more aspects of git, rather than just branching/merging14:14
→ mckendricks joined14:15
← devoldmx left14:15
→ devoldmx joined14:15
← pdelvo left14:16
cmn you'll have to define what you mean14:16
← AWinterman left14:16
charon Stavros: umm, i'm not sure what you are looking for. legit seems to have a 'legit install' way of wrapping its commands as git aliases. branching is solved in legit afaics. i'm not sure what wrapping you could do around revert, it's one of the better-defined commands14:16
Stavros something like "git uncommit", "git revert <sha>", things like that14:16
charon: isn't reverting handled by reset --soft and --hard?14:16
charon !revert14:16
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 uncommitted changes in the working tree [git reset --hard]; c) undo committing [git reset --soft HEAD^]; d) restore staged versions of files [git checkout -p]; e) move the current branch to a different point(possibly losing commits)[git reset --hard $COMMIT]?14:16
charon revert as in git-revert has a clean UI, so i don't see what you'd fix there14:17
← BSaboia left14:17
← PeterBooker left14:17
← flaguy48 left14:17
← rchavik left14:18
Stavros yeah, that's the problem: "revert" has four different meanings, for none of which git has an intuitive command14:18
albel727_ → albel72714:18
Stavros (b) and (d) could be one command14:19
(a) could be called something more intuitive14:19
(c) could be called uncommit, or rollback14:19
"uncommit" is much more intuitive than "reset --soft HEAD^"14:19
i might have to write my own aliases, but that list helps, thanks14:19
charon as for (c), welcome to pandora's box. that command only uncommits in one specific case, and does other fun things in other cases14:19
Stavros that's true :/14:19
→ flaguy48 joined14:20
Stavros i guess everything breaks if you've pushed beforehand14:20
charon (b) and (d) are actually covered by 'git checkout -f -- .' as far as they coincide, but again, they might not14:20
canton7 Stavros, you might like add -i - that's got some better nomenclature imo14:20
← devoldmx left14:20
Stavros canton7: for adding files to the index, you mean?14:20
→ devoldmx joined14:21
→ wesside joined14:21
Stavros basically, what i want is to assign more memorable names to the various commands, rather than reset --soft HEAD^14:21
← huoxito left14:21
Stavros this would probably easily be handled by a few aliases14:21
lov yes, it probably would.14:21
canton7 Stavros, if you run git add -i, you'll see it uses the terms update, revert, and 'add untracked' for manipulating the index14:22
Stavros canton7: let me try that14:22
canton7 someone did attempt to do that a while back... made a reasonable-sized project out of it14:22
charon as for the naming of e.g. revert, i think you'll just have to accept that some things are as they are. some of the terminology is drawn from p4, which isn't as well known as e.g. subversion, but refusing to adapt to this different terminology seems a silly way to proceed, since you just won't speak the same language as anyone you could turn to for help14:22
Stavros canton7: ah, right14:22
canton7: hmm, why did it grow so much?14:22
canton7 Stavros, it was almost all him iirc14:22
→ alunduil joined14:22
canton7 there also been a number of threads on the mailing list to do with reforming the UI - lots of very interesting and in-depth discussion there14:23
Stavros charon: that could probably be solved by having the tool print what it's doing14:23
← mikecmpbll left14:23
Stavros git uncommit --print: > git reset --hard14:23
then i'd know14:23
→ daniel_- joined14:23
Stavros canton7: ah, but why did it require that much code? do you know the name?14:23
canton7 Stavros, I meant 'reasonable-sized' as in 'comprehensive'14:24
Stavros oh14:24
that sounds promising14:24
charon Stavros: well, my point still stands even then. for example, (b) and (d) above are partly because many people come to git after prior svn exposure. there's really no reason why they wouldn't call it "undoing the changes" instead. not to mention that the difference between (b) and (d) is a git-specific thing.14:24
→ angelsl joined14:24
charon no reason other than 'svn revert', i mean.14:24
→ frem joined14:25
charon so if you prefer svn's language (or any other tool's language), meh; it's just different, not better, than git's terminology14:25
Stavros charon: this difference in language is just because git exposes its internals when there's no need to, though... as you said, there's no reason (b) and (d) should be so different14:25
charon mind you, this is purely about 'revert' and similar. there are *many* things wrong with git's ui14:25
Stavros charon: i see your point, i just think many of those use cases for 'revert' should have their own names14:26
at least to make them easier to refer to14:26
→ nbouscal joined14:26
Stavros the fact that git doesn't currently name them doesn't mean nobody should try14:26
→ acharlieh joined14:26
canton7 Stavros, there's http://www.saintsjd.com/2012/01/a-better-ui-for-git/, but I'm pretty sure the one I'm remembering was larger14:26
Stavros canton7: maybe http://pyrite.sophiasuchtig.com/Screenshots?14:26
canton7: ah, that post is spot on14:27
canton7 Stavros, it's probably worth going through the list - a lot of this discussion has already taken place in a lot of depth14:27
→ charliesome_ joined14:27
Stavros unstaging is another thing i'm afraid to do because reset HEAD looks scary14:27
canton7 Stavros, no, it wasn't a GUI14:27
← charliesome left14:27
charliesome_ → charliesome14:27
Stavros canton7: i imagine it would have, i'll look at that post, thanks14:27
ah14:27
→ rudak joined14:27
← nbouscal left14:28
Stavros hmm, for some reason i thought pyrite was a porcelain14:28
→ johnnyfuchs joined14:28
Stavros looks like it's a web ui14:28
charon canton7: heh, i think except for 'git retreat' (yay, he found a word that was completely unused in SCMs so far!) all of those things have been discussed on list14:28
Stavros maybe http://repo.or.cz/w/yap.git then14:28
→ nbouscal joined14:28
→ jatt joined14:28
charon in fact we have 'git stage' as a synonym for 'git add', not sure where 'git unstage' got lost14:29
→ Piggy_ joined14:30
Stavros and will git rm unstage a file?14:30
charon huh?14:30
Stavros git add stages a file14:30
git rm deletes it14:30
it's insane14:30
charon it will be removed from the index, but i wouldn't say "unstage" is a meaningful word in this sense since it is about "staging the deletion" in human terms14:30
cmn unstage can mean different things to different people14:30
Stavros cmn: like what?14:31
cmn removing completely can mean unstage, as well as resetting the state to a previous one14:31
← vpopov left14:31
Stavros "unstage" should be symmetrical to "stage", i.e. add/remove from the staging area14:31
"git stage" doesn't change the file, so "git unstage" shouldn't either14:31
cmn but stage doesn't always add to the staging area, sometimes it replaces14:32
charon well, obviously the correct definition of 'unstage', if we had it, would be 'git reset'14:32
Stavros hmm, i'm not familiar with that14:32
cmn what do you mean it doesn't change the file?14:32
← sebbi left14:32
charon i'm just wondering why 'git stage' made it into upstream, but 'git unstage' did not14:32
cmn because stage only means one thing14:32
Stavros cmn: i might be confused about what you mean by "resetting the state to a previous one"14:32
cmn doing what reset does in file mode14:32
← charliesome left14:32
→ thl0 joined14:32
Stavros right, so the state of git14:32
cmn 'of git'?14:33
Stavros rather than the file, i.e. reset --hard14:33
cmn the index, right14:33
when you stage a file, it replaces whatever state it had in the index before14:33
→ tchan joined14:33
→ ojacobson joined14:34
cmn and as charon says, the only realisitic version of unstage would be reset, at which point you can just alias it yourself if you want to use that name14:34
Stavros i'm not saying git doesn't DO these things, i'm just saying its nomenclature is nuts14:34
charon cmn: a proper git-unstage should refuse to take --mode flags, though14:34
ojacobson Stavros: "git's UI is awful" is hardly novel ground :)14:34
Stavros 90% of the insanity could probably be fixed by aliases14:35
ojacobson: haha, i'm pretty sure, i'm just pondering improvements :p14:35
← nbouscal left14:35
Stavros although i'm also pretty sure that's hardly novel ground as well14:35
ojacobson indeed :)14:35
cmn charon: that's true14:35
ojacobson I don't think the underlying problem is fixable, since the *common* language is still going to be git(1)14:35
moritz Stavros: I'm pretty sure alternative user interfaces exists. The hard part is agreeing on one, and teaching it to all the newbies14:35
canton7 it's not nuts, it's just based around internal concepts (yet ones with which most gitters are familiar). aliases only get you some of the way, since external and internal concepts don't map 1:114:35
ojacobson even if your new, improved UI is better, when it breaks, people are going to use git to fix it14:35
gan What baffles me is why --cached was kept when --staged was added. staged is a good terminology.14:35
ojacobson your alternative is to throw compat with git's porcelain out the window entirely14:36
cmn gan: backwards compatibility14:36
Stavros moritz: definitely, i'm just wondering what the most widespread/commonly agreed to be saner alternative is14:36
→ chrisshattuck joined14:36
cmn gan: you can't just remove options people are relying on14:36
gan Still, should have been caught earlier, I am sure "staging tree" and such were part of gits initial nomenclature, no?14:36
→ raijin joined14:36
charon umm, nobody calls it "staging tree" even today14:36
→ dsdeiz_ joined14:37
cmn staging tree has never been part of any nomenclature14:37
charon in fact it's an important detail in some operations that the index does not map to a tree during merge conflicts14:37
gan fine, but "staging" is used14:37
cmn yes, later14:37
the index was first called dircache, then index14:37
and you can call it staging area, as that's what you use it for14:37
gan yeah index isn't perfect either, but it's known14:37
cmn it's what we have, adding names doesn't fix it14:38
→ dvide joined14:38
← dsdeiz left14:39
→ jhoepken_ joined14:39
gan Yeah I don't really advocate adding names, that was others in this discussion14:39
→ dougsland joined14:40
← dougsland left14:40
→ dougsland joined14:40
gan cmn: but actually you can remove options, if done relatively early and you go through a "deprecated" period. But now I agree it's too late. Removing any git feature would be too intrusive.14:40
cmn you can only change the api after giving users enough time to migrate14:41
gan yes14:41
canton7 options are removed even now (branch --set-upstream is on its way out, for instance). just very carefully and very slowly14:41
← bmac left14:41
cmn which is at least several months to a year, and some things you can't change before switching to a newer major version14:41
it's on its way out after 2.014:41
gan well there you go, so --cached should be scrapped ;-)14:41
→ devth joined14:42
cmn so send the patch to deprecate it14:42
→ AWinterman joined14:42
gan no I think it's probably used too much14:42
cmn but you need to know how you want to do it14:42
so you don't want to remove it14:42
→ cirenyc joined14:42
→ lisak joined14:43
Stavros we clearly need a git200014:43
gan uh yeah, hence the three non-alphabetic characters at the end of my sentence.14:43
charon Stavros: so it seems git-unstage got lost in endless discussions about the merit. i'm not sure this is newcomer-friendly territory, but if you want to prod this again, you should probably send a summary of previous discussion, especially also what led to git-stage14:43
→ nilg` joined14:43
Stavros charon: that sounds intimidating14:43
→ Plouj joined14:43
Stavros plus, i think that's the least of our worries, really14:43
← mricon left14:43
lisak hey, I pull, resolved conflict, and suddenly I'm at this state http://i.imgur.com/N7GbrFT.png whereas my colleague at this one http://i.imgur.com/9tvM3ak.png14:43
Stavros i think this is a good starting place, at least for my use: https://github.com/saintsjd/gum14:44
Plouj shouldn't 'git hash-object .git/objects/XX/YYYY' return "XXYYYY"?14:44
lisak http://pastebin.com/Jwkt0wHE14:44
Stavros oh wait, it doesn't exist14:44
ojacobson Plouj: no.14:44
Stavros what the hell?14:44
lisak any idea how to get out of this ?14:44
Plouj If not, what command would return XXYYYY?14:44
← JosephSilber left14:44
ojacobson The files on disk are compressed. :)14:44
lisak I haven't been using git for 2 years14:44
forgot everything14:45
→ jgoss joined14:45
ojacobson Plouj: I'd use sed, if I already had the path to a loose object and wanted its hash.14:45
charon Plouj: sounds like a massive !xy problem, anyways14:45
gitinfo Plouj: 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?14:45
Plouj charon: :)14:46
ojacobson (separately: I wonder how long it takes before repeatedly git hash-object-ing the same input converges)14:46
→ Rixius joined14:46
← rath left14:46
← Eryn_1983_Fl left14:46
charon lisak: sounds like your coworker should run 'git fetch origin' and/or in fact is using a different origin repo14:46
Plouj well, git says a particular XXYYYY object is corrupt and I want to confirm and I want to be able to find a cloned copy which isn't corrupt - so I need a test command.14:46
ojacobson git cat-file TYPE HASH14:47
lisak charon, he is using the same14:47
gan lisak: What does "get out of" mean. What do you want? Looks like before the merge you locally had two commits your colleague does not (yet) have.14:47
charon ojacobson: finding a fixpoint of the hash function should be as hard as finding a collision14:47
lisak we double checked14:47
ojacobson charon: indeed, it should :)14:47
→ axl_ joined14:47
ojacobson and I'm reasonably convinced git doesn't go out of its way to wreck that property14:47
Stavros does anyone know what the easiest way to extend git with simple python script(s) would be?14:47
→ Dougie187 joined14:47
ojacobson Stavros: by doing so14:48
Stavros haha, thanks14:48
charon Stavros: write a mygit that does some work itself, but runs git <args> instead if it doesn't know about that command, then alias git=mygit in the shell?14:48
ojacobson I'm not kidding, git's invocation system is simple enough that there's nothing to it14:48
put git-foo on PATH or on git's libexec path14:48
'git foo' is now a command14:48
Stavros oh14:48
→ tjbiddle joined14:48
lisak gan, the screen shot was taken before those last 2 commits14:48
Stavros well that sounds easy14:48
thanks14:48
ojacobson for integrating with existing git bits, use libgit2 or use the subprocess module to invoke git plubming commands14:48
charon lisak: so did you fetch? i mean, there's got to be an explanation why origin/master doesn't agree. i'd start looking there14:48
gan your colleague's screen shot14:48
charon lisak: and i mean exactly 'git fetch origin', no more, no less14:49
→ marhaban1 joined14:49
charon lisak: you can also use 'git remote -v' to check the urls14:49
Stavros ojacobson: i'm just going to use git's commands, so it should be safe to run "git <whatever>", i just want to do some things like "back up files before reset --hard"14:49
lisak http://pastebin.com/kn29vtVC14:49
urls are correct14:49
charon <charon> lisak: and i mean exactly 'git fetch origin', no more, no less14:50
← devoldmx left14:50
lisak git fetch origin does the same14:50
Plouj ojacobson: Umm. I want to be able to confirm that, given a blob file, it is not corrupt. I don't think cat-file can be passed an arbitrary file for verification.14:50
charon on both repos obviously14:50
lisak I just tried14:50
→ devoldmx joined14:50
lisak status is the same14:50
→ jamesez joined14:50
Stavros charon: that's how legit works, isn't it?14:50
← sebrock left14:51
Stavros charon: although it aliases git <something> to "!legit <something>"14:51
nerdfiles legit is quite handy.14:51
← cirenyc left14:51
lisak charon, what now :-)14:51
charon Stavros: no idea. but note that ojacobson's way is just as good14:51
Stavros nerdfiles: indeed, i've found it a bit buggy, though, and the docs are frequently wrong14:51
← epicowl left14:51
nerdfiles Hm.14:51
Stavros charon: yep, i'm just trying to decide between the relative merits of the two14:51
nerdfiles Our workflows must not overlap much then. I'm still slowly bringing it into my daily routines.14:51
Stavros charon: ojacobson's way is better for single, simple scripts, your way is better for a more complicated approach14:51
nerdfiles I mostly use switch and branches.14:52
There's something else in there too, but I have way too many aliases.14:52
← alunduil left14:52
→ _BJfreeman joined14:52
Stavros nerdfiles: ah, if you try "git harvest <branch>", like in the docs, it won't work, it needs both branches14:52
← theshadow left14:52
→ bmac joined14:52
Stavros actually i see they're updated now, so they might be current14:53
→ nbouscal joined14:53
charon lisak: well, i can tell you what i would do to get back to nice history, but i'm not sure there's a point in doing that before you know why you have a different view of origin/master14:53
Stavros also graft doesn't work for me, it just leaves me on my branch, but i might need to update14:53
← AWinterman left14:53
nerdfiles I think you're right actually. That's why I've become so minimalistic with git plugins in general.14:53
→ refried_ joined14:53
lisak charon,14:53
http://pastebin.com/qpdFq81U14:53
nerdfiles I tried harvest and found myself frustrated one morning.14:53
lisak we all have this14:54
origin14:54
charon lisak: and you have all run 'git fetch origin', and you still don't have the same commit at the tip (git show origin/master)?14:54
← marhaban1 left14:54
nerdfiles We have a lot of branches in one of our repos. Many tiny bug fixes since the client has no, well, experience developing software.14:54
Stavros nerdfiles: yeah, if it worked properly it would be great, but it looks like kenneth hasn't given it the attention it needs14:54
← BJfreeman left14:54
nerdfiles Hm.14:54
Stavros nerdfiles: i've found git-flow very handy, and legit helps a lot with it14:54
but, for example, "harvest" currently does nothing for me14:55
so i just do merge --no-ff and leave it at that14:55
nerdfiles Ah, thank you !14:55
charon lisak: my money is on somebody running 'git fetch origin master' (or git pull origin master) instead.14:55
→ cilly joined14:55
lisak git fetch origin is different from git fetch origin master ?14:55
nerdfiles This is exactly it. We need a "feature" convention.14:55
lisak origin has only master14:55
charon !fetch414:55
gitinfo [!fetchfour] We recommend against using 'git fetch/pull <remote> <refspec>' (i.e. with branch argument), because it doesn't update the <remote>/<branch> ref. The easy way to fetch things properly is to get everything: 'git fetch' or 'git pull' are sufficient if you have one remote; otherwise we recommend 'git fetch <remote>' (plus 'git merge <remote>/<branch>' if you wanted to pull/merge).14:55
ojacobson lisak: yes, it is14:55
Stavros nerdfiles: how do you mean?14:55
ojacobson 'git fetch origin' updates refs/heads/origin/* (assuming you haven't changed the fetch spec)14:56
'git fetch origin master' does not, and does create/update FETCH_HEAD14:56
nerdfiles Well, I'm managing the front-end, and I often get sidebar requests.14:56
Plouj well, `git cat-file -p XXYYYY | git hash-object --stdin` unsurprisingly outputs XXYYYY so I'm out of ideas of how to test if a blob is corrupt.14:56
ojacobson if you need to care about FETCH_HEAD at all, you need to re-examine your life choices14:56
Stavros nerdfiles: oh, yeah, git-flow definitely helps with that14:56
nerdfiles So it'd be nice to have git provide a convention for me — egg-zachyly.14:56
→ tatsuya_o joined14:56
Stavros especially if you have multiple releases, you can backport patches really easily14:56
or, well, forward-port14:56
ojacobson Plouj: loose objects are zlib compressed streams14:56
the underlying stream is type, \0, contents14:57
← pitzips left14:57
ojacobson so blob\0THE FILE CONTENTS14:57
nerdfiles I like "hotfixes." Heh.14:57
ojacobson the hash of that stream should be the hash of the object14:57
charon lisak: once you get a consistent view of origin/master, there is the nice option where you'll have to re-resolve the conflict (see also git-rerere for next time). make sure you have no uncommitted changes, and a backup, then: git reset --hard HEAD^; git rebase origin/master; git push14:57
lisak: or the not-so-nice option, which is just 'git merge origin/master; git push'14:57
← abrkn left14:57
charon that's assuming you have 'git push' configured to do the right thing; otherwise use 'git push origin master'14:57
alfplayer Hello. Do you know a tool to animate graphs, to help visualize?14:58
← alorence left14:58
Stavros nerdfiles: i just work on the release branch for hotfixes, they're usually small14:58
← Rangarao left14:58
Plouj ojacobson: is "blob\0THE FILE CONTENTS" compressed, or is "THE FILE CONTENTS" the compressed part?14:58
nerdfiles Makes sense.14:58
Plouj ojacobson: I wasn't able to zcat a blob file straight up.14:58
ojacobson Plouj: the whole stream is compressed.14:59
← _BJfreeman left14:59
ojacobson it's zlib, not gzip14:59
(subtle distinction)14:59
→ _BJfreeman joined15:00
Plouj :/ I don't know of a simple program to extract those.15:01
charon python's .decode('zlib') can do it15:01
→ polymar_ joined15:01
← polymar_ left15:01
charon so you can open('foo').read().decode('zlib')15:01
← juril left15:02
charon it'll probably fail the same way git did though15:02
Plouj http://stackoverflow.com/questions/3178566/deflate-command-line-tool15:02
→ sekati joined15:02
← jatt left15:02
← devth left15:02
→ devth_ joined15:03
Plouj Oh. I thought it would just give me corrupt results which I can, hopefully, compare to the original (if I ever find it)15:03
ojacobson I appear to have lied about the header, too15:03
no, but you can re-hash the stream you get and compare it to the hash you read the stream from15:03
→ mercy____ joined15:03
→ cirenyc joined15:03
← bodom left15:03
→ BadDesign joined15:04
← BadDesign left15:04
→ BadDesign joined15:04
mercy____ morning all; i find myself routinely typing "git push origin <featurebranchname>" because my default settings for "git push" seem to push all branches i have local commits on. is there a shortcut for pushing to origin just the current feature branch i am on?15:04
ojacobson mercy____: man git-config15:04
gitinfo mercy____: the git-config manpage is available at http://jk.gs/git-config.html15:04
ojacobson have a look at push.default15:04
← devoldmx left15:05
← polymar left15:05
→ devoldmx joined15:05
← codeful left15:05
← kewah left15:06
→ fairuz joined15:06
mercy____ ojacobson: ty15:06
← Piggy_ left15:06
← axl_ left15:07
Plouj ojacobson: yeah, that's what I'm hoping to do15:08
← Kwpolska left15:08
Plouj So far, I don't have the correct command. Using a non-corrupt object I don't get the same hash with: python -c "import zlib,sys;print zlib.decompress(sys.stdin.read())" < ~/work/linux-2.6/.git/objects/22/572f036afacea9a296de2f7b0cc67541302876 | git hash-object --stdin15:09
7f4e4e9dbe8a168c1aea15fced8a9d251dcff68d15:09
→ mikecmpbll joined15:09
→ barneybook joined15:09
ojacobson oh, ick15:09
Plouj I wonder if python's print is adding an unnecessary newline15:09
→ epicowl joined15:09
ojacobson well, it is that15:09
you want sys.stdout.write15:09
→ axl_ joined15:09
← ryez left15:10
charon umm no. why are you piping the raw object, with blob <size>\0 in front, back to hash-object? you should sha1sum it15:10
← zsergua__ left15:10
← r4ven left15:10
grawity Or better, compile http://www.zlib.net/zpipe.c :D15:10
charon (the newline is spurious too, so avoid that)15:10
← cilly left15:11
→ kewah joined15:11
← devth_ left15:12
Plouj Ah, simply: `python -c "import zlib,sys;sys.stdout.write(zlib.decompress(sys.stdin.read()))" < ~/work/linux-2.6/.git/objects/22/572f036afacea9a296de2f7b0cc67541302876 | sha1sum` works!15:12
→ ntus1017 joined15:13
→ devth joined15:13
→ alunduil joined15:13
→ kumbayo joined15:13
→ pradi joined15:13
Plouj charon: I didn't know I was doing what you said, but I'd like to understand what you mean. Does hash-object need to be told what the size of the blob is?15:14
← Stavros left15:14
← acharlieh left15:14
charon Plouj: no, git stores the compressed stream "blob <size>\0<actual data>" for an object of type blob15:14
Plouj replacing sha1sm with git hash-object --stdin produces a different hash, strangely15:14
← __main__ left15:14
charon and the hash that hash-object gives you is the hash over *that*15:14
grawity Plouj: because `git hash-object` adds the header charon mentioned15:14
Plouj OOOH15:15
I just noticed "blob 221429/*" at the top of that stream...15:15
the "/*" being the start of a comment in the file15:16
→ b-rye joined15:16
Plouj ok, thanks15:16
grawity and a null byte in between, which your terminal doesn't display15:16
→ __main__ joined15:16
Plouj yes15:16
→ acharlieh joined15:16
grawity it's also why `git hash-object` has an option for specifying the type15:16
← devoldmx left15:16
gitinfo set mode: +v15:16
b-rye i have a simple git usage question15:16
→ devoldmx joined15:17
← BadDesign left15:17
b-rye I did a commit to head, then checked out to an older version15:17
_BJfreeman → BJfreeman15:17
b-rye now I'm not sure how to checkout to my last commit15:17
→ huoxito joined15:17
grawity tell it to check out the branch that you were on15:17
!detached15:18
charon b-rye: check out the branch it was on; you can probably just use 'git checkout -' which goes back to the last branch15:18
gitinfo A detached HEAD(aka "no branch") occurs when your HEAD does not point at a branch. New commits will NOT be added to any branch, and can easily be !lost. This can happen if you a) check out a tag, remote tracking branch, or SHA; or b) if you are in a submodule; or you are in the middle of a c) am or d) rebase that is stuck/conflicted. See !reattach15:18
Plouj thanks guys15:18
→ SJS joined15:18
← SJS left15:18
→ SJS joined15:18
b-rye I'm sorry, maybe I meant I committed to master15:18
but let me try that15:18
← lisak left15:19
→ AWinterman joined15:19
b-rye ah well, that's not my last commit15:20
how do I undo that, charon?15:20
charon b-rye: look in 'git reflog'; if you can't read that, pastebin it so we can help15:20
→ ErKa joined15:21
charon odds are you were already detached before the first checkout, so it's worth investigating a little to avoid the risk of losing *that* commit in the mists of unreachable history15:21
b-rye I don't see my last commit in reflog15:22
how could a commit dissapear?15:23
charon that's not possible; pastebin 'git reflog' output15:23
Plouj Humm, so I'm seeing "error: sha1 mismatch 226fa137fa7cc941e9b72a3531fe3a4bf7660050" from fsck, but checking the sha1sum manually tells that the blob is not corrupt :/15:24
charon b-rye: that reflog (of HEAD, which is what you get without args) records all movements of HEAD. by definition, a commit will be recorded in the HEAD reflog15:24
← gan left15:24
← _Qman left15:24
b-rye http://pastebin.com/uuYeStSe15:24
it's missing the commit I just made, and maybe one from yesterday15:25
→ jfinn joined15:25
b-rye maybe HEAD is a red herring15:25
I was on master15:25
← jargon left15:25
charon !HEAD15:26
gitinfo HEAD is a 'pointer' to the currently checked out branch (or commit, if HEAD is !detached). In bare repositories it serves a different function: it tells clients which branch to checkout initially after cloning. HEAD is *not* something that exists separately for every branch; that's a common misunderstanding. It also is *not* the newest commit in the repo (which is hard to define in a DVCS anyway)15:26
b-rye did a commit,15:26
← devoldmx left15:26
← devth left15:26
→ devth_ joined15:26
b-rye then switched to an old version15:26
← Sigma left15:26
→ devoldmx joined15:26
b-rye now I can't get back to my previous commit15:26
*last commit15:26
→ doug_carmichael joined15:26
charon so that reflog tells us that you committed while on master, and the commit you made last was 9b1f121. try 'git show 9b1f121' to verify, and then 'git checkout master'15:26
Nevik wanna see how bad people advise other people? http://stackoverflow.com/q/16860076/176149915:27
that's why you come to irc when you'Ve messed up your git repo15:27
← ThomasLocke left15:27
charon Nevik: sturgeon's law...15:28
← alunduil left15:28
Nevik charon: ?15:28
charon 90% of everything is crap15:28
Nevik ah15:28
→ hobodave joined15:28
Nevik sounds about right :D15:28
b-rye git checkout maser still doesn't have my latest commit15:28
Nevik love the last comment that recommends the cherry-pick the dangling commit15:28
← jiku left15:28
charon b-rye: what does !lol say on master?15:29
gitinfo b-rye: A nifty view of branches, tags, and other refs: git log --oneline --graph --decorate --all15:29
← DoniRS left15:29
← devth_ left15:29
→ devth joined15:30
← AWinterman left15:30
← acharlieh left15:30
→ thiago joined15:30
charon b-rye: and perhaps 'git reflog show master', if you mucked about with 'branch -f' or similar15:30
b-rye I don't see my changes on that, gitinfo15:30
← devth left15:30
→ devth joined15:31
charon Nevik: well, cherry-pick is probably still saner than resetting there ;)15:31
b-rye I don't see my commit on that, charon15:31
charon b-rye: gitinfo is just a bot... and if you want effective help, you'll unfortunately have to pastebin the output again15:31
Nevik charon: why? the guy wants to restore his branch :P and i think that he did more than one commit that he lost15:32
← tbosua left15:32
→ acharlieh joined15:32
Nevik charon: i probably wouldnt reset there either, but make a new branch at it, heh15:32
b-rye sry, pastebin which output?15:32
← lunarjar left15:32
charon b-rye: that funky log command, and 'git reflog show master'15:32
Nevik: because cherry-pick at least won't risk losing more commits (that are currently part of the branch)15:33
granted, there are cases where it won't do the job either, but it doesn't make it any worse at least15:33
← devth left15:33
b-rye http://pastebin.com/7XVMtaqL15:33
→ devth joined15:34
charon b-rye: well, i'm starting to get curious as to what commit you think is missing...15:34
→ pdelvo joined15:34
b-rye I just made it a few minutes ago15:34
Nevik charon: well, sicne the guy wasnt smart enough to come ask us, it's something of a moot point, heh15:34
b-rye and likely another one yesterday15:34
→ devth_ joined15:35
← devth left15:35
b-rye made it right before switching to an old version on master15:35
charon b-rye: so what does 'git show 9b1f121' say? is it not that newest commit?15:35
← devth_ left15:35
← mishok13 left15:35
grawity `git fsck --unreachable |& grep commit`15:35
charon because apart from your insistence that 9b1f121 (which is the tip of master) is not the right commit, everything adds up and exactly matches your story15:35
Nevik grawity: is that |& right?15:36
grawity Nevik: for stderr, yes15:36
Nevik ah15:36
b-rye http://pastebin.com/7c1h2htL15:36
Nevik see, didnt know thatone15:36
← pdelvo_ left15:36
grawity bash shorthand for 2>&1 |15:36
b-rye maybe I accidentally committed it to a different branch?15:37
← sekati left15:37
charon b-rye: no, the HEAD reflog would have shown that15:37
are you sure you finished committing?15:37
b-rye yes sir15:37
← devoldmx left15:37
charon go with grawity's advice and look through the unreachable commits ('git show <sha1>' each)15:37
→ lunarjar joined15:37
→ solrize joined15:37
→ devoldmx joined15:38
b-rye ok15:38
charon but i fail to see, short of a crash with ensuing failure to write the reflog completely (and i thought we synced that), how the situation could have come about ... if you really did commit15:38
Nevik grawity: it's good to know, though i probably wont remember it when i need it :<15:38
← kermit left15:39
→ scientes_ joined15:39
← scientes_ left15:39
→ scientes_ joined15:39
← j0ran1 left15:39
← yann|work left15:39
← tjbiddle left15:40
→ kermit joined15:40
b-rye sorry, I don't know what happened15:40
I did commit though15:40
→ theshadow joined15:40
b-rye didn't want to lose my changes before switching15:40
shit15:40
maybe I should have pushed too15:40
charon nothing from the unreachable commits?15:40
you can try digging up your changes from the unreachable blobs if the commits don't yield anything15:41
← marhaban left15:41
b-rye in bash, 'git fsck --unreachable |& grep commit' gives me a syntax error15:41
charon but it's a much bigger haystack15:41
← mercy____ left15:41
charon hmm, works for me... you could also use git fsck --unreachable 2>&1 | grep commit15:42
(i.e. spell out the |&)15:42
grawity: that trick is really nice15:42
b-rye I just removed the &15:42
http://pastebin.com/Pzsz4A8Y15:43
charon so now you can look at 'git show <sha1>' for both commits15:43
b-rye ok15:43
← Sonderblade left15:44
b-rye found it!15:44
grawity [ah, right, the "unreachable" lines are printed to stdout. so the |& in my example was just left over from earlier attempts with fsck -v.]15:44
← lamarus left15:44
charon then maybe do: git log --decorate --oneline --graph <that sha1> --all15:44
and show that... i'm really curious what it shows15:44
b-rye ok15:45
charon depending on the result, you'll want to merge or cherry-pick it to master to recover it, and then you're done15:45
is that a public repo/open source stuff btw?15:45
→ Macaveli joined15:46
← daniel_z left15:46
charon if so, it would be nice if you could share a tarball of your repository for further investigation. i'm a bit curious what brought this about, since it really seems like an impossible situation15:46
b-rye http://pastebin.com/749gS01L15:46
charon the reflog of HEAD *should* mention this commit15:46
err... so you ran 'git stash'?15:47
b-rye yes, then I popped it15:47
→ dreiss joined15:47
b-rye then did a commit, I believe15:47
charon well, apparently not, but at least you found the stash15:47
in that case, proceed by 'git stash apply <sha1>' instead of the other options15:47
b-rye so I didn't commit?15:47
canton7 check your shell history15:47
→ mercy____ joined15:48
← VinceThePrince left15:48
charon shrugs15:48
← d3pwnd left15:48
charon perhaps you ran 'git stash pop; git commit' and it complained because you had no staged changes. or some such.15:49
b-rye so I would do - git stash apply 10305da15:49
← narcos left15:49
charon correct15:49
b-rye ok15:49
→ aartist joined15:49
→ w0ng joined15:49
b-rye ok, sorry about that15:49
I could have sworn I did a proper commit of that15:50
comment and everything15:50
hmm15:50
charon well, it's not that bad. it would have been more frightening if git or your OS would have been so broken as to let that commit disappear.15:50
→ pro_metedor joined15:50
b-rye alright charon, thanks a great deal for helping me with that!15:51
→ juril joined15:51
canton7 maybe you committed but didn't add those changes?15:51
b-rye I don't know15:51
→ freestyl3r joined15:51
← psoo left15:52
b-rye I will pay more attention next time, and will know better too!15:52
brb15:52
← devoldmx left15:52
charon i wonder how the stash message came about, it sort of indicates that you spent more time on that than is usual... and perhaps thought it was used for a commit15:52
→ devoldmx joined15:52
→ mikewintermute joined15:53
canton7 point15:53
← relipse left15:53
→ pitzips joined15:53
← juril left15:53
→ howlymowly_ joined15:54
← thiago left15:54
← acharlieh left15:54
→ hwrd|work joined15:55
→ Chillance joined15:55
← w0ng left15:55
← kewah left15:56
→ AWinterman joined15:57
→ acharlieh joined15:57
→ alunduil joined15:57
← devoldmx left15:58
→ devoldmx joined15:58
Nevik canton7: dash15:58
b-rye yes, that sounds right15:58
I must have thought the stash operation was a commit15:58
→ Kadmos joined15:58
b-rye I was using TortoiseGit15:59
interesting15:59
← Balliad left15:59
canton7 ouch16:00
← mikecmpbll left16:00
canton7 that has a checkbox for stash iirc? very old memory there16:00
← shadok left16:00
← raijin left16:00
← fairuz left16:00
b-rye I don't think so16:00
← mercy____ left16:01
b-rye I just thought the box where I entered my comment for for committing rather than stashing, somehow16:01
Friday error16:01
← kaictl left16:01
← zoresvit left16:02
→ Adlai joined16:03
← konundra left16:03
→ _iron joined16:04
← devoldmx left16:06
→ devoldmx joined16:06
→ marshall joined16:07
marshall hey git16:07
← AWinterman left16:07
marshall I'm writing a post-receive hook and I'd like for it to check if a certain file has changed in any of the new commits. How might I go about this?16:07
→ AWinterman joined16:07
→ zoresvit joined16:07
→ kaictl joined16:08
← Macaveli left16:08
→ boombatower joined16:08
→ jdunck joined16:09
ojacobson marshall: git diff --name-only, with special cases for "new branch created" and "branch removed" cases16:09
and for root commits16:09
osse marshall: if you know the name of the file you can do e.g. if git diff -q $old $new -- path/file.txt; then blablbalba; fi16:10
with the special cases taken care of, yes16:10
← SJS left16:10
← devoldmx left16:10
marshall osse: oh, thanks!16:11
ojacobson (there's also the special-ish case of non-fastforward pushes, though it's less special than the other two)16:11
→ drizzd joined16:11
marshall are $old and $new supplied as arguments to the post-receive script automatically?16:11
ojacobson marshall: man githooks16:11
gitinfo marshall: the githooks manpage is available at http://jk.gs/githooks.html16:11
marshall I mean, environment vars16:11
ojacobson pre-receive and post-receive get the entire push list as input on stdin in a fixed format16:11
marshall ok16:12
→ SJS joined16:12
← SJS left16:12
→ SJS joined16:12
ojacobson (you only have the option of allowing or preventing a whole push, in pre-receive, whereas 'update' can reject single branches )16:12
← epitron_ left16:12
→ epitron_ joined16:12
epitron_ → epitron16:12
ojacobson (post-receive and post-update can't affect the results of a push at all)16:12
osse marshall: so start the script with e.g. read old new ref; to read stdin into those variables. In a robust script you would need to use a loop16:13
← kaictl left16:13
→ raijin joined16:14
→ japandroid joined16:14
→ sekati joined16:14
→ alansaul joined16:14
← ntus1017 left16:14
← tvw left16:14
← epicowl left16:14
← danicampa901 left16:15
alansaul Hey guys, does anyone know why I might be getting a error: fatal: The remote end hung up unexpectedly when I try push pull or even remote show origin? My origin is defined and it says origin git@github.com:SheffieldML/Saul.git (fetch)16:15
Which is my private repo, its worked in the past, just stopped working today?16:15
ojacobson [email@hidden.address] then16:16
github had some outages yesterday, there may be some hangovers or something16:16
→ noxd joined16:16
alansaul Ahh i see, just checked their status: 16:13 UTC Investigating SSH access problems on one fileserver pair16:16
Maybe I'm one of the unlucky ones16:16
→ m0viefreak joined16:17
→ bcow_ joined16:17
gitinfo set mode: +v16:17
bcow_ Hi, I forked a repo in github but forgot to set a remote upstream before I made changes and committed, now I am trying to sink with the original repo but when I do: "git remote -v" there is an "upstream" remote listed with no remote address/path it is just blank, it won't let me add a new remote named "upstream" and when I check the git.config there is no remote in there named "upstream" any ideas ?16:19
→ OliverJAsh joined16:19
← rostam left16:20
jokajak bcow_: use git remote set-url upstream <new path> ?16:20
← GodEater left16:20
← pradi left16:20
→ akiress joined16:20
→ RaceCondition joined16:20
canton7 bcow_, nothing in .git/config ? how about ~/.gitconfig ?16:20
← grizzle left16:21
→ grizzle joined16:21
→ macabre joined16:21
← sharpshooter left16:22
← andref_dias left16:22
← uidzer0 left16:22
← jceb left16:22
← ferdev_ left16:23
→ danicampa901 joined16:23
← fitztrev left16:24
← mklappstuhl left16:24
→ rostam joined16:24
← jhoepken_ left16:25
← babilen left16:25
→ GodEater joined16:25
→ ntus1017 joined16:26
→ thierryp joined16:26
← dreiss left16:26
bcow_ ah that did it jokajak ..... I thought the equiv command was --set-upstream-to which wasn't working16:26
thanks guys!16:26
→ tjbiddle joined16:26
← hobodave left16:26
← tjbiddle left16:26
← bvi_ left16:26
← RaceCondition left16:27
→ catcher joined16:27
← GodEater left16:27
→ GodEater joined16:27
← alunduil left16:27
→ kaictl joined16:27
← ojacobson left16:27
→ ojacobson_ joined16:27
ojacobson_ → ojacobosn16:27
← OOPMan left16:28
← sekati left16:28
→ influx joined16:28
← bmac left16:28
influx → pyrameter16:28
← acharlieh left16:29
← Droolio left16:30
→ acharlieh joined16:30
→ relipse joined16:30
← thierryp left16:30
← kumbayo left16:30
→ Artpicre joined16:31
→ ferdev joined16:31
→ AdrienBrault joined16:32
← axl_ left16:32
← pkrzywicki left16:33
← hwrd|work left16:33
catcher I'm hoping to convert from svn to git. My workflow involves private svn repos on a dev server with 1-4 contributors per project. They commit code back to these repos, which auto-updates a WC on the dev server, and they can obviously get each others' changes.16:33
Will a similar workflow involve me creating my own git server? Is there a better way to take advantage of git's distributed nature?16:33
→ lamarus joined16:33
← davidchubbs left16:34
DolphinDream http://ndpsoftware.com/git-cheatsheet.html :) nice16:34
ojacobosn catcher: !workflow !deploy16:34
gitinfo catcher: Finding the right workflow for you is critical for the success of any SCM project. Git is very flexible with respect to workflow. See http://sethrobertson.github.com/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.16:34
catcher: Git is not a deployment tool, but you can build one around it(in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/the-list-and-irc/deploy.html16:34
→ davidchubbs joined16:34
← LimeBlast left16:35
← leeN left16:35
therealklanni OK, so to cherry-pick a merge you need to use -m and provide the mainline number. Where do you get that information from?16:35
Also, when cherry picking a merge that replays all the commits from that merge, is that right?16:35
← Plouj left16:35
← doug_carmichael left16:36
← flaguy48 left16:36
therealklanni Or is it only recording the merge commit, not the commits that made up the merge?16:36
→ bmac joined16:36
therealklanni (it's a --no-ff merge commit)16:37
ojacobosn therealklanni: git cherry-pick -m 1 <some-merge> copies the diff from <some-merge>^1..<some-merge> and splats it out as the new commit.16:37
So it goes with -m2 (^2), etc16:37
Nothing about the merge-ness of the source commit is preserved.16:37
therealklanni So is there any point to using -m versus just cherry picking the commits themselves?16:38
ojacobosn the merge commit can have diffs not present on either side of the merge, so yes16:38
→ flaguy48 joined16:38
ojacobosn (eg. conflict resolutions)16:38
therealklanni True16:38
→ rednaks joined16:38
therealklanni Makes sense16:38
So the mainline number is how many commits I want to go backward in the merge?16:39
→ rwestphal joined16:39
→ ferdev_ joined16:39
ojacobosn no16:39
← PigeonFriend left16:39
→ dreiss joined16:40
ojacobosn When you create a merge commit C, the commit you had checked out becomes C^1, the first branch you merged becomes C^2, the second branch you merged becomes C^3, and so on16:40
man gitrevisions16:40
gitinfo the gitrevisions manpage is available at http://jk.gs/gitrevisions.html16:40
DolphinDream is there a way (terminal/gui) to get a simplified version of the repo branches ? (not include every single commit but just branch nodes and merge nodes16:40
← thl0 left16:40
grawity there are no "branch nodes"16:40
ojacobosn DolphinDream: probably something in man git-log's history simplification section16:40
gitinfo DolphinDream: the git-log manpage is available at http://jk.gs/git-log.html16:40
← ferdev left16:41
ferdev_ → ferdev16:41
DolphinDream grawity, i meant merge commits16:41
→ epicowl joined16:41
osse DolphinDream: there's the '--merges' option to git log, but it looks a little weird16:42
therealklanni ojacobosn: OK I think I understand the revision part. So when I cherry-pick the merge, and specify mainline 1, what exactly am I telling git to do? It takes the version of that file before the merge, not the result of the merge, right?16:43
← raijin left16:44
→ ryanisnan joined16:44
→ raijin joined16:44
ryanisnan → Guest2666916:44
← Artpicre left16:44
ojacobosn therealklanni: git cherry-pick in general takes the diff of some commit and creates a new commit using that diff to generate the new commit's tree16:45
← akiress left16:45
ojacobosn for "normal" commits with one parent, there's only one choice of diffs to use16:45
for merge commits, there are multiple choices: one for each parent of the merge16:45
therealklanni Right16:45
ojacobosn How those parents are numbered depends *entirely* on how you run git-merge, so I can't tell you something more generally useful than that16:46
you can see what it's going to use with git-diff though16:46
'git diff C^1..C' will tell you what 'git cherry-pick -m 1 C' will use16:46
→ w0ng joined16:46
therealklanni OK, that's good to know16:46
→ breck7 joined16:46
← Asenar left16:47
→ Artpicre joined16:47
ojacobosn for merge commits the respective parent diffs (those C^n..C expressions for various values of n) tell you how the merge result differs from each parent's immediately-prior-to-merge state16:47
← pyrameter left16:47
ojacobosn so if you create a merge commit where origin/master is C^1 and some-branch is C^2, then 'git diff C^1..C' will tell you how the merge differs from origin/master16:47
→ cjz joined16:48
cjz hi if i cherry pick from one branch to master, then later merge the branch into master, git will do the right thing right?16:48
→ deathshadow joined16:48
cjz not double merge16:48
→ dorkmafia joined16:49
jast cjz: in some cases, if other stuff in the branch modifies the same area of the file, git will see it as a conflict16:49
→ rath joined16:49
jast the only thing that is a real merge is 'git merge'... anything else is essentially the same as copying changes over16:49
and real merges are what 'git merge' is smartest about16:49
cjz hmm ok16:49
ill hold off as long as i can16:49
← rwestphal left16:50
← dreiss left16:51
← spuz left16:51
→ vpopov joined16:52
← cirenyc left16:52
← raijin left16:52
xrq` → xrq16:52
xrq is it safe to have a git repo within a repo? does git just dtrt?16:53
quazimodo hugs whoever created git bisect16:53
→ jhasse joined16:53
ojacobosn xrq: depends on how you define "right", but generally it does things that are unsurprisng16:53
← suy left16:53
← AWinterman left16:54
→ ferdev_ joined16:54
jhasse I want to rename a bunch of files, but after git-add'ing them git status shows them as new files and the old ones as deleted. How to avoid this without using git mv?16:54
→ thillux joined16:55
ojacobosn nil16:55
→ tvw joined16:55
ojacobosn git doesn't know about renames16:55
'git mv' will produce the same results16:55
(when git claims that a file was renamed, it's guessing based on content similarity)16:55
→ raijin joined16:55
ojacobosn (the threshold is adjustable, but there's no way to spell it out explicitly in the history)16:55
→ hwrd|work joined16:56
← ferdev left16:56
ferdev_ → ferdev16:56
jhasse ojacobosn, git mv results in 'renamed: a -> b', renaming them by hand and adding them doesn't16:56
ojacobosn did you also 'git rm' the originals?16:56
→ Kwpolska joined16:56
← Kwpolska left16:56
→ Kwpolska joined16:56
ojacobosn if not, do so and note 'git status' afterwards.16:56
→ etcetera joined16:56
jhasse ojacobosn, ah thx!16:56
worked thanks again :)16:57
← dsdeiz_ left16:57
→ AWinterman joined16:59
→ dsdeiz joined16:59
→ msekleta joined17:00
← jhasse left17:01
← tatsuya_o left17:02
← DolphinDream left17:02
← angelsl left17:03
← Artpicre left17:04
→ homer91 joined17:04
← lahwran left17:04
→ konundra joined17:04
← epicowl left17:04
→ momomomomo joined17:04
← relipse left17:05
→ lahwran joined17:05
→ etcetera_ joined17:06
→ system64 joined17:06
← etcetera left17:08
etcetera_ → etcetera17:08
→ sekati joined17:08
causasui_ → causasui17:08
← aartist left17:08
← dhruvasagar left17:08
← breck7 left17:08
← Dougie187 left17:10
→ Dougie187 joined17:10
← ferdev left17:10
→ akiress joined17:10
← pdelvo left17:11
→ ferdev joined17:11
← brocken left17:12
← davidchubbs left17:12
→ davidchubbs joined17:13
← aendrew left17:13
→ pr0ton joined17:13
→ devoldmx joined17:15
← pro_metedor left17:15
→ Sigma joined17:15
→ SrPx joined17:16
SrPx How exactly do I fix something in someone else's project? Should I fork the project, fix it, then what? Or just clone the original project, modify stuff and pull?17:16
jast 1) pull is for getting updates *from* somewhere. 2) in case you're talking about github, they have a feature called 'pull request' that you can use to ask someone to merge a change you made in your fork.17:17
→ Artpicre joined17:17
← Dougie187 left17:18
→ Dougie187 joined17:18
→ marhaban joined17:19
SrPx I see, so I shouldn't clone the original project at all?17:19
bambams You would 'push' if you had push access to their repository, but most likely you won't so instead you publish your repository and then submit a pull request (as jast poitned out).17:19
→ markovh joined17:19
mokau What do you do when git hangs on something like "Auto-merging eventcalendar/admin.py"17:19
Do I just wait?17:19
→ stliu joined17:19
jast mokau: discreetly check if the world outside still exists17:19
bambams SrPx: Yes, you would clone their repository. If using GitHub, "forking" their project basically does that too.17:19
jast more helpfully, abort the merge and retry, I suppose17:20
→ rakl joined17:20
SrPx bambams: hm you are confusing me, I cloned it but I had no fork of it on my account. So I just pressed the Fork buttom on the site and cloned my fork..17:20
bambams SrPx: What are you confused about? Sounds like you got it. :)17:20
SrPx nothing then, thanks17:21
→ marhaban1 joined17:21
← pmdarrow left17:21
← zoresvit left17:22
← Artpicre left17:22
bambams mokau: ^C, git reset --hard (will throw away changes in your working directory and reset you to the original branch head), and then try again. Or if you're in a rebase, git rebase --abort. Or if in a cherry-pick, git cherry-pick --abort. Or, just wait. :)17:22
→ jceb_ joined17:23
→ Artpicre joined17:23
bambams It probably shouldn't hang though. Could there be other reasons for it being so slow, like operating over Samba or something? :)17:23
← marhaban left17:23
← sgambino left17:23
← BJfreeman left17:24
← AdrienBrault left17:24
← devoldmx left17:24
→ devoldmx joined17:24
→ tmcmahon joined17:24
→ sgambino joined17:25
→ pdelvo joined17:25
ojacobosn → ojacobson17:25
← mikewintermute left17:26
← dorkmafia left17:26
→ mikewintermute joined17:26
mokau man merging is just not working...17:27
it just hangs no matter what I do17:27
← Artpicre left17:27
→ Artpicre joined17:27
← alansaul left17:27
→ babilen joined17:28
← achadwick left17:28
mokau I can do "git diff HEAD..branch" and it shows me the differences17:28
which are ridiculously simple17:28
why wouldn't it be able to merge17:28
← pitzips left17:28
← Artpicre left17:29
→ Artpicre joined17:30
mokau Even if I reclone17:31
and try to merge...17:31
bambams mokau: Can you share the repository?17:32
← Artpicre left17:32
→ dreiss joined17:32
mokau bambams: unforutnately not :(17:32
bambams mokau: What version of Git, platform, etc.?17:33
mokau Is there a way to turn on profuse debugging information?17:33
→ alunduil joined17:33
← noxd left17:33
→ Artpicre joined17:33
mokau Ubuntu 12.04 Git 1.7.10.417:33
bambams Maybe try setting GIT_TRACE and GIT_MERGE_VERBOSITY? :-/17:34
→ DolphinDream joined17:34
→ normanrichards joined17:34
mokau aha!17:34
Its trying to use sublime text as the git merge editor17:34
lmao17:34
bambams xD17:34
mokau And I use virutal dekstops so I didn't see it poping up on a different screen17:35
→ fitztrev_ joined17:35
bambams It's a good day when it turns out to be something simple. :)17:35
← devoldmx left17:35
← marshall left17:35
→ devoldmx joined17:36
← hwrd|work left17:37
mokau bambams: thanks for the moral support17:37
→ uidzer0 joined17:37
→ txomon|fon joined17:38
→ LuxuryMode joined17:38
txomon|fon anyone knows what to do when git gives Out of memory while packing an object? I have a 4GB image I want to versionize...17:38
← dwmw2_gone left17:39
→ dsdeiz_ joined17:39
← dsdeiz left17:39
← w0ng left17:39
Kwpolska txomon|fon: get more RAM?17:40
txomon|fon done17:40
and if it is not enought?17:40
Kwpolska txomon|fon: honestly, git sucks at binary files.17:40
← aspiers left17:40
→ DolpheenDream joined17:41
txomon|fon any idea on how to versionize a filesystem?17:41
← tommick left17:41
jast there are filesystems that support versioning, e.g. ext3cow, zfs, btrfs, ...17:41
→ nebuchadnezzar joined17:41
Kwpolska txomon|fon: not do it? you are likely to need 4gb/version anyways, unless you can do delta.17:41
→ breck7 joined17:42
nebuchadnezzar Hello17:42
jast and of course there's always rdiff-backup17:42
jaalto How do I stash only 2 files from 5 modified?17:42
txomon|fon rdiff-backup?17:43
txomon|fon is looking at that17:43
jast jaalto: git stash save has a -p option to pick individual hunks to stash17:43
← rakl left17:43
jaalto jast: Huh, by hunk basis?17:44
jast I believe that's what I said, yes17:45
→ rakl joined17:45
← rudak left17:45
Kwpolska jaalto: “With --patch, you can interactively select hunks from the diff between HEAD and the working tree to be stashed.” --man git-patch17:45
→ Macaveli joined17:45
← Macaveli left17:45
← rakl left17:45
Kwpolska git-stash*17:45
→ Macaveli joined17:45
txomon|fon jast, but I need to versionize!17:45
→ rakl joined17:46
Kwpolska txomon|fon: !xyproblem?17:46
!xy17:46
gitinfo 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?17:46
txomon|fon I am trying to versionize a whole system image17:46
→ noxd joined17:46
Kwpolska txomon|fon: why?17:46
← Katafalkas left17:47
← jkremser left17:47
← Macaveli left17:47
SamB what is this "versionize" thing17:48
← Artpicre left17:48
txomon|fon Kwpolska, because I am doing system images, and I have to distribute them17:48
Kwpolska txomon|fon: and?17:48
txomon|fon SamB, introduce in a version control system (for me)17:48
bambams Binary generally doesn't work well with version control.17:48
txomon|fon and I want to have the different versions I changed17:48
Kwpolska txomon|fon: ftp/http server with images is the way to go.17:49
← didlix left17:49
txomon|fon for that, I use dd, and put git over the img, which is exactly what I am trying to do17:49
→ JosephSilber joined17:49
← devoldmx left17:49
← nkambo left17:49
txomon|fon but I would prefer putting all the system under git17:49
← liquid-silence left17:50
← mckendricks left17:50
→ codeful joined17:50
→ devoldmx joined17:50
ojacobson txomon|fon: are these images you can build from other files already under git?17:50
versioning generated outputs using an SCM tool is generally a poor practice, regardless of tool, is why I ask17:50
SamB wouldn't you want some kind of snapshotted, possibly de-duplicated, FS for this?17:50
→ h0bbit joined17:50
ojacobson (keep what you need, regenerate the rest, archive the builds if you want to save time later)17:51
→ mckendricks joined17:51
txomon|fon ojacobson, the problem is that many things are picked from here and ther17:51
Kwpolska txomon|fon: this is still nonsense. github will kill you, bitbucket will too. You will need lots of RAM to get it into git.17:51
← yacks left17:51
SamB fossil/venti comes to mind, though I may have misspelled it17:51
→ grizzle_ joined17:51
txomon|fon Kwpolska, I am using my own server, not public17:51
→ Artpicre joined17:52
Kwpolska txomon|fon: does your server have tons of free space?17:52
txomon|fon Kwpolska, it does17:52
the idea of saving the image is to be able to know the differences17:52
that is why I wanted to save it in fs format and not in binary format17:52
← grizzle left17:52
txomon|fon the rdiff-backup is great17:53
Kwpolska txomon|fon: what is “fs format”?17:53
← DolphinDream left17:53
DolpheenDream → DolphinDream17:53
← Artpicre left17:53
→ DolpheenDream joined17:53
txomon|fon Kwpolska, filesystem format17:53
→ dhruvasagar joined17:53
catcher If I want 4 devs collaborating on a project independently, I need them to push code to a central repo of some sort, right?17:53
→ Artpicre joined17:54
txomon|fon now I am searching all the hardlinks I can and substituting them for symlinks17:54
→ relipse joined17:54
txomon|fon catcher, independently and central repo... are opposite there17:54
Kwpolska catcher: you can make a huge mess by pushing/pulling from each other instead17:54
ojacobson catcher: !workflow17:54
gitinfo catcher: Finding the right workflow for you is critical for the success of any SCM project. Git is very flexible with respect to workflow. See http://sethrobertson.github.com/GitBestPractices/#workflow for a list of references about choosing branching and distributed workflows.17:54
catcher ojacobson, been reading those links, just looking for a sanity check.17:55
← Artpicre left17:55
txomon|fon another way, following ojacobson's idea would be to create a sort of script that reproduces all steps...17:55
ojacobson framing it as "I need" ensures that the answer is "no" no matter what you ask, around git workflows17:56
txomon|fon but it is insane17:56
→ Artpicre_ joined17:56
ojacobson git's too flexible to force a specific workflow on users17:56
txomon|fon: that's called a "build script" and it's a necessary part of most software projects17:56
catcher Kwpolska, for a very basic workflow where 4 devs contribute to a main line of code with occasional branches, does it make sense for us to push to a bare repo on a central server? Or am I missing the point of being distributed that way?17:57
txomon|fon ojacobson, but it would be a build-configure-install-create-andtoomanythings script17:57
ojacobson create scripts for each of those stages17:57
create driver script that ties the stages together17:57
ship it17:57
Artpicre_ → Artpicre17:57
Kwpolska catcher: most people do it this way, and most people do github.17:57
ojacobson (Obviously, I'm not you and I've got no visibility into your project, but my general attitude towards builds is "fucking automate it" because people are unreliable)17:58
Kwpolska (unless, of course, you don’t do open source)17:58
ojacobson (at length: http://grimoire.ca/dev/builds)17:58
→ cirenyc joined17:58
→ d3pwnd joined17:58
← devoldmx left17:58
→ devoldmx joined17:58
txomon|fon the good thing is that it is a controlled environment...17:59
← dr_lepper left17:59
← d3pwnd left17:59
txomon|fon but is just for testing, we will drop this platform in 5 weeks17:59
→ dsdeiz joined17:59
Kwpolska ojacobson: computers can sometimes go apeshit when it comes to builds. Which reminds me I should do an automated version number generator. I once turned 3.0.1 into 3.0.l (three.nought.ell — blame consolas) and 2.1.3.4 into 2.1.3.35.17:59
s/builds/builds, too/17:59
→ marshall joined18:00
← nilg` left18:00
catcher Kwpolska, much appreciated. Last q - with private projects, any reason to use github over a bare repo on a private server?18:00
← pr0ton left18:00
← cirenyc left18:00
txomon|fon catcher, because it has a very nice graphical interface and lots of tools for your project18:00
Kwpolska catcher: “private” = FLOSS but nobody is going to use it? go github.18:00
← dsdeiz_ left18:01
→ epicowl joined18:01
→ pr0ton joined18:01
← etcetera left18:01
← ferdev left18:02
→ dr_lepper joined18:02
→ ferdev joined18:03
← dr_lepper left18:04
→ dr_lepper joined18:04
← alunduil left18:04
→ etcetera joined18:05
← epicowl left18:05
→ eijk_ joined18:06
← tmcmahon left18:07
← konundra left18:07
→ konundra joined18:07
← dijonyummy left18:08
← rakl left18:08
← jonconley left18:08
← tinti left18:09
← momomomomo left18:09
→ beaky joined18:09
← etcetera left18:10
← eijk left18:10
→ douglaslassance joined18:10
catcher txomon|fon, Kwpolska private meaning the code shouldn't be available to the public, and I have ~30 repos. github could get a little expensive, but probably worth it18:10
→ momomomomo joined18:10
→ cirenyc joined18:11
← khushildep left18:11
beaky are there alternatives to github for hosting repos?18:11
ojacobson bitbucket18:11
rath bitbucket.org18:11
bambams Arguably all code should be available to the public. In any case, Bitbucket offers free private repos.18:11
beaky wow thanks guys18:11
← grizzle_ left18:12
Kwpolska bitbucket is uglier and less functional IMO, but to each their own…18:12
gitinfo set mode: +v18:12
DolpheenDream so.. how do i list the repo graph with only the branching and merging transitions (no other commits)18:12
beaky wow this is perfect for uni code projects18:12
→ grizzle joined18:12
DolpheenDream i just want to see a brief illustration of how branches are related18:12
→ ferdev_ joined18:13
→ etcetera joined18:13
← devoldmx left18:14
← ferdev left18:14
→ PaulCapestany joined18:14
→ devoldmx joined18:14
ojacobson DolphinDream: git does not know where one branch separated from another18:15
→ adiabatic joined18:15
ojacobson play with the history-simplification options for git log, though, and see if one of them gives you what you want18:15
→ _BJfreeman joined18:16
→ epicowl joined18:16
← ferdev_ left18:17
← majuscul1 left18:17
← albel727 left18:18
→ ferdev joined18:18
DolpheenDream ojacobson, thx18:18
→ albel727 joined18:19
ojacobson EugeneKay: "caution: this end upstream"18:19
_BJfreeman → BJfreeman18:19
→ Shusshu joined18:20
← Shusshu left18:20
DolpheenDream ojacobson, i was looking for the typical -graph log but one that has the intermediary commits on a branch reduced so taht the graph is not as tall18:20
← dhruvasagar left18:21
← pr0ton left18:21
← devoldmx left18:22
← DolphinDream left18:22
→ devoldmx joined18:22
→ dijonyummy joined18:22
← dijonyummy left18:23
← solrize left18:24
← japandroid left18:25
→ kadoban_ joined18:25
→ dorkmafia joined18:26
→ pmdarrow joined18:26
→ skelterjohn joined18:27
← kadoban left18:27
← eijk_ left18:27
← jason237 left18:28
→ eijk_ joined18:28
offby1 DolpheenDream: for "gitk", you can hit F4 and then choose "simple history". I think this corresponds to "--simplify-by-decoration".18:28
→ felds joined18:29
← pdelvo left18:31
← grizzle left18:32
← eijk_ left18:32
→ didlix joined18:32
→ grizzle joined18:32
← scampbell left18:33
→ pdelvo joined18:33
→ felds_ joined18:33
→ eijk_ joined18:33
← felds left18:35
→ alvaro_o joined18:35
← normanrichards left18:36
← alvaro_o left18:37
→ jcromartie joined18:37
← eijk_ left18:37
→ eijk_ joined18:38
← _goTAN left18:39
→ netiul joined18:40
← thillux left18:40
← tsdh left18:40
DolpheenDream thx offby118:41
← jceb_ left18:41
→ andref_dias joined18:42
→ eijk joined18:42
← dorkmafia left18:42
← davidchubbs left18:43
→ dhruvasagar joined18:43
← daniel_- left18:43
← eijk_ left18:44
→ jarray52 joined18:44
→ sq-one joined18:45
jarray52 when using a git merge tool, what is the difference between files marked LOCAL and files marked BASE? (I'm using meld)18:46
← epicowl left18:46
← grizzle left18:46
← beaky left18:47
← nerdfiles left18:47
← eijk left18:48
→ adam-_-_ joined18:48
→ maqr joined18:49
offby1 I think BASE is the commit that's common to both branches18:49
→ eijk joined18:49
offby1 and LOCAL is the branch that you had checked out before you said "git merge"18:49
← jimi_ left18:49
offby1 that leaves REMOTE as the branch whose name you passed to "git merge"18:49
just guessing but I suspect I'm right18:49
jarray52 So, BASE is the version where the fork in the code occurred?18:50
And, LOCAL and REMOTE are the two branches being merged?18:51
← relipse left18:51
← codeful left18:51
← eijk left18:52
← herlo left18:52
→ pdelvo_ joined18:52
→ eijk joined18:52
← devoldmx left18:54
← curvedmark left18:54
→ devoldmx joined18:54
→ normanrichards joined18:55
← eijk left18:55
← pdelvo left18:56
→ mcwulf2 joined18:56
→ eijk joined18:56
→ Vert joined18:58
jarray52 offby1: Thanks.18:58
← h0bbit left18:58
← mcwulf2 left18:58
→ neilk_ joined18:58
← neilk_ left18:58
→ neilk_ joined18:58
← eijk left18:59
neilk_ Is it fair to say that git subtree has made tools like braid mostly obsolete?18:59
→ h0bbit joined18:59
→ jceb_ joined19:00
→ eijk joined19:00
← homer91 left19:01
→ homer91 joined19:01
→ pr0ton joined19:01
→ rakl joined19:02
← imMute left19:03
← eijk left19:03
← kadoban_ left19:03
← rakl left19:03
→ eijk joined19:03
← jceb_ left19:04
← sampiero left19:05
← freestyl3r left19:05
← system64 left19:06
← pr0ton left19:06
→ knite joined19:06
← eijk left19:06
→ eijk joined19:07
← marthinal left19:07
← devoldmx left19:08
→ devoldmx joined19:09
→ rudak joined19:09
← eijk left19:10
→ eijk joined19:10
← rudak left19:11
→ tureba_ joined19:11
→ lb joined19:11
← lb left19:11
→ lb joined19:11
→ infynyxx joined19:12
→ imMute joined19:12
← imMute left19:12
→ imMute joined19:12
→ Cykey joined19:12
← devoldmx left19:12
→ devoldmx joined19:12
→ jonconley joined19:13
← homer91 left19:13
← eijk left19:13
← tureba left19:14
→ nkambo joined19:15
→ eijk joined19:15
→ homer91 joined19:15
← Guest26669 left19:17
← etcetera left19:17
→ alunduil joined19:18
← esc left19:18
← jcromartie left19:19
← eijk left19:19
← tureba_ left19:19
→ jceb joined19:19
← knite left19:19
→ eijk joined19:19
→ etcetera joined19:20
→ tureba joined19:20
adhawkins-away → adhawkins19:21
→ poseidon joined19:22
→ devth joined19:22
→ w0ng joined19:22
poseidon I want to migrate from svn to git. Is there a way to import all of the svn logs into the git repo?19:22
← m8 left19:22
← eijk left19:22
→ pinage404 joined19:23
arand !svn19:23
gitinfo [!svn_upstream] git-svn determines the upstream branch to push to by finding the topmost commit in your current branch that has a "git-svn-id" line (i.e., 'git log --grep=^git-svn-id: --first-parent -1'), and using the branch mentioned in it.19:23
→ eijk joined19:23
ddv poseidon: https://github.com/nirvdrum/svn2git19:24
→ mniip joined19:24
arand poseidon: svn2git for a one-time offline conversion, git-svn for continous two-way bridging.19:24
← h0bbit left19:25
poseidon arand: thanks19:25
mniip git just barked at me for not having user.email set, so i went on and have set it, but commit --amend set the author email to something even more weird, which isn't even what my hostname is,19:26
though reset-add-commit worked fine19:26
← neilk_ left19:26
← pinage404 left19:26
← eijk left19:26
→ Droolio joined19:27
← Droolio left19:27
→ lacrymology joined19:27
→ grizzle joined19:27
→ codeful joined19:27
lb mniip, git commit --amend ..reset-author19:27
i meant --19:28
mniip, git commit --amend --reset-author19:28
← Tuna-Fish left19:28
→ codeful01 joined19:28
mniip ah19:29
ok19:29
github tut didn't mentioned that, though19:29
→ pr0ton joined19:29
← w0ng left19:30
← Akufen left19:30
mniip didn't mention*19:30
→ Tuna-Fish joined19:32
← danicampa901 left19:32
→ dmarkham_ joined19:32
← jceb left19:32
→ pinpin404 joined19:32
← codeful left19:33
→ nicedice joined19:33
→ tplaner joined19:33
← nbouscal left19:34
← infynyxx left19:34
← ferdev left19:35
→ pitzips joined19:35
skelterjohn starting from scratch, is there a way to get only a single commit from a remote git repo?19:35
that is, don't download the whole history19:35
ojacobson skelterjohn: "yes, but"19:36
"but the resulting repo can't push or be cloned", "but it has to be a branch tip or a tag", "but an bunch of other limitations"19:36
lov Oh, you can't do it for arbitrary commits?19:36
skelterjohn well, the two limitations you mentioned are acceptable19:37
can i look somewhere to see the rest of the limitations?19:37
ojacobson I don't believe so, no19:37
man git-clone19:37
gitinfo the git-clone manpage is available at http://jk.gs/git-clone.html19:37
ojacobson read the bits about shallow clones19:37
skelterjohn ok, thanks19:37
← b-rye left19:37
← devoldmx left19:37
→ devoldmx joined19:38
→ aartist joined19:38
sonOfRa how do I change the email adress only for a certain repo? I tried `git config --local user.email="my@other.email`, but looking at $reporoot/.git/config doesn't show any changes19:38
do I have to apply this manually to the config each time?19:38
← quetz left19:39
← cortes left19:39
← felds_ left19:39
therealklanni Each time you clone it?19:40
EugeneKay You're setting it wrong19:40
man git-config19:40
gitinfo the git-config manpage is available at http://jk.gs/git-config.html19:40
EugeneKay `git config <key> <value>19:40
→ japandroid joined19:41
EugeneKay No = involved19:41
→ infynyxx joined19:41
← mniip left19:41
sonOfRa oh19:41
→ sonOfRa joined19:42
sonOfRa thanks, EugeneKay19:42
→ btanaka joined19:43
→ epicowl joined19:43
← normanrichards left19:43
← pmdarrow left19:44
→ jason237 joined19:44
← mikewintermute left19:45
→ felds joined19:45
← ntus1017 left19:45
→ nerdfiles joined19:45
← noxd left19:46
← dmarkham_ left19:47
→ dmarkham_ joined19:47
→ BadDesign joined19:47
→ b-rye joined19:47
→ Davey joined19:48
← b-rye left19:48
→ rurufufuss joined19:51
→ Prazanna joined19:52
→ _BJfreeman joined19:52
← maslowbeer left19:52
← rath left19:52
→ solrize joined19:53
← lb left19:53
→ mencito joined19:53
← marhaban1 left19:53
← BJfreeman left19:54
_BJfreeman → BJfreeman19:54
← didlix left19:55
→ ferdev joined19:57
→ marhaban joined19:57
← Jackneill left19:58
← devoldmx left19:58
→ devoldmx joined19:58
→ therealGent joined19:58
→ normanrichards joined19:58
← bcow_ left20:00
→ Stavros joined20:00
Stavros hello20:00
← therealGent left20:00
Stavros does anyone know how i can perform the equivalent of git rm --cached with GitPython?20:00
← devoldmx left20:01
→ devoldmx joined20:01
← acharlieh left20:01
→ tommick joined20:01
→ acharlieh joined20:02
→ PleasureTown joined20:02
← acharlieh left20:03
→ therealGent joined20:03
← PleasureTown left20:03
→ PleasureTown joined20:05
gitinfo set mode: +v20:05
→ Chikore joined20:05
← juanmabc left20:05
PleasureTown Hey internet friends20:05
→ jceb joined20:05
PleasureTown Could anyone help me with a post-recieve hook question by chance/20:06
← devoldmx left20:06
→ tinti joined20:06
← Stavros left20:07
→ tjbiddle joined20:07
← alunduil left20:08
tjbiddle Anyone here have a Gitlab installation? Been having trouble paginating commits for their API - I get a 500 error on two entirely different installations. The docs say I should be able to pass per_page and page as GET arguments - Even used a ruby wrapper for the gitlab api and run into the same issue. Someone mind testing on theirs?20:09
HedgeMage PleasureTown: Just ask, don't ask to ask...it's rude20:09
PleasureTown hah, thanks for the notice20:10
tjbiddle HedgeMage: Well, I wouldn't say rude - Just not necessary. IRC is for community help and collaboration so it goes without saying all are welcome to ask questions :-)20:10
PleasureTown I've created a post-recieve hook from ruby to move the files to a dir /etc/puppet , but I can't tell if it's working. I made sure to chmod 755 it, but I can't say I know what I'm doing20:11
tjbiddle PleasureTown: Came in late - What exactly are you trying to do here?20:12
→ Stavros joined20:12
tjbiddle I work with Puppet daily - Might be of some help.20:12
Stavros hello20:12
→ acharlieh joined20:12
tjbiddle Hi Stavros !20:12
Stavros i am creating a git alias, called mygit, that calls git if it doesn't find a command20:12
→ aspiers joined20:12
Stavros i'm guessing it will recurse if i alias mygit to git, huh20:12
→ nbouscal joined20:12
← pdelvo_ left20:13
→ jhoepken_ joined20:13
EugeneKay Probably.20:13
→ pdelvo joined20:13
HedgeMage tjbiddle: Meh, I'm pretty old-school -- I dislike inefficient behaviors in large channels. See the first bullet point under "Habits": http://binaryredneck.net/support-leech/20:13
PleasureTown tjbiddle: I just got started with Puppet and figured the first step was to put the configs under VCS. So I created a repo in /opt/puppet and set up a ruby script as a post-recieve hook to move the files to /etc/puppet/environments/production but i'm not sure how to test if it's working. It's not moving any files, so I suspect not.20:13
← etcetera left20:14
→ LucasTizma joined20:14
← japandroid left20:14
← whoami left20:15
tjbiddle HedgeMage: Heh - I can understand. But I think we're all there at some point - especially when learning something new. Most will have questions but don't have anything to contribute back until they find their way around.20:15
PleasureTown Also, what are some good mIRC addons. I haven't used mIRC in 15 years...20:15
EugeneKay The rest of us are just experts in everything.20:15
PleasureTown :)20:15
EugeneKay HexChat is a great addon20:15
If by addon you mean throw it the fuck out.20:15
PleasureTown haha20:16
ojacobson PleasureTown: zero'th suggestion: put something unconditional at the top of your hook that spews out information about the hook20:16
see if that output appears on push20:16
if not, no hook.20:16
EugeneKay Because mirc is, lol, mirc.20:16
HedgeMage tjbiddle: Absolutely. It's simply my philosophy that acculturating newbs to the old-school way of handling things makes them easier and more pleasant to support, which in turn gives them a better shot at getting meaningful help. Then, as they "grow up" in the tech in question, they can pay it forward.20:16
→ etcetera joined20:16
PleasureTown ojacobson: what do you recommend?20:16
← tplaner left20:16
HedgeMage tjbiddle: Hacker culture and technical skill compliment and enhance one another.20:16
→ faen joined20:16
ojacobson echo "Starting $0"20:16
tjbiddle PleasureTown: Check out librarian-puppet - It will make things a lot easier down the road. What I do is I have a repository that holds my librarian-puppet info - It has a "Puppetfile" much like a Gemfile in ruby where you specify dependencies - In this case - Puppet modules. You can store each module in it's own repo so it stays isolated and clean - And run librarian-puppet update when ready to push changes (You'd just need to do a git pull on the repo to get t20:16
new Puppetfile first - Which you can either do manually, or have cron do it)20:16
→ OnlyMax joined20:17
HedgeMage anyway, back to coding20:17
tjbiddle Or, you could setup a post-receive hook to do that as well.20:17
→ whoami joined20:17
← SrPx left20:17
PleasureTown hah, I'll take it slow and do the basics right now20:17
tjbiddle Just have it ssh, run git pull, run librarian-puppet update - voila you have your puppet manifests set to go :-) Personally I'd do it in a Jenkins job or something so you have more granular control of when you want to push it out.20:17
← etcetera left20:18
tjbiddle PleasureTown: Well let me know if you run into any issues - And #puppet is pretty friendly as well!20:18
faen → farn20:18
← jday_ left20:18
tjbiddle Just don't try to overcomplicate things if you don't need to - as with anything.20:18
ojacobson PleasureTown: I have https://github.com/ojacobson/gitolite-local-code/blob/master/hooks/common/post-receive.d/50-run-command set up on my puppet repo, rigged (via git-config) to run a script that updates the puppet manifests/modules/files dirs20:18
tjbiddle HedgeMage: Godo point :-)20:19
→ _doug joined20:19
← LucasTizma left20:19
_doug $Id$ .... does git automagicall replace that on commit?20:19
HedgeMage tjbiddle: thanks :D20:19
PleasureTown interesting20:19
tjbiddle _doug: $ld$ ? Where are you seeing that?20:19
→ LucasTizma joined20:19
ojacobson The actual update script (/usr/local/bin/pupdate) is https://gist.github.com/ojacobson/71ebabbed59dd410e10220:20
→ LNDN joined20:20
ojacobson I should replace that 'git pull' with something saner20:20
← jceb left20:20
_doug tjbiddle http://git-scm.com/book/ch7-2.html20:20
www.g20:20
oops20:20
→ jjasinski joined20:21
ojacobson _doug: it's possible to set up sccs/rcs-style markers like $Id$, but no, Git doesn't support them on its own20:21
that is rather the _point_ of the page you linked, though, so I'm not sure where your confusion is coming from20:21
_doug tjbiddle: Context is that I want to replace a value in a source file in checkin so that the script can report it's version number20:21
PleasureTown ojacobson: I see the issue. It appears to be executing my script as bash and not ruby20:22
ojacobson Then the shebang is wrong20:22
git just calls exec..() eventually20:22
← jgoss left20:23
PleasureTown hmm, i have #!/usr/bin/env ruby20:23
charon _doug, ojacobson: note that the dater used there is even subtly wrong in some cases. if you check out across commits that happen to share a blob, but that each modified this blob (e.g. if you jump between rebased versions of the same topic this can easily happen) git will *not* re-checkout the file, thus the dater never gets a chance to look at it again.20:23
EugeneKay Ah, /usr/bin/env.20:23
_doug charon: The quantum effect20:23
PleasureTown should probably just /usr/bin/ruby eh20:23
← sgambino left20:24
charon _doug: anyway, the real answer is to use a build rule that substitutes `git describe --always` somewhere20:24
EugeneKay Give it a try20:24
← JosephSilber left20:24
_doug I'll have script read source file... use that... and have that get thrown into the build package20:24
→ _BJfreeman joined20:24
charon _doug: e.g. you could look at GIT-VERSION-GEN in git.git20:24
and the associated Makefile rules20:25
← jfinn left20:25
_doug charon: Yah but in actuality, git id's suck. A integer build number is much more useful20:25
charon basically the idea is to have the packaging scripts package a little version file to be used in case you build from package instead of from repo20:25
PleasureTown No love Eugene. Should I threaten the script with a knife?20:25
→ pantsman joined20:25
← pantsman left20:25
→ pantsman joined20:25
charon _doug: well, whatever floats your boat :)20:25
_doug charon: *gets floaty*20:26
EugeneKay Drink a beer, then shatter the bottle and hold the broken-off neck threateningly.20:26
PleasureTown Hah, just poured one. Spent the day fixing a nortel phone system and a Domino server. It felt like the early 90's all over again20:26
← marshall left20:26
← vpopov left20:26
ojacobson oh nortel20:27
PleasureTown Yeah, set up by an idiot. It's a nightmare20:27
tjbiddle Anyone here have a Gitlab installation? Could use help trying to troubleshoot if an API call is only happening on my installations or not.20:27
← BJfreeman left20:27
PleasureTown Nortel in Austin, TX goes out, so do the nortels in Sao Paolo and Belgrade...20:27
Thanks whoever set this up. Pro job20:28
→ pppaul joined20:28
EugeneKay Does it work if you make a bash script that does `echo "Butts lol"` ?20:28
← Stavros left20:28
PleasureTown lol20:28
→ jfinn joined20:28
← _BJfreeman left20:29
→ _BJfreeman joined20:30
→ btree joined20:31
EugeneKay Serious question20:32
PleasureTown 10 inches20:32
→ Samual_ joined20:32
← Samual_ left20:32
→ Samual_ joined20:32
← chamunks left20:32
→ chamunks joined20:32
← btree left20:32
→ daniel_- joined20:33
← pppaul left20:34
→ styx_ joined20:35
→ destiny joined20:35
PleasureTown I'm beginning to think it's my ruby Eugene20:35
← Samual left20:35
← destiny_ left20:35
← AlekxE left20:35
← styx__ left20:35
← diegoviola left20:35
gitinfo set mode: +v20:35
← Klowner left20:35
← Zeelot3k left20:35
← bgerber left20:35
← w0m left20:35
← jexmex left20:35
← chuckf left20:35
← bpayton left20:35
→ dwmw2_gone joined20:35
→ _AlexE_ joined20:35
→ w0m joined20:35
→ Klowner joined20:35
→ chuckf joined20:35
→ Zeelot joined20:36
→ jexmex joined20:36
→ bgerber joined20:36
→ Guest71468 joined20:36
← jason237 left20:36
→ didlix joined20:36
EugeneKay Ah, ruby.20:36
→ bazz joined20:36
→ bpayton joined20:36
PleasureTown Yeah, I don't know anything about anything, so it makes it all the more difficult20:36
comps_ → comps20:36
← hinst left20:37
→ leeN joined20:39
PleasureTown I think I'm going to look at this from a different... less confusing, angle20:39
← konundra left20:39
PleasureTown So, I set up a bare repo on the server in say... /opt/puppet and I want to to move those files to /etc/puppet every time they're updated. What's the best way of going about that? Just a bash script?20:40
← didlix left20:40
comps PleasureTown: !deploy20:41
gitinfo PleasureTown: Git is not a deployment tool, but you can build one around it(in simple environments) or use it as an object store(for complex ones). Here are some options/ideas to get you started: http://gitolite.com/the-list-and-irc/deploy.html20:41
→ CodeShark joined20:42
PleasureTown oooh thanks20:43
← jfinn left20:43
→ jday_ joined20:44
→ jfinn joined20:46
← Davey left20:46
← LuxuryMode left20:46
← ojacobson left20:46
PleasureTown Anyone here like frizbees?20:46
→ ferdev_ joined20:47
← ferdev left20:48
ferdev_ → ferdev20:48
← PanicG left20:49
→ BiggFREE joined20:50
EugeneKay My dog likes to eat them20:50
← adam-_-_ left20:53
→ codepython777 joined20:53
← jonascj left20:53
← lamarus left20:53
→ codepython7771 joined20:54
← geekbri left20:54
← LNDN left20:55
PleasureTown https://www.youtube.com/watch?v=Zxgxa1GJlnI20:56
→ esc joined20:57
← esc left20:57
→ esc joined20:57
← hellopat left20:57
→ vmiklos joined20:57
← codepython777 left20:58
← wesside left20:58
← skorgon left21:00
← baneM left21:00
← asakura left21:00
← KillYourTV left21:00
← trungqtran left21:00
→ etcetera joined21:01
→ thierryp joined21:01
← jamesez left21:01
← theshadow left21:02
← etcetera left21:02
← txomon|fon left21:02
→ thiago joined21:03
← dhruvasagar left21:04
→ nicedice_ joined21:04
← Slush- left21:04
← seymour left21:05
← BiggFREE left21:05
→ marshall joined21:05
← coopsh left21:05
← nick_h left21:05
→ seymour joined21:05
← ruskie left21:06
← heytrav left21:06
→ heytrav joined21:06
→ akurilin joined21:07
← bmac left21:07
← m0ltar left21:07
→ nick_h joined21:07
→ irqq_ joined21:07
← nicedice left21:07
← hector_salamanca left21:07
→ frew joined21:07
→ Slush- joined21:08
→ hector_salamanca joined21:08
→ tureba_ joined21:08
← zequence left21:09
← irqq left21:09
→ Mamoru joined21:10
→ coopsh joined21:10
→ zequence joined21:10
← Artpicre left21:11
Dougie187 When you push to a remote, is there a way to push unreachable history?21:11
← style left21:11
← tureba left21:12
→ jgoss joined21:13
← jgoss left21:13
→ jgoss joined21:13
→ barneybook|2 joined21:13
← pr0ton left21:13
→ pr0ton_ joined21:13
← Dougie187 left21:14
→ _ffio_ joined21:15
← daanavitch left21:15
JPeterson if the last commit is a merge commit, how do i replay its commits on top? i.o.w. how do i replace the merge commit with a reply on top?21:16
← PleasureTown left21:16
← ffio_ left21:16
→ jaequery joined21:16
← fooey left21:17
← barneybook left21:17
→ ruskie joined21:17
← tureba_ left21:17
← kakos left21:18
← BadDesign left21:19
→ felds_ joined21:19
← felds left21:19
← Mamoru left21:19
← threesome left21:20
→ tureba joined21:20
→ plumpkin joined21:21
→ jason237 joined21:23
← AaronMT left21:23
← codepython7771 left21:25
→ codepython777 joined21:25
→ revyd joined21:25
← netiul left21:26
→ Suprano joined21:27
→ krik joined21:27
→ cotton joined21:28
→ w0ng joined21:29
← therealGent left21:30
← codepython777 left21:30
← Marenz left21:31
→ pdelvo_ joined21:31
→ JCii8 joined21:32
← JCii left21:32
_BJfreeman → BJfreeman21:32
← ErKa left21:32
→ atmosx joined21:32
atmosx is it possible to move all changes from a specific commit and on to a different branch?21:33
→ JaredMiller joined21:33
→ kereno_ joined21:33
cbreak atmosx: you can cherry-pick commits21:33
← v0n left21:33
atmosx cherry-pick21:34
cbreak that'd create a new commit with the same changes where ever you want21:34
atmosx what does that mean? I'm using a program called 'tig' that says something about cherry-picking21:34
cbreak but to remove the original commit, you'd have to rewrite history, with git rebase -i or git reset --hard21:34
atmosx: man git-cherry-pick21:34
gitinfo atmosx: the git-cherry-pick manpage is available at http://jk.gs/git-cherry-pick.html21:34
← chazcafe left21:34
← pdelvo left21:34
atmosx nice thanks21:34
← Guest71468 left21:35
← JCii8 left21:36
→ JCii8_ joined21:36
← JaredMiller left21:36
→ threesome joined21:36
← revyd left21:37
← longbeach left21:37
→ codepython777 joined21:38
← marshall left21:40
poseidon Anone know of any good reading material for using git in a development lifecycle. ie we have a release every two weeks, tickets for each, some dependent on other releases, etc.21:41
→ ghostmachine joined21:42
← msekleta left21:43
← rednaks left21:43
← irqq_ left21:44
← codeful01 left21:44
→ irqq joined21:44
← pr0ton_ left21:45
← natechan left21:46
← ColoradoPods left21:46
← TeckniX left21:48
→ AaronMT joined21:49
← nerdfiles left21:49
← ghostmachine left21:49
→ JaredMiller joined21:49
← mmitchell left21:49
← codepython777 left21:50
← felds_ left21:51
← JaredMiller left21:51
← john2496 left21:52
→ nerdfiles joined21:52
cbreak poseidon: there's that "a successfull branching model" thing21:52
git flow21:53
→ parhs joined21:54
← parhs left21:54
→ parhs joined21:54
→ pr0ton joined21:54
← parhs left21:54
← irqq left21:54
← m0viefreak left21:54
→ m0viefreak joined21:54
→ ch3mical joined21:54
→ maslowbeer joined21:54
→ frogonwheels_ joined21:54
← jhoepken_ left21:54
→ etcetera joined21:54
← Chikore left21:55
→ andres_f joined21:55
← andref_dias left21:55
→ esc_ joined21:56
→ jamesd joined21:56
→ chuckf_ joined21:56
→ mz|` joined21:56
← mz|` left21:56
→ mz|` joined21:56
→ Klowner_ joined21:56
→ sqz__ joined21:56
← meLon left21:56
→ meLon_ joined21:56
→ lacrymol1gy joined21:56
→ kortti_ joined21:56
→ maqr_ joined21:57
→ kesshi_ketsugo joined21:57
→ crash\\ joined21:57
← Klowner left21:57
← Klowner_ left21:58
→ skoopty joined21:59
→ parhs joined21:59
← parhs left21:59
→ parhs joined21:59
→ epsylon` joined21:59
← breakingthings left22:00
→ adhawkins_ joined22:00
← parhs left22:00
→ parhs joined22:00
← parhs left22:00
→ parhs joined22:00
← parhs left22:01
← fitztrev_ left22:01
→ parhs joined22:01
← parhs left22:01
→ parhs joined22:01
← parhs left22:02
← meLon_ left22:02
← adiabatic left22:03
→ jonconle_ joined22:03
→ meLon joined22:03
← jonconle_ left22:03
← AaronMT left22:03
← seymour left22:03
← esc left22:03
← ferdev left22:03
← chuckf left22:03
← pantsman left22:03
← tjbiddle left22:03
← normanrichards left22:03
← mencito left22:03
← epicowl left22:03
← lacrymology left22:03
← jonconley left22:03
← maqr left22:03
← PaulCapestany left22:03
← momomomomo left22:03
← uidzer0 left22:03
← Sigma left22:03
← jdunck left22:03
← Adlai left22:03
← Kadmos left22:03
← Chillance left22:03
← decotii left22:03
← arand left22:03
← kemyd left22:03
← OliverJAsh left22:03
← mz|`_ left22:03
← DonRichie left22:03
← jamesd_ left22:03
← crash\ left22:03
← MetaCosm left22:03
← andres left22:03
← shentino left22:03
← kesshi left22:03
← bbloom left22:03
← rue left22:03
← EthanCooler left22:03
← segher left22:03
← kortti left22:03
← therealklanni left22:03
← adhawkins left22:03
← kanch left22:03
← frogonwheels left22:03
← epsylon left22:03
← alexwhitman left22:03
← sqz left22:03
← Jobbe left22:03
← nevyn left22:03
← cian1500ww left22:03
← lrvick2 left22:03
← skoop left22:03
← EToS left22:03
← Evanlec left22:03
← Caius left22:03
← DrZaius_ left22:03
← awx left22:03
← thierryp left22:04
→ galaris joined22:04
← JPeterson left22:04
→ Kadmos joined22:04
→ amcsi joined22:04
→ fBirD joined22:04
→ thierryp joined22:04
→ JPeterson joined22:05
← nedbat left22:05
← pinpin404 left22:06
gitinfo set mode: +v22:06
tommick in general, given an object hash, what's the plumbing for reproducing the object content with that hash?22:07
← tirengarfio left22:07
tommick neither of these two is it:22:07
git cat-file commit 18205b318d1bc337e2fa3626967003cc8d2d74e8 | sha1sum22:07
sha1sum .git/objects/18/205b318d1bc337e2fa3626967003cc8d2d74e822:07
jast that's because the hash includes the object's header, too (which isn't part of the object itself)22:08
← homer91 left22:08
→ dsdeiz_ joined22:08
jast the git-generated hash, that is22:08
and the file in the objects dir is zlib-compressed22:08
← krik left22:08
← dsdeiz left22:08
→ SBoudrias joined22:08
← R1KRDO left22:08
← nbouscal left22:08
← adhawkins_ left22:09
→ felds joined22:09
→ dambler joined22:09
→ adhawkins joined22:09
← meLon left22:09
tommick +jast: yes, the git-generated hash (I'm walking a tree as part of producing a serialisation format)22:09
adhawkins → adhawkins-away22:09
← Tectu_ left22:10
→ Tectu joined22:10
tommick +jast: okay, but should the git cat-file one work?22:10
→ tirengarfio joined22:10
jast for getting the raw content of an object? sure22:10
← dsdeiz_ left22:11
→ dsdeiz joined22:13
→ Bass10 joined22:13
← cannonball left22:13
→ Vendethiel1 joined22:14
← Vendethiel1 left22:14
→ Vendethiel joined22:14
Vendethiel pls22:14
If I create a branch to work off, why are the changes applied to my "previous" branch?22:15
← felds left22:15
← AWinterman left22:15
jast Vendethiel: related to !float, presumably?22:15
gitinfo Vendethiel: If you have made a change in your working directory and have NOT YET COMMITTED, you may "float" that change over to another (`git checkout oldbranch`) or new (`git checkout -b newbranch`) branch and commit it there. If the files you changed differ between branches, the checkout will fail. In that case, `git stash` then checkout, and `git stash apply` and go through normal conflict resolution.22:15
← p4tux left22:15
Vendethiel jast: I love you22:16
← Anders_J left22:16
Vendethiel but seriously, how come it works like that?22:16
andres_f → andres22:17
Vendethiel seems anti-intuitive to me22:17
← tirengarfio left22:17
Vendethiel and if you're replying that fast with a command, I guess this question is asked rather frequently22:17
→ AWinterman joined22:17
→ MetaCosm joined22:17
← MetaCosm left22:17
→ MetaCosm joined22:17
→ s0undt3ch joined22:18
→ rue joined22:18
← MrRoot left22:18
← nerdfiles left22:18
→ BSaboia joined22:18
→ Chillance joined22:18
→ EthanCooler joined22:18
→ segher joined22:18
→ tirengarfio joined22:18
→ tirengarfio_ joined22:18
→ arand joined22:19
→ AaronMT joined22:19
→ jdunck joined22:19
← jaequery left22:19
→ cian1500ww joined22:20
→ Jobbe joined22:20
→ shentino joined22:20
→ seymour joined22:20
→ ferdev joined22:20
→ tjbiddle joined22:20
→ normanrichards joined22:20
→ mencito joined22:20
→ epicowl joined22:20
→ PaulCapestany joined22:20
→ uidzer0 joined22:20
→ OliverJAsh joined22:20
→ Adlai joined22:20
→ decotii joined22:20
→ DonRichie joined22:20
→ bbloom joined22:20
→ lrvick2 joined22:20
→ kanch joined22:20
→ alexwhitman joined22:20
→ Evanlec joined22:20
→ Caius joined22:20
→ DrZaius_ joined22:20
→ awx joined22:20
→ kemyd joined22:20
← cian1500ww left22:20
→ knite joined22:20
← AWinterman left22:20
← cmn left22:20
← acharlieh left22:20
shentino → Guest727822:20
→ cian1500ww joined22:20
→ EToS joined22:21
→ felds joined22:22
→ cmn joined22:22
← stliu left22:22
← dreiss left22:23
→ mfulz1 joined22:23
← sekati left22:23
→ Mamoru joined22:23
← mfulz1 left22:25
← felds left22:25
→ mfulz1 joined22:25
← mfulz1 left22:26
← tvw left22:26
→ mfulz joined22:26
← noname001 left22:26
mfulz Hi22:26
gitinfo mfulz: hi! I'd like to automatically welcome you to #git, a place full of helpful gits. Got a question? Just ask it — chances are someone will answer fairly soon. The topic has links with more information about git and this channel. NB. it can't hurt to do a backup (type !backup for help) before trying things out, especially if they involve dangerous keywords such as --hard, clean, --force/-f, rm and so on.22:26
← tirengarfio_ left22:27
← tirengarfio left22:27
mfulz does anybody know about the following problem: I've gitolite-2.2 ubuntu 12.04 lts running via https over Apache which is working perfectly with git since version 1.8.0.x but since 1.8.1.x over https I'll always get info/refs not valid: is this a git repository?22:27
→ dsdeiz_ joined22:28
→ tirengarfio_ joined22:28
→ tirengarfio joined22:28
→ jaequery joined22:28
← Mamoru left22:28
cmn this is the way VCSs work22:29
← sq-one left22:29
← _iron left22:29
← dsdeiz left22:30
→ trungqtran joined22:30
mfulz cmn: this is the way VCSs work <- addressed to my question?22:30
← kermit left22:31
→ ColoradoPods joined22:31
→ krazyj joined22:31
← trungqtran left22:31
← breck7 left22:32
→ kermit joined22:32
→ noname001 joined22:32
krazyj hey all… i'm writing some bash automation and i was wondering… what's the SAFEST method to extract any diffs, so a fetch/merge can be done?22:32
thiago mfulz: does it work over git or ssh?22:32
krazyj use case: automation which, at midnight, pulls the latest master and merges it in22:32
i want the automation to adapt for any uncommitted diffs22:33
thiago krazyj: fetch and merge do not work with diffs22:33
krazyj: merge are never safe22:33
← jonatasnona left22:33
krazyj thiago: right.. but is there some combination/configuration of stash/etc22:33
mfulz yes over ssh its working fine22:33
thiago mfulz: is that the smart git or the dumb one?22:33
mfulz thiago: what do you mean exactly?22:33
thiago mfulz: there are two HTTP serving methods: the dumb and the smart one22:34
which one are you using?22:34
← knite left22:34
→ relipse joined22:34
mfulz ScriptAlias / /usr/share/gitolite/gl-auth-command/22:35
thiago krazyj: do you want to automatically fetch changes, merge them and keep your uncommitted changes?22:35
krazyj thiago: yes22:35
← Vendethiel left22:35
krazyj to the maximum possible extent22:35
→ KillYourTV joined22:35
thiago krazyj: that's not safe22:35
jast Vendethiel: well, uncommitted changes don't "belong" to any branch. the working tree is just switched around to match whatever branch you're checking out. if there are uncommitted changes, it's infinitely more sensible (and useful) to keep them than to quietly destroy them.22:35
krazyj thiago: what's The Next Best Thing, you think?22:35
thiago krazyj: I strongly recommend against them22:35
krazyj: fetch every night, without merging22:35
krazyj it'd be a shame for my automation to not fetch/merge anytime there are diffs22:35
thiago krazyj: also, commit your changes22:35
krazyj *sigh* ok22:35
thiago krazyj: and still do not merge22:36
krazyj: that's still not safe22:36
krazyj thiago: if there's no diffs? why?22:36
thiago krazyj: because it's not safe22:36
krazyj: merges can fail22:36
krazyj what about it?22:36
jast oh, I was actually a few seconds too late there. whoops.22:37
thiago krazyj: merges can also silently fail: git succeeds, but the result is incorrect22:37
← braisarias left22:38
→ s0undt3ch__ joined22:38
krazyj thiago: how often can that happen?22:38
← s0undt3ch left22:38
← decotii left22:38
krazyj thiago: and i don't understand how that could happen...22:38
thiago krazyj: any time22:38
← jaequery left22:39
thiago krazyj: the merge result might be correct, even without conflicts, but the resulting file is not syntactically correct anymore.22:39
mfulz thiago: do you asking about the http-backend? $ENV{GIT_HTTP_BACKEND} = "/usr/lib/git-core/git-http-backend";22:39
thiago git does not know what purpose your files may have, what syntax is correct or not. It does line-based merging, not semantic.22:39
krazyj that's an acceptable outcome in my scenario. all commits are expected to be syntactically sound, lest someone's ass is on the line ;)22:40
thiago you don't get it22:40
suppose on one side of the merge there's a function getting renamed22:40
→ FreezingCold joined22:40
← Gentlecat left22:40
thiago on the other side of the merge, a new call to that function is placed22:40
when the two sides are merged together, there's no conflict, but the code will not work because the new call is placed to the old name22:40
→ baneM joined22:40
krazyj right ok22:41
i was thinking in a scenario with no diffs22:41
thiago I was too22:41
→ sekati joined22:41
thiago both changes are, individually, correct. But together they are wrong.22:41
krazyj ah ok22:41
→ aendrew joined22:41
krazyj thiago: is there any way to merge a branch that you're not checked out to22:42
→ ErKa joined22:42
thiago no22:42
you can push an update22:42
← atmosx left22:42
krazyj ?22:42
thiago git push . newhead:branchname22:42
that is safe22:42
→ knite joined22:42
krazyj thiago: what if… i'm on branch foobar and i want to fetch/merge my local master from origin:master22:42
foobar has diffs, etc22:43
thiago merge is *always* a human-driven operation22:43
← tirengarfio left22:43
← tirengarfio_ left22:43
thiago to understand: do you want to update your local master with origin's master?22:43
krazyj thiago: right22:43
without checking out master to do it22:44
→ AdrienBrault joined22:44
thiago krazyj: git fetch origin && git push . origin/master:master22:44
← cirenyc left22:44
thiago or, in one command: git fetch origin master:master22:44
krazyj thiago: ahh ok22:44
→ AWinterman joined22:44
thiago that operation can fail, but it is safe22:44
krazyj thiago: i think my automation might do as such....22:44
thiago if it fails, that's because your local master has commits that aren't present in origin's master22:44
mfulz thiago: could it be because of old git version on ubuntu 1.7.9.522:44
thiago you will require a manual merge22:44
mfulz ?22:44
thiago mfulz: I don't know. You're using gitolite in that web server, not git.22:45
krazyj thiago: 1. check if diffs on local 'master' branch. 2. if no diffs, `git fetch origin master:master`22:45
else.. do nothing22:45
→ skorgon joined22:45
→ tirengarfio_ joined22:45
→ tirengarfio joined22:45
← rnsanchez left22:45
← tirengarfio left22:45
mfulz thiago: ah ok I'll give it a try because the git-http-backend is comming from git-core22:46
thiago krazyj: let's use proper terms. What do you mean by diffs? uncommitted changes? Or do you mean commits not present in origin's master?22:46
krazyj thiago: either22:46
so if EITHER OF THOSE EXIST, do nothing22:46
else.. git fetch origin master:master22:46
← frem left22:46
← BSaboia left22:46
thiago krazyj: the git fetch command will fail if either condition is true22:46
→ nbouscal joined22:47
krazyj thiago: wait, what? i thought fetch can always happen.22:47
thiago krazyj: if the local master has commits not found in origin's master, that fetching operation would be not a fast-forward.22:47
krazyj: if the local master has uncommitted changes, it is the branch that is checked out. You're not allowed to do that operation.22:47
krazyj thiago: right, i'm saying if the local master has commits not found in origin's master, automation does nothing22:47
ditto for any uncommitted on the local master22:47
→ pdelvo joined22:47
thiago that git fetch command will fail on its own22:47
→ guiambros joined22:48
→ guiambros_ joined22:48
krazyj the automation script will skip that repo, and continue22:48
ELSE22:48
git fetch origin master:master22:48
thiago the git fetch command, alone, is enough22:48
krazyj for?22:49
thiago for updating master, provided it is not checked out22:49
krazyj if the local master branch is, essentially, untouched, i want to fetch/merge the origin master branch22:49
that's my goal22:49
thiago that's what it will do, provided the branch is not checked out22:50
← skorgon left22:50
← lacrymol1gy left22:50
← pdelvo_ left22:50
krazyj thiago: are you saying if local:master is checked out, `git fetch origin master:master` doesn't fetch/merge commits from origin:master?22:51
else, it does22:51
thiago it downloads, but it refuses to update22:51
you cannot update the checked out branch using push or fetch22:51
mfulz not working :(22:52
← tirengarfio_ left22:52
krazyj thiago: any ideas what command i could use in automation for this?22:52
thiago krazyj: git fetch origin && git merge --ff-only22:52
krazyj thiago: what's ff-only?22:53
thiago "fast forward only"22:53
→ cirenyc joined22:54
← sekati left22:55
← nbouscal left22:56
← knite left22:57
→ dagerik joined22:57
← mfulz left22:57
dagerik when i clone this repo and type git log, why do i only see two commits? https://github.com/huceke/omxplayer/commits/master22:57
→ knite joined22:57
dagerik i need to check out an earlier commit22:57
← cirenyc left22:59
cbreak different branches? :)22:59
tommick +jast: apologies if I'm being slow, but then why doesn't "git cat-file commit 18205b318d1bc337e2fa3626967003cc8d2d74e8 | sha1sum" give an output of 18205b318d1bc337e2fa3626967003cc8d2d74e8 ?22:59
→ jaequery joined23:00
→ Joshi3s` joined23:00
cbreak tommick: why should it? git doesn't use sha directly23:00
tommick: man git-hash-object23:00
gitinfo tommick: the git-hash-object manpage is available at http://jk.gs/git-hash-object.html23:00
← pitzips left23:00
tommick cbreak: thanks, that's exactly the pointer I needed23:02
→ tirengarfio joined23:02
→ tirengarfio_ joined23:02
tommick "git cat-file commit 18205b318d1bc337e2fa3626967003cc8d2d74e8 | git hash-object --stdin -t commit" gives me the hash back23:02
krazyj thiago: sorry, do you think you can explain a ff merge versus any other merge23:03
thiago krazyj: ff merge isn't a merge. It's "update the branch to point to the new commit and check it out"23:04
← johnnyfuchs left23:04
thiago krazyj: it's only permitted because the current branch has no new commits that the remote didn't have23:04
→ asakura joined23:05
← Thra11 left23:06
← tirengarfio left23:07
← devth left23:07
← chamunks left23:07
← chrisshattuck left23:07
→ chamunks joined23:08
tommick next question: given just an object id (a hash), is it possibly to reliably determine its type? I'm guessing not, cos a blob's content could look like a tree or a commit, right?23:08
cbreak tommick: man git-cat-file23:09
gitinfo tommick: the git-cat-file manpage is available at http://jk.gs/git-cat-file.html23:09
tommick thanks :)23:10
→ pac1 joined23:10
→ FabTG_ joined23:10
← epicowl left23:10
← tirengarfio_ left23:10
gitinfo set mode: +v23:11
← FabTG left23:11
← AaronMT left23:11
pac1 Where can I find information on managing a team clone.23:11
?23:11
← knite left23:11
→ tirengarfio joined23:12
→ tirengarfio_ joined23:12
→ AaronMT joined23:12
← infynyxx left23:15
→ LekeFly joined23:15
→ curvedmark joined23:16
→ alunduil joined23:16
← jaequery left23:16
← ColoradoPods left23:17
← dmarkham_ left23:20
thiago pac1: what do you mean by managing?23:20
pac1: and what do you mean by team clone?23:20
← catcher left23:21
← aendrew left23:21
pac1 thiago, a team clone is a clone on the gitorious.org server that has been setup for a specific team.23:21
thiago ok23:21
pac1 managing= keeping a clone in sync with a repository I do not have commit rights to.23:21
← AaronMT left23:22
← s0undt3ch__ left23:23
thiago so you want to fetch from that source repository and push to the team's clone?23:23
pac1 that sounds right.23:23
→ s0undt3ch joined23:23
← etcetera left23:23
pac1 so to work with this I'll need a local clone.23:25
→ nbouscal joined23:25
pac1 can a local clone pull from the main repository and push to the team one?23:25
thiago yes23:25
pac1 Is that how the team repository will stay in sync?23:25
with the main one.23:25
?23:25
thiago yes23:26
→ etcetera joined23:26
thiago it stays in sync by your making it be in sync23:26
pac1 ok.23:26
I pull from the main repository and push to the team clone. How do I prevent a pull from the main repository from overlyaing my changes?23:27
will a stage do it?23:27
thiago don't pull, just fetch23:28
pac1 ok.23:28
← akiress left23:28
pac1 so fetch from the main repository will update my local repository. If I do that before staging or commiting a change, do I lose that change?23:28
← tirengarfio left23:29
← tirengarfio_ left23:29
thiago fetch never affects the working tree23:29
pac1 aha! Does pull affect the working tree? Don't answer....23:29
thiago not answering :-)23:29
← PaulCapestany left23:30
→ PaulCapestany joined23:30
← nizeguy left23:31
← Kadmos left23:32
← pr0ton left23:32
← etcetera left23:33
← boombatower left23:33
pac1 protection is by creating a branch!23:33
→ devoldmx joined23:33
→ etcetera joined23:33
← Kvothe left23:33
← normanrichards left23:34
← Prazanna left23:36
→ johnnyfuchs joined23:38
→ ojacobson joined23:38
← grizzle left23:38
← drizzd left23:38
→ lresende joined23:38
→ grizzle joined23:38
← quazimodo left23:39
→ infynyxx joined23:39
← devoldmx left23:39
→ devoldmx joined23:40
→ SrPx joined23:40
→ JaredMiller joined23:40
← JaredMiller left23:41
→ AWinterm_ joined23:42
← scientes_ left23:42
→ BSaboia joined23:43
← AWinterman left23:43
→ tirengarfio_ joined23:43
→ tirengarfio joined23:43
← tirengarfio left23:43
→ supnagga joined23:44
→ quazimodo joined23:45
→ mixel joined23:45
← cjz left23:46
← CodeShark left23:46
→ jxpx777 joined23:49
← AWinterm_ left23:49
→ sekati joined23:49
← BSaboia left23:50
← leeN left23:52
← curvedmark left23:53
← galaris left23:54
← guiambros_ left23:55
→ jonconley joined23:55
← supnagga left23:55
→ nedbat joined23:55
→ Orbitrix joined23:55
← lifelike left23:56
→ LuxuryMode joined23:56
→ Olipro joined23:57
← nbouscal left23:58
← redrocket left23:59
← ktosiek left23:59

Logs Search ←Prev date Next date→ Channels Documentation