IRCloggy #git 2011-02-16

Logs Search ←Prev date Next date→ Channels Documentation

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

2011-02-16

← dreiss left00:00
devians i really need to just trim all the whitespace and fix all the line endings, commit that and move forward. its only 4 people on the repo, shouldn't be too painful00:00
← pheaver left00:01
pheaver_ → pheaver00:01
SethRobertson devians: Do you have multiple branches?00:01
→ verto joined00:01
SethRobertson public branches00:01
devians not in upstream no00:01
SethRobertson Then you can (publicly) avoid the fake merge horror I talked about.00:01
← chrischambers left00:03
← pheaver left00:03
→ pheaver joined00:03
← eletuchy left00:03
→ Quex joined00:06
← Quex left00:06
← d0k left00:06
→ btipling joined00:06
← btipling left00:06
→ btipling joined00:06
← kuhrt left00:07
← SimonP86 left00:07
→ orospakr joined00:08
→ chrischambers joined00:09
← thiago_home left00:09
← _psychic_ left00:10
→ zivester_ joined00:10
→ eletuchy joined00:10
← LiamH1 left00:11
← _Vi left00:11
→ derek joined00:11
← derek left00:12
→ vanmik joined00:12
devians mmm00:14
← maattd left00:14
devians im still getting that whole, working dir is dirty/unstaged changes crud, even when theres nothing there, the dir is clean00:14
its gotta be some sort of file permission or time or something diff that doesnt show up00:15
→ ka1ser joined00:17
→ jmorris joined00:18
← ka1ser left00:18
ayust devians: are you on windows00:18
→ ka1ser joined00:18
ka1ser hi00:18
← verto left00:19
devians ayust: osx00:19
ka1ser how to checkout a branch in a remote's tag? suppose tag "GOOD" and remote name "origin".... I'm trying: git checkout -b test origin/GOOD but it doesn't work...00:19
ayust devians: what's the output of "git diff"00:20
ka1ser any ideas?00:20
cirwin ka1ser: step 1, clone the entire repository00:21
ayust ka1ser: tags aren't pulled to refs/remotes/, they're pulled to refs/tags00:21
cirwin step 2. checkout the tag00:21
← normanrichards left00:21
ka1ser oh nvm... when I do the fetch I get the tags ^^.... my bad00:21
SethRobertson devians: The ctime thing didn't help?00:21
ayust so just "git checkout -b test GOOD"00:21
would work00:21
ka1ser cirwin I just fetched it00:21
thanks =]00:21
devians SethRobertson: it did not :(00:21
→ stice joined00:22
← jmil left00:22
ChanServ set mode: +v00:22
SethRobertson Try core.ignoreCygwinFSTricks true00:22
← airborn left00:23
← j416 left00:23
SethRobertson devians: I had a typo in my first suggestion. it should have been core.trustctime w/o an e in the middle.00:23
stice Hey chan, little question. When i clone a git repo, all files are in there.. but when i del one file and then enter 'git pull' it shows me its still up to date, is that normal behaviour?00:23
SethRobertson stice: what does git status say?00:24
NfNitLoop "up to date" means you have everything that's on the server. (and more)00:24
→ jmil joined00:24
NfNitLoop as SethRobertson mentioned... try 'git status'. it'll tell you that you have more than the server.00:24
← ben225 left00:27
stice it says changed but not updated, is there a way to 'clone' again or ignore the actual .git informations without removing the folder and cloning by hand?00:27
SethRobertson stice: I hesitate to recommend this: `git fetch; git reset --hard @{upstream}`00:27
cirwin what do you mean by "clone again"? if you want to get the working directory into the same state as the last commit on the server, just use checkout00:28
(or reset if you have already made commits)00:28
SethRobertson stice: More safely, `git status` tells you how to undo each change you made00:28
stice thanks for the help :)00:29
← lirakis left00:29
← mjf left00:30
→ devn joined00:31
← dropdriv1 left00:31
devn Is there any way to specify multiple committers?00:31
→ vadimkk joined00:31
← linusoleander left00:32
SethRobertson Multiple authors makes more sense than committers00:32
Essentially no, though you could temporarily reset your user.* information00:33
← engla left00:33
→ engla joined00:33
engla devn: multiple for one commit? no.00:33
← vadimkk left00:34
← HopsNBarley left00:34
← xiong left00:37
→ dropdrive joined00:37
→ linusoleander joined00:38
← j0shua left00:38
→ cutmail joined00:39
← ironfroggy__ left00:39
→ ironfroggy__ joined00:39
→ Vaddim joined00:39
← ironfroggy__ left00:40
← Heimidal left00:40
→ MikhX joined00:40
→ Exodist joined00:42
← leavittx left00:43
→ robinbowes joined00:46
→ dermoth|home joined00:46
← mattly left00:47
← Vaddim left00:47
← dyer left00:47
← steph021 left00:48
→ daum_ joined00:48
← mando left00:48
daum_ hey guys when i am trying to do my original push to a new repository i get the following: http://pastebin.com/u9QuaJKR00:49
FauxFaux daum_: Did you read what it tells you to read?00:49
ayust you don't have a local master00:49
daum_ hmm weird..00:50
ayust first, type 'git status' - did you actually commit?00:50
second, type 'git branch' - are you actually on the branch you think you're on?00:50
daum_ ah doh forgot to first commit...man been a long day00:50
← stice left00:50
← eijk__ left00:53
→ xiangfu joined00:55
← spearce left00:55
← obrienmd left00:56
← linusoleander left00:58
← tewecske left00:59
→ Engin joined01:00
← MikhX left01:01
→ MikhX joined01:02
← harcourtfentonm1 left01:08
← ddkilzer left01:08
→ ProperNoun joined01:09
← chrischambers left01:09
← ProperNoun left01:10
← Engin left01:10
← hobodave left01:10
→ lirakis joined01:10
→ _Vi joined01:12
→ Anthro joined01:13
← __name__ left01:13
← noahcampbell left01:14
Anthro Anyone know how to get the equivalent of svn log -v? I want to see the changed files in each commit, but not a full diff.01:15
cirwin git whatchanged01:15
→ shiba_yu36 joined01:15
← przemoc left01:16
← alvaro_o left01:16
Anthro cirwin: Looks like what I want, but I'm not entirely sure what I'm looking at now. There's a list of files, but I don't know what the numbers mean.01:17
← yxz97 left01:18
cirwin the first two columns are file permissions, you can ignore them01:18
the second two columns are the start of the SHA-id of the file as it appeared in that commit01:18
again, you can ignore them01:18
← jrmuizel left01:19
cirwin maybe git log --stat is less confusing?01:19
→ shadowhand joined01:19
→ dyer joined01:19
← dyer left01:19
→ dyer joined01:19
shadowhand so if "git ls-files -t" is deprecated, how can one get a list of all files that are currently tracked?01:20
"git status --untracked-files=no --short" is start, but that only shows HEAD01:20
cirwin git ls-files shows you what is tracked with no arguments?01:20
← smolyn left01:21
← rns left01:21
Anthro cirwin: Ah, git log --stat is exactly the information I want. Thanks!01:21
← tvw left01:22
shadowhand cirwin: correct, but i need the status01:22
which -t provides01:22
cirwin ah, ok01:22
← daum_ left01:23
→ _psychic_ joined01:23
→ przemoc joined01:24
← cutmail left01:25
→ bcardarella joined01:26
← SirFunk left01:28
← Anthro left01:28
→ kuhrt joined01:29
RJD22 → RJD22|away01:29
→ Utkarsh joined01:30
→ coderdad joined01:32
← shabble_ left01:32
← Utkarsh_ left01:32
→ urkud_ joined01:32
← urkud left01:33
← justin-george left01:33
← BobJelly left01:34
← Araxia_ left01:34
→ bburhans_ joined01:34
→ rns joined01:34
← bburhans left01:35
→ coderdad_ joined01:36
← coderdad left01:37
← xerora left01:37
→ BobJelly joined01:38
← kenichi left01:38
← eMBee left01:40
→ eMBee joined01:40
← doener left01:40
→ doener joined01:40
← Arrowmaster left01:40
← oriba left01:41
← coderdad_ left01:42
→ coderdad joined01:42
→ pleonasm joined01:42
harleypig Is there a way I can migrate a project in a subversion subdirectory to a git repo? If I try to migrate the whole subversion repository I seem to run out of memory (I think).01:44
→ jcromartie joined01:45
jcromartie Is this a good place to talk about workflow?01:45
FauxFaux Here!01:47
harleypig: Without branches or tags? Just give it the full path to the folder you want within trunk.01:47
← setmeaway left01:48
→ setmeaway joined01:48
→ mdel joined01:49
← _psychic_ left01:51
← museic left01:51
← ningia left01:52
→ obrienmd joined01:53
RJD22|away → RJD2201:54
← n3m left01:54
← warthog9 left01:55
← pdelgallego left01:55
→ Araxia_ joined01:56
← mithun left01:56
← Guest91470 left01:57
→ brosner joined01:57
→ yuwang joined01:59
→ n3m joined01:59
← amerine left01:59
→ stliu joined01:59
← stliu left01:59
→ stliu joined01:59
→ jrmuizel joined01:59
→ Maletor joined02:00
shadowhand hmm, why does "git log --reverse -n1" not show me the original commit?02:00
seems like a bug02:00
"git log --reverse --format='%H' |tail -1" and "git log --reverse -n1 --format='%H'" should have the same output, no?02:01
cirwin well, -1 limits the search and then --reverse changes the output02:01
← ericlee left02:01
→ smolyn joined02:02
shadowhand right, and i am saying shouldn't --reverse be noticed before limit?02:02
→ ericlee joined02:02
shadowhand it's like doing an SQL LIMIT before ORDER BY02:02
your result will be wrong 99.9999% of the time02:02
cirwin what about git log -10 --reverse?02:03
← lol_h left02:03
← obrienmd left02:03
cirwin it seems in that case, it's much more likely that you'll want the ten most recent commits02:04
→ marc_kdab_ joined02:04
← Maletor left02:05
shadowhand cirwin: i see your point02:05
→ gretch joined02:05
shadowhand still seems a bit odd to me02:05
cirwin it's consistent at least02:05
shadowhand so far as i can tell, there is no log option to get the root commit02:05
→ samhassell joined02:06
cirwin what's the root commit useful for?02:07
← aleray left02:07
← eletuchy left02:07
→ mando joined02:07
→ hobodave joined02:07
← marc_kdab left02:07
hachi_ → hachi02:08
← mando left02:08
← tjholowaychuk left02:08
cirwin (and I suppose, what would you expect to happen if you had more than one?)02:08
→ justin-george joined02:09
← jmorris left02:10
shadowhand cirwin: this all comes down to "git ls-files -t" being deprecated02:10
→ jmorris joined02:11
shadowhand so i am trying to find alternatives02:11
cirwin what are you trying to do?/02:11
shadowhand such as: git diff --name-status --ignore-submodules=all $(git log --reverse --format="%H" | head -1) | tr '\t' ' '02:11
cirwin at a high level?02:11
← jcromartie left02:11
shadowhand cirwin: i'm patching http://github.com/resmo/git-ftp02:11
← sh4rm4 left02:11
shadowhand it currently fails with submodules, so i am trying to use "git diff" rather than "ls-files"02:12
← lirakis left02:12
← sudonotplease left02:12
cirwin is this to find the list of files that have changed since you last deployed, or what?02:12
shadowhand is it actually possible for a root commit to be two commits?!02:12
cirwin: well, yes, and no02:13
cirwin you can have two commits with no parents, yes02:13
shadowhand as a root?02:13
← coderdad left02:13
cirwin a commit with no parents is a root02:13
shadowhand when doing "git ftp init" it has to send all the files02:13
but sending a directory (submodule) will fail02:14
miserably02:14
hence, i need a "ls-files" that excludes submodules02:14
→ jstein joined02:14
← pheaver left02:14
→ wliao joined02:15
→ Arrowmaster joined02:15
jstein hi in my project is a config file. I want to exclude this from usual git commits (even if changed) but sometimes if we have new options i want to commit a new version. How would you do that?02:15
cirwin I have no further help to give — but it sounds like you're trying to over-complicate things.02:16
← jrmuizel left02:16
cirwin jstein: you hope that your configuration file language lets you include a second configuration file that you can store locally, if not; you're just careful when you do git add02:16
→ NickyP joined02:16
shadowhand cirwin: maybe just a test on sending the file that checks if it is a directory?02:16
→ anjen joined02:18
← yuwang_ left02:18
jstein cirwin: no i only have one file. i thought i could put it into the exclude list and overwrite the excludelist with git commit my.conf manual if necessary02:18
bburhans_ → bburhans02:18
cirwin you can't ignore a file that exists in the repository, git suckth02:18
← gretch left02:18
cirwin there's a hack with update-index02:19
← bcardarella left02:19
→ _psychic_ joined02:20
← anjen left02:20
jstein how do ohters solve that? many coders will have their own configs in the project...02:20
← _psychic_ left02:21
cirwin we have a config file template, and eveery developer creates their own copy that is not in the repository02:21
which is fine as the default config has never changed02:21
→ rgr joined02:21
← jfkw_ left02:22
→ Heimidal joined02:22
→ ben_alman joined02:22
monokrome cirwin: I usually maintain a secondary configuration in my projects. If you have an application config, there can be a local config that is used to override the former's settings.02:23
shadowhand monokrome: ditto02:23
← jaeckel left02:23
cirwin that works providing you have a good configuration language02:23
← cwillu_at_work left02:23
monokrome The language really isn't significant02:23
→ sh4rm4 joined02:24
monokrome If you can't use two configurations, then you can always have your software check if file A is there before it uses file B02:24
→ Swimming_bird joined02:24
monokrome But I don't see any reason why you wouldn't be able to use two configurations.02:24
Here's an example: https://github.com/monokrome/django-base/blob/master/bin/application.wsgi02:25
→ amerine joined02:26
monokrome gets back to fighting with git-daemon02:26
shadowhand you can always include a config, and then another config02:26
most languages will overwrite the options in the original with the 2nd02:27
monokrome thinks so, also,02:27
shadowhand we do that in Kohana02:27
← mw_ left02:27
shadowhand cirwin: thanks, btw02:27
i just added a simple "if [ -d ... ]" check and am skipping directories02:28
cirwin :)02:28
← SunilThaha left02:28
shadowhand then i just need to add some git submodule foreach magic02:28
← anathematic left02:28
→ jaeckel joined02:28
← jaeckel left02:28
→ jaeckel joined02:28
monokrome shadowhand: Kohana makes it even more easy, though - with it's cascading filesystem loading02:28
→ Auro1 joined02:29
Auro1 hi02:29
shadowhand monokrome: yes indeed02:29
woot02:30
monokrome shadowhand: Why don't you use .gitignore?02:30
→ rayners joined02:31
← Titosemi left02:31
← Schmallon left02:31
← chessguy left02:31
← Araxia_ left02:32
→ Titosemi joined02:33
→ Remixman joined02:34
← justin-george left02:34
→ justin-george joined02:35
← smolyn left02:35
harleypig FauxFaux: I'm following the steps found at http://pauldowman.com/2008/07/26/how-to-convert-from-subversion-to-git/ and passing svn+ssh://user@svn.repo.com/pdev/project/trunk/devbranch for the repo url ... it churns for a few minutes, then leaves me with an empty directory. So, what am I misunderstanding?02:35
← Titosemi left02:35
devians SethRobertson: i think that ctime flag might have worked02:36
→ obrienmd joined02:37
→ mrevd joined02:37
mrevd can i add filemode = false to my user .gitconfig?02:37
→ hyperair joined02:39
→ TanEk joined02:40
← werdan7 left02:40
← ben225_ left02:41
← Yuuhi left02:42
ddilinger according to git status, my branch is ahead of 'origin/master' by 12 commits, but i havn't made any commits to master, and certainly not 12. How can i figure out whats going on?02:42
err, my master branch02:43
git log shows no mention of my username02:43
→ Vampire0 joined02:45
kpreid ddilinger: presumably the branch you're on thinks the corresponding remote branch is master. check .git/config. also, git log origin/master..HEAD will list those 12 commits02:45
← b00p3r left02:46
→ dl9pf joined02:46
← dl9pf left02:46
→ dl9pf joined02:46
→ Booper joined02:46
ddilinger kpreid: the branch i'm on is master02:46
← bronson left02:47
kpreid ok, just do the git log then.02:47
← ki__ left02:47
ddilinger git log shows no mention of my username02:47
← Vampire0_ left02:47
ddilinger i've never commited to master02:47
thats only done by lead02:47
kpreid doesn't matter, those are still commits which are in master but not origin/master02:47
← dl9pf_ left02:47
← mrevd left02:48
ddilinger how do i see just those commits? there is nothing in git log with my name attached, i dont know what else to search for02:48
kpreid the command I gave should list exactly those commits02:48
...but I don't know git thoroughly02:48
ddilinger git log is 5k lines long02:48
its not just the 12 commits :P02:49
kpreid git log origin/master..master is 5k lines long?02:49
shadowhand monokrome: sorry?02:49
monokrome: use .gitignore for what?02:49
monokrome ignoring a file/directory02:49
ddilinger git log origin/master..master is a diff command, thats 75 lines, which according to unfuddle are the changes i just pulled from origin02:50
its the stuff the lead merged in today and yesterday02:50
→ gretch joined02:50
kpreid sounds like your remotes/origin/master isn't updated for some reason!02:50
→ karstensrage joined02:50
ddilinger hmm, so what do i do?02:51
kpreid well, I would try: git fetch origin master02:51
karstensrage how come git branch --set-upstream mybranch origin/mybranch says origin/mybranch is ambiguous?02:51
→ bcardarella joined02:51
cirwin karstensrage: did you create a local branch called origin/mybranch by accident?02:51
(I wish git branch would forbid that, but it doesn't)02:52
karstensrage cirwin, i don't think i did, it wasnt in my .git/config for sure02:52
ddilinger kpreid: hmm, fetch must be different than pull some way? i ran another git pull earlier and it only said i'm up to date. Now it pulled some objects, said i was up to date, and git status is back to like it should be02:52
cirwin branches don't have to be in .git/config if you've not set any configuration for them02:52
ddilinger kpreid: so summary, i'm not sure why but its fixed :)02:52
kpreid: thanks02:52
→ jfkw joined02:52
karstensrage hmm02:52
cirwin `git branch` will show you02:52
karstensrage ill check tomorrow02:52
kpreid ddilinger: well, pull is fetch and merge/rebase, so ... I dunno, ask someone who knows more about exactly what git chooses to fetch or not02:53
← chrissbx left02:55
→ mando joined02:56
← mdel left02:57
→ justin_george joined02:57
← alip left02:57
← mando left02:58
← justin-george left02:58
→ alip joined02:59
→ smolyn joined03:01
→ mdel joined03:02
→ phaed joined03:02
phaed hi, I'm trying to cherry-pick a commit from another branch and am seeing the error:03:03
← Remixman left03:03
phaed Your local changes to the following files would be overwritten by merge: xxx/xxx/xxx03:04
but if I run `git diff xxx/xxx/xxx` nothing is returned03:04
← mdel left03:04
phaed s/returned/shown03:04
→ pheaver joined03:04
← smolyn left03:04
phaed any idea? am I reading the error wrong?03:04
← sontek left03:06
→ sontek joined03:06
phaed ah well, whatever was going on, I avoided it by just running `git checkout -- xxx/xxx/xxx` and then cherry-picking again03:06
monokrome last03:06
← jstein left03:07
← jjuran left03:07
← c00kiemon5ter left03:07
← obrienmd left03:08
→ c00kiemon5ter joined03:08
← c00kiemon5ter left03:08
→ c00kiemon5ter joined03:08
→ steph021 joined03:09
← steph021 left03:09
→ steph021 joined03:09
→ sudonotplease joined03:09
← rgr left03:10
→ obrienmd joined03:12
← nadavoid left03:12
→ dwaite joined03:13
← phaed left03:14
→ Araxia_ joined03:14
← pheaver left03:15
→ nadavoid joined03:15
← Guest28785 left03:17
→ devanjan joined03:22
← Utkarsh left03:23
→ juangiordana_ joined03:24
→ jony joined03:24
← bcardarella left03:25
← juangiordana left03:25
→ normanrichards joined03:26
→ noahcampbell joined03:27
← doener left03:27
← RaceCondition left03:29
→ Remixman joined03:31
← kuhrt left03:31
→ eletuchy joined03:32
→ mattly joined03:32
→ pheaver joined03:33
→ doener joined03:33
→ Utkarsh joined03:33
← shiba_yu36 left03:34
← vanmik left03:35
→ intripoon joined03:36
→ Schmallon joined03:37
← mattly left03:37
← Elfe left03:38
← intripoon_ left03:39
→ mdel joined03:39
→ mw_ joined03:40
← mw_ left03:40
→ mw_ joined03:40
← Remixman left03:40
← up_the_irons left03:41
← BobJelly left03:41
← saysjona1han left03:42
← juangiordana_ left03:42
← devanjan left03:42
→ shiba_yu36 joined03:43
→ simplechat joined03:43
← simplechat left03:43
→ simplechat joined03:43
→ BobJelly joined03:45
← torvalamo left03:45
← Swimming_bird left03:47
← jmil_ left03:48
→ torvalamo joined03:50
→ vitorpc joined03:50
← obrienmd left03:51
← scelis left03:51
← vitor-br left03:51
→ timj_ joined03:52
← tk8 left03:54
← cylence left03:55
→ ISF_ joined03:56
→ tek0 joined03:56
→ Elfe joined03:56
← timj left03:56
← WebDragon left03:57
← xiangfu left03:58
← ISF left03:58
→ hasrb joined03:58
→ smolyn joined03:59
← yogione left03:59
← ISF_ left04:01
← tchan left04:02
→ tchan joined04:03
← torvalamo left04:03
DrNick2 → DrNick04:03
← vitorpc left04:05
→ bloopletech joined04:05
← VladDrac left04:06
→ VladDrac joined04:06
bloopletech I'm trying to use git-apply to apply some patches but it doesn't seem to be doing anything. I just get a bunch of whitespace warnings, and git diff / git status / git log report no changes04:06
→ Swimming_bird joined04:07
→ cylence joined04:07
→ torvalamo joined04:08
→ ISF joined04:08
← sternocera left04:08
bloopletech is there some trick to git-apply? I'm not supplying any options to git-apply either04:08
→ devanjan joined04:10
← cylence left04:10
← eletuchy left04:10
← zorzar_ left04:10
→ jayzawrotny joined04:10
jayzawrotny Hello, I'm a git noob and I have a small question. What does git commit -a do?04:10
cirwin git apply --whitespace=fix04:10
^ bloopletech04:11
jayzawrotny: it commits all changes to all files that you have added to the repository04:11
jayzawrotny oh, so the ones it's tracking changes on from a previous add statement then, right cirwin?04:11
monokrome avoids -a04:11
jayzawrotny Oh, is it a bad practice?04:12
cirwin jayzawrotny: yes, or from the previous commit04:12
jayzawrotny Ah04:12
monokrome Not really, as long as you use `git status` before you use -a to make sure that you don't commit things that you didn't expect04:12
cirwin I use it all the time, I had to write myself a "git uncommit" though bceause I often committed too many changes by accident04:12
← leo2007 left04:12
jayzawrotny Ah, I see, that makes sense. So use it with care then :)04:13
monokrome yep04:13
ka1ser Is there an easy way to create a commit that reverts another one that already exists?04:13
cirwin git revert <commit>04:13
→ ki__ joined04:14
ka1ser =o04:14
nice04:14
lol04:14
thanks cirwin04:14
cirwin thank git, not me :p04:14
monokrome lol04:14
ka1ser yeah04:15
they think of everything04:15
=P04:15
→ cutmail joined04:15
← torvalamo left04:15
← bloopletech left04:15
ka1ser s/of/about04:15
→ dreiss joined04:18
← ki__ left04:19
jayzawrotny Anyways, thanks very much for your help. I really appreciate it :)04:20
← jayzawrotny left04:20
← pabst left04:21
← neorab left04:21
← pheaver left04:22
→ zorzar_ joined04:23
→ eletuchy joined04:23
→ Yuffster joined04:23
← eletuchy left04:23
→ OOPMan joined04:24
→ BlindOracle joined04:26
← zombor left04:27
← Swimming_bird left04:27
← dwaite left04:28
→ xerothermic joined04:29
← ceej left04:29
← smolyn left04:31
← friskd left04:31
← devanjan left04:31
→ saysjonathan joined04:32
← tro left04:32
← stepnem left04:33
→ coderdad joined04:33
→ neorab joined04:34
← javaanse_jongens left04:34
→ vanmik joined04:34
→ obrienmd joined04:34
← ka1ser left04:34
→ stepnem joined04:36
← saysjonathan left04:36
→ javaanse_jongens joined04:37
← shadowhand left04:37
→ saysjonathan joined04:38
→ tro joined04:38
→ hunterloftis joined04:38
hunterloftis Okay so, made some changes, a few commits, decided to abort them, checked out maybe 3 commits previous, want to commit this to the head now to erase those later changes... why is it so hard?04:38
→ project2501a joined04:38
← obrienmd left04:39
cirwin hunterloftis: because you want to use git reset, not git checkout04:39
(or git revert, if you want to keep the aborted commits in history for posterity)04:39
← dreiss left04:40
hunterloftis cirwin: thanks... I just did revert.04:40
Why would that not just be implemented as a checkout then commit?04:40
I guess that's probably a deep discussion04:40
Using git just out of necessity, it gets frustrating04:40
cirwin so checkout checks files out of git into the current directory04:41
and sometimes changes HEAD to point to a different place04:41
whereas reset moves the current branch's pointer to a different commit04:41
it's quite simple underneath, but the UserInterface is full of leaky abstractions and yuck04:41
← fr0sty left04:42
← saysjonathan left04:42
hunterloftis I seem to have totally lost some files04:43
No idea how04:43
Damn it git04:43
cirwin did you commit them?04:43
hunterloftis I reverted to a time when those files existed, but they're still not there04:43
cirwin if so they are not lost, just hiding :p04:43
what does git status say?04:43
hunterloftis on branch master, nothing to commit04:43
← nadim__ left04:43
cirwin ok — try "git log --all -- file_name.youve_lost"04:44
it will tell you which commits that file is used it04:44
*in04:44
→ nadim__ joined04:44
→ bad_mantis joined04:44
hunterloftis Yep04:45
← Heimidal left04:45
hunterloftis should I revert to that commit?04:45
cirwin no, "reset"04:45
ChanServ set mode: +v04:45
jspiros_ → jspiros04:45
cirwin revert creates new commits that undo old commits — reset moves your current branch to a better place04:45
hunterloftis I don't know that I follow the difference04:47
but I'll just reset (hash)04:47
my file still isn't there04:47
→ fr0sty joined04:48
cirwin now if you do git status04:48
← pleonasm left04:48
cirwin you will see that it should be, and you can (finally) do git checkout to get it back04:48
→ pleonasm joined04:48
← pleonasm left04:48
← ISF left04:48
hunterloftis I'm not sure I follow04:49
should I checkout that same revision?04:49
that I just reset to?04:49
cirwin yes04:49
← mw_ left04:49
cirwin you will probably need -f04:49
→ lessthanzero joined04:50
cirwin there are optimisations, but this way is safer given that I can't see what I'm doing :p04:50
hunterloftis I just did git checkout -f (that hash) and still no dice04:51
I appreciate your help though04:51
← kanru1 left04:51
→ kanru joined04:52
← noahcampbell left04:52
cirwin no worries — if it all gets too much, do a tutorial04:52
hunterloftis I may look for a svn or mercurial version of github04:53
This is costing my team huge amounts of time04:53
Just trying to figure out simple things in git that should be simple04:53
Like, there's a file I deleted 3 commits ago. I just want the file back.04:53
cirwin ah, so this commit you just reset to was the commit in which you deleted the file?04:53
you need to go one back from that then..04:54
git is simple04:54
→ ISF joined04:54
cirwin and if you take out half an hour to learn it, you will get a very long way04:54
but it's not easy to pick up from the "I have task X I want to do" direction04:54
which is a shame04:54
← TanEk left04:54
cirwin bitbucket.org is mercurial github04:55
→ zeshoem joined04:55
hunterloftis awesome, thanks - and I just followed your directions again on the previous commit, and got my files back04:55
cirwin that04:55
that's my fault04:55
hunterloftis But honestly it just seems like... software that's not good from a "I have a task" direction is just poorly designed software04:55
Without your help I would never have figured that out04:55
cirwin it's poorly designed user interface04:55
← ISF left04:55
hunterloftis I'm still not sure exactly why it worked04:55
cirwin so you have to understand what goes on underneath04:56
→ ISF joined04:56
cirwin try reading "git for computer scientists"04:56
it's a blog post that everyone raves about04:56
it'll hopefully tell you the "why", and then you can work out the "how"04:56
hunterloftis hmm ok thanks04:57
← kukks left04:57
← hobodave left04:57
→ hobodave joined04:57
hunterloftis damn it04:59
it restored the friggin directory04:59
not the files in it04:59
I've gotta give up on this04:59
it's midnight and I've been fighting with git for hours04:59
cirwin I have a shell script that can tell you which commit each file was deleted in05:00
maybe that would be useful?05:00
or would you prefer to sleep05:00
hunterloftis hah that would be awesome05:00
← ISF left05:00
cirwin http://dpaste.org/qcnA/05:00
in order to get the files back, you'll need to go to the commit before they were deleted05:00
which is <sha1>^ with a caret after the hash05:01
← jfkw left05:01
← engla left05:01
← kanru left05:02
→ saysjonathan joined05:03
→ kanru joined05:03
← jkprg left05:05
← Marquel left05:07
← coderdad left05:07
← saysjonathan left05:07
← nullvariable left05:08
→ saysjonathan joined05:08
→ infid joined05:08
→ Heimidal joined05:08
← lessthanzero left05:09
→ avinashhm joined05:10
← Heimidal left05:12
→ jkprg joined05:12
← jkprg left05:13
← saysjonathan left05:13
→ _1094kms joined05:13
→ saysjonathan joined05:13
hunterloftis cirwin: is there a way to get git to sort of "refresh?"05:15
I removed some submodules05:15
But it won't pick up those files to add them05:15
cirwin ahhh submodules ahhhh05:15
hunterloftis I sort of need it to "git init" again05:15
Yeah I fucking hate submodules05:15
cirwin I've no idea how they work, sorry.05:16
hunterloftis hah thanks05:16
neither do we :/05:16
cirwin I don't think anyone does..05:16
← avinashhm left05:17
← saysjonathan left05:18
→ mdance joined05:18
→ pheaver joined05:18
→ hobodave_ joined05:19
→ Utkarsh_ joined05:19
→ warthog9 joined05:20
← hobodave left05:21
→ dreiss joined05:22
hobodave_ → hobodave05:22
← Utkarsh left05:23
← mdance left05:23
→ mdance joined05:23
mdance i have an opensource project tracked with GIT and was not sure which files were cache/tmp files, so my initial commits are tracking the temp files. I would like to stop these files from being tracked, so I added them to .gitignore, and tried a git rm --cached cache/tmp items. The problem I am running into is if I do a git pull on the live server, it removes these files, is there any way to have git pull ignore deletions for one pull or so05:23
mething similar?05:23
→ nullvariable joined05:24
cirwin mdance: heh, interesting question — not that I know of.05:24
can you just temporarily move the temp files elsewhere, then git pull, then move the files back?05:24
infid how can i checkout my git repo from my desktop to my laptop? when i do 'git clone infid@infidsnetwork:~/git/test' it says 'warning: you appear to have cloned an empty repository'05:24
but it's not empty05:25
← Eridius left05:25
→ BiggFREE joined05:25
cirwin you want to clone the .git directory05:25
:~/git/test/.git05:25
→ up_the_irons joined05:25
infid i tried that05:25
← Nayruden left05:25
cirwin and?05:25
infid says the same thing05:25
cirwin you definitely have committed your changes?05:25
infid yes05:26
→ natek joined05:26
cirwin bizarre05:26
→ xiangfu joined05:26
wereHamster infid: what does 'git show-ref' say in he repo on the server?05:26
← natek left05:26
infid er i take it back05:26
git status said none of them were committed, weird, thought i did05:27
← project2501a left05:27
cirwin ah :)05:27
infid thanks05:27
why is git so much harder to use than svn05:27
hunterloftis infid: seconded05:27
cirwin because it's more powerful05:27
infid i like that it's more powerful but it's kinda crazy05:27
wereHamster infid: it's not harder to use, but it's harder to learn05:28
infid wereHamster: yeah that's more accurate05:28
→ harshu_ joined05:29
← hobodave left05:32
→ avinashhm joined05:33
← nadavoid left05:33
→ saysjonathan joined05:33
→ hobodave joined05:34
→ project2501a joined05:37
← saysjonathan left05:38
→ atmos joined05:38
← project2501a left05:39
← jony left05:40
mdance hmmmm just tried an git update-index --assume-unchanged and thats not going to work as it doesnt work on directories05:40
→ LongBeach joined05:40
← hornswoggles left05:40
atmos is there any way to do something like 'git ls-remote [email@hidden.address] refs/heads/master@{1day}' ?05:41
→ HG` joined05:43
← cirwin left05:46
← BobJelly left05:49
← tty2 left05:50
→ tannern joined05:51
→ cenuij joined05:51
← cenuij left05:51
→ cenuij joined05:51
→ tty1 joined05:51
→ freimatz joined05:52
← sivy left05:52
tannern I ran into this once before I just dont remember how i fixed it, Im using `git clone ...` in a deploy script over ssh and I'm getting "host key verification failed"05:54
→ Titosemi joined05:54
→ BobJelly joined05:55
← madewokherd left05:56
→ saysjonathan joined05:56
← mdance left05:56
→ slonopotamus joined05:57
→ Pupeno joined05:58
← rayners left05:58
→ psankar joined05:59
← psankar left05:59
→ psankar joined05:59
← saysjonathan left06:00
← Textmode left06:02
→ captproton joined06:05
captproton hi all06:05
Auro1 PARTY06:05
← Auro1 left06:05
← sftp left06:05
captproton how do I pull the remote master and have it overide the local master?06:06
atmos git fetch origin; git reset --hard origin/master06:07
captproton howdy atmos! How's Hancock?06:08
→ saysjonathan joined06:08
atmos shit i dunno, haven't touched it in forever :)06:08
→ ThomasLocke joined06:09
captproton atmos: funny I mention an historiclal name and someone named Locke enters06:09
→ boombatower joined06:10
← boombatower left06:10
→ boombatower joined06:10
captproton atmos: would you still recommend using Hancock?06:10
atmos: is there something better now?06:10
atmos i dunno, there's hmac auth that a lot of people use06:10
similar to the aws api06:10
hancock works fine if you're ok with openid as a protocol and not calling it openid to your users06:10
captproton atmos:what's up with openid anyway? DHH doesn't like it anymore06:11
→ csanz joined06:12
atmos that's kind of a loaded question but as a protocol it's not the worst06:12
→ wack joined06:12
atmos openid on the internet is kind of a shit show because it's too difficult for users06:12
← saysjonathan left06:12
→ werdan7 joined06:12
_rane indeed06:12
atmos things like facebook and google auth trump them because epople can say "yeah i have one of those"06:12
DrNick there's also the fact that everybody is an openid provider, but nobody is an openid consumer06:13
← bad_mantis left06:14
devians the stack overflow openid isnt so bad. like 2 clicks.06:15
→ TomyLobo joined06:16
→ bousquet joined06:16
→ tiagobutzke joined06:17
→ project2501a joined06:17
← shiba_yu36 left06:19
← mdel left06:20
captproton devians: it's cool if you're a tech head, but atmos is right.06:20
→ friskd joined06:21
devians its kind of like the story of how amazon 1 click came about.06:21
← atmos left06:22
captproton bye atmos06:22
his hip hop preso is pretty good06:22
→ saysjonathan joined06:25
→ sechrist joined06:26
← slonopotamus left06:26
→ Goliath joined06:27
→ jjuran joined06:27
→ cirwin joined06:27
← Khisanth left06:29
← saysjonathan left06:30
→ saysjonathan joined06:30
→ Khisanth joined06:30
← jmorris left06:30
← BiggFREE left06:32
← saysjonathan left06:35
← gbacon_ left06:35
← Yuffster left06:35
→ saysjonathan joined06:36
← psankar left06:36
→ mdel joined06:36
→ gbacon_ joined06:37
← mikepack left06:37
→ hakunin joined06:38
→ Zhonghua joined06:38
← n3m left06:38
← halstead left06:38
← Zhonghua left06:38
→ Remixman joined06:39
← saysjonathan left06:40
→ saysjonathan joined06:41
← xiangfu left06:43
← yuwang left06:44
→ pznate joined06:46
← saysjonathan left06:46
← pznate left06:46
→ saysjonathan joined06:46
→ unreal-dude joined06:47
unreal-dude any suggestions here for setting up and configuring gitosis on debian lenny? google is finding nothing that is especially useful...06:48
→ Halerazematt joined06:50
← mdel left06:50
← freimatz left06:51
→ TrineLocke joined06:51
→ ludde joined06:51
← saysjonathan left06:51
→ freimatz joined06:51
→ jony joined06:53
← project2501a left06:53
← captproton left06:54
← Bass10 left06:54
→ project2501a joined06:58
← project2501a left06:58
→ xrogaan joined07:00
→ slonopotamus joined07:00
xrogaan sup07:00
← cutmail left07:01
xrogaan is there a way to use git svn with an already existing git repository by commiting the current history into the svn repository ?07:02
I know, it's odd, but I need to put the sources into a svn repository07:02
→ envi joined07:03
← zivester_ left07:03
← Ratler_ left07:05
→ bronson joined07:05
→ Ratler joined07:05
→ cutmail joined07:07
→ JoeCoolDesk1 joined07:10
← JoeCoolDesk1 left07:10
← slonopotamus left07:11
← tiagobutzke left07:12
→ cbreak_work joined07:12
← guardian1 left07:12
← jony left07:14
← Soliton left07:14
→ project2501a joined07:14
← urkud_ left07:15
← hakunin left07:16
← marc_kdab_ left07:17
← MikhX left07:17
→ leavittx joined07:17
→ thiago_home joined07:18
→ Zinuzoid joined07:18
→ tk8 joined07:18
→ saysjonathan joined07:18
→ Utkarsh joined07:18
→ Soliton joined07:18
→ opalka joined07:19
← opalka left07:19
→ opalka joined07:19
← Utkarsh_ left07:20
→ dc5ala joined07:20
← pheaver left07:21
← c00kiemon5ter left07:21
← sudonotplease left07:21
→ c00kiemon5ter joined07:22
← c00kiemon5ter left07:22
→ c00kiemon5ter joined07:22
← saysjonathan left07:22
→ saysjonathan joined07:23
→ Engin joined07:24
← Will| left07:24
← boombatower left07:26
→ Will| joined07:27
← saysjonathan left07:28
← hunterloftis left07:28
→ shiba_yu36 joined07:28
← LouisJB left07:28
→ boombatower joined07:28
← boombatower left07:28
→ boombatower joined07:28
→ rchavik joined07:29
→ hunterloftis joined07:29
← dreiss left07:29
→ joel_falcou joined07:30
→ tiagobutzke joined07:32
← hasrb left07:32
→ t_b joined07:36
→ ngng joined07:36
→ DrNick2 joined07:36
← DrNick left07:38
→ ph^ joined07:40
→ Marquel joined07:42
← jmil left07:42
← emptyflask left07:42
→ ayanes joined07:44
← ayanes left07:44
→ ayanes joined07:44
→ kipras_ joined07:44
→ mziulu joined07:46
← kipras_ left07:46
→ kipras joined07:47
→ chrischambers joined07:47
← Mage left07:48
← gordonjcp left07:48
→ petrux joined07:48
→ disappearedng joined07:48
← disappearedng left07:48
→ disappearedng joined07:48
→ saysjonathan joined07:49
← wack left07:51
cebewee xrogaan: I don't know if you can do git svn init on an already used repository (look at the man page); but you could07:52
→ Badiss joined07:52
← phantomcircuit left07:52
→ phantomcircuit joined07:53
cebewee init a new repository, put the svn stuff there, pull in the commits from your existing repository and svn dcommit it07:53
← saysjonathan left07:53
→ jony joined07:53
cebewee but i don't know if git-svn can handle merges and stuff07:53
→ skoop joined07:54
← sechrist left07:57
→ saschpe joined07:57
← steph021 left07:58
→ saysjonathan joined08:00
→ EricInBNE joined08:00
← Badiss left08:02
← saysjonathan left08:04
→ guardian joined08:05
← Schmallon left08:05
→ pfrenssen joined08:08
→ henrikbjorn joined08:11
→ Badiss joined08:13
← vanmik left08:13
→ vanmik joined08:13
→ teweWork joined08:14
← NickyP left08:16
← stepnem left08:16
→ psankar joined08:16
← psankar left08:16
→ psankar joined08:16
→ MikhX joined08:17
← acid_ left08:17
→ saysjonathan joined08:17
← tiagobutzke left08:18
→ sechrist joined08:18
→ marc_kdab_ joined08:19
→ nilesh joined08:20
nilesh How do I allow a group to write to all repos except gitosis-admin?08:20
→ linusoleander joined08:21
← saysjonathan left08:21
← cirwin left08:22
→ NickyP joined08:22
→ airborn joined08:23
→ shruggar joined08:24
← ericlee left08:24
← wliao left08:25
→ ericlee joined08:25
→ serard joined08:25
→ wliao joined08:25
→ rendar joined08:26
← wliao left08:26
← lanthan_ left08:27
← linusoleander left08:27
← _rane left08:27
→ tewecske joined08:27
→ wliao joined08:27
← cutmail left08:28
→ runde joined08:28
← wliao left08:28
→ wliao joined08:29
← wliao left08:30
→ wliao joined08:31
→ larsrh joined08:31
← spike left08:32
← wliao left08:32
→ thierryp joined08:32
→ wliao joined08:32
→ cilly joined08:32
→ cutmail joined08:32
← mziulu left08:33
→ saysjonathan joined08:33
← ph^ left08:34
← iAsteris_ left08:35
→ kreemer joined08:36
→ Halerazematt_ joined08:36
← jony left08:36
→ jony joined08:36
→ radarek joined08:36
adamholt → adamholt_away08:37
adamholt_away → adamholt08:37
← opalka left08:37
→ hegge joined08:37
← saysjonathan left08:38
→ nevyn^ joined08:38
→ Gunni joined08:38
→ petrux_ joined08:39
→ ph^ joined08:39
← Halerazematt left08:39
→ mastro joined08:41
← t_b left08:41
← petrux left08:41
→ goshawk joined08:41
→ deuterium joined08:41
→ bradleyayers joined08:41
→ iAsterisk joined08:42
→ dpino joined08:43
→ tvw joined08:43
→ psoo joined08:43
adamholt → adamholt_away08:44
adamholt_away → adamholt08:44
→ saysjonathan joined08:45
← amerine left08:46
← logankoe1ter left08:46
→ linusoleander joined08:47
← ph^ left08:48
← jjuran left08:49
← saysjonathan left08:49
← thiago_home left08:50
→ rane_ joined08:50
→ ph^ joined08:50
← OOPMan left08:51
adamholt → adamholt_away08:52
adamholt_away → adamholt08:52
← shruggar left08:52
← bradleyayers left08:53
adamholt → adamholt_away08:53
→ makkes joined08:53
adamholt_away → adamholt08:54
→ logankoester joined08:55
→ priidu joined08:55
← vanmik left08:56
→ saysjonathan joined08:57
← justin_george left08:57
← makkes left08:57
← rns left08:58
→ davzie joined08:58
→ stepnem joined09:00
→ Nickeeh joined09:00
→ tatsuya joined09:01
← saysjonathan left09:01
→ ch077179 joined09:02
→ saysjonathan joined09:02
← avinashhm left09:02
→ jjuran joined09:02
→ fr0sty|mobile joined09:03
Nickeeh I accidently merged A with some other local branch B without --no-ff, but I actually want to. How do I redo my merge? Should I reset HEAD to the head of A to where it was before fast-forwarding B on it, then git merge --no-ff?09:05
I haven't pushed my changes, of course.09:05
cbx33 that's what I would think09:05
but I'm no expert09:05
cbreak_work it's a reasonable thing to do09:05
← robotarmy left09:05
cbreak_work you could also chose to live with a nicely linear history09:06
← harshu_ left09:06
← saysjonathan left09:06
→ saysjonathan joined09:07
→ leo2007 joined09:09
→ doub joined09:09
Nickeeh cbreak_work: well, I like to know where my work was done. All the commits on branch A are commit's we're gonna release.09:10
I think it looks nice.09:10
→ harshu joined09:10
doub Hi! I have some trouble related to msysgit, is this the right place to ask for help?09:11
← saysjonathan left09:11
→ rns joined09:11
→ ExtraSpice joined09:12
← dr_lepper left09:12
→ saysjonathan joined09:12
→ dr_lepper joined09:13
← dr_lepper left09:13
→ dr_lepper joined09:13
→ sechrist_ joined09:13
← shiba_yu36 left09:13
→ j416 joined09:15
→ shiba_yu36 joined09:17
← saysjonathan left09:17
← sechrist left09:17
→ RobertMeta joined09:18
← RobertMeta left09:18
→ RobertMeta joined09:18
→ saysjonathan joined09:18
→ Saur joined09:18
fr0sty|mobile doub: it's _a_ place to ask...09:19
doub ok, i'll have a try then :-)09:19
← Saur left09:19
doub i'm accessing a private gitorious repository with msysgit over ssh, and i can do most operations without problem09:20
← MetaCosm_ left09:20
→ nd___ joined09:20
→ dpino_ joined09:21
← dpino left09:21
doub but when I try to push, the command ends in a "fatal: The remote end hung up unexpectedly". the push actually succeeded (the remote repository is in the correct state), but the remote branch in my local repository has not been updated (i need to do a git fetch to update it)09:21
← saysjonathan left09:22
Arrowmaster sounds more like a gitorious problem than a msysgit problem09:22
doub I think the problem appeared when the remote repo admin added some push hooks that display sends some text to the client09:22
← nd___ left09:22
Arrowmaster sounds extremely likely09:22
doub all my colleagues can push without problem (even some on windows), so there is a problem on my computer09:23
→ nd__ joined09:23
→ saysjonathan joined09:23
nd__ Aloha… I wonder, how can I use a separate ssh key for pushing? I do not want to use my default key for one project…09:23
Nickeeh ...09:24
nd__: why would you not want to do that?09:24
charon nd__: man ssh_config09:25
nd__ charon: Only for working with one repository...09:26
charon nd__: you can define aliases of sorts that refer to some real hostname, and use that09:26
nd__ Nickeeh: Because it seems I cannot add my public key to two separate projects on github and so far I don't see the possiblity to have commit access on both accounts.09:26
charon nd__: cf. the "hostname" setting09:27
nd__ charon: Ah, nice... that helps. Thanks09:27
← saysjonathan left09:27
→ Farsinsen joined09:28
← Farsinsen left09:28
← nevyn^ left09:28
→ nevyn^_ joined09:28
← samhassell left09:29
→ spaceonline joined09:29
→ darkredandyellow joined09:30
→ benishor joined09:31
→ ChUrP joined09:32
→ f00li5h joined09:32
← nilesh left09:33
← boombatower left09:34
← pfrenssen left09:35
→ mabrand joined09:35
← darkredandyellow left09:38
→ darkredandyellow joined09:39
← petrux_ left09:39
→ t_b joined09:39
→ avinashhm joined09:39
→ ningia joined09:40
→ saysjonathan joined09:41
← ketas left09:43
← BobJelly left09:43
→ jhaddad_ joined09:44
← jhaddad_ left09:44
← jhaddad left09:45
← MikhX left09:45
← saysjonathan left09:45
→ jhaddad joined09:45
→ smuf joined09:45
→ bluenovember joined09:48
→ BobJelly joined09:48
← ch077179 left09:49
← logankoester left09:50
→ opalka joined09:50
← opalka left09:50
→ opalka joined09:50
→ OOPMan joined09:52
→ urkud_ joined09:54
← marc_kdab_ left09:54
← opalka left09:55
→ randomsort joined09:56
→ pfrenssen joined09:57
← urkud_ left09:58
→ saysjonathan joined09:59
→ logankoester joined10:01
← saysjonathan left10:03
→ saysjonathan joined10:04
← a9b left10:05
→ shruggar joined10:07
→ marc_kdab_ joined10:08
← saysjonathan left10:08
→ saysjonathan joined10:09
← ngng left10:12
→ curtana joined10:12
→ ngng joined10:12
← ngng left10:12
→ bradleyayers joined10:13
→ mkramer joined10:13
← saysjonathan left10:14
← angelsl left10:17
← hyperair left10:18
← bousquet left10:18
→ Behi joined10:18
→ saysjonathan joined10:21
→ angelsl joined10:21
→ fr0sty|android joined10:23
← nevyn^_ left10:25
→ nevyn^ joined10:25
← saysjonathan left10:25
→ saysjonathan joined10:26
← Buglouse left10:26
← serard left10:26
← fr0sty|mobile left10:26
← cenuij left10:27
→ sudonotplease joined10:28
← jds left10:29
← saysjonathan left10:30
→ drizzd_ joined10:31
→ saysjonathan joined10:31
← hyperboreean left10:32
→ hyperboreean joined10:32
← Utkarsh left10:32
← drizzd left10:34
← cutmail left10:34
← saysjonathan left10:36
← Remixman left10:36
→ saysjonathan joined10:37
→ xiong joined10:38
→ Halerazematt joined10:38
← tvw left10:38
← bgerber left10:40
← saysjonathan left10:42
← Halerazematt_ left10:42
→ saysjonathan joined10:42
→ bgerber joined10:42
→ ketas joined10:43
← wliao left10:44
← kipras left10:44
→ kipras_ joined10:44
← saysjonathan left10:46
→ mizar99 joined10:47
← Weust`_ left10:47
→ icwiener joined10:48
← mizar99 left10:48
← stliu left10:49
← shiba_yu36 left10:50
→ saysjonathan joined10:53
→ jhoepken joined10:54
← saysjonathan left10:58
← dwmw2_gone_ left10:58
→ dwmw2_gone_ joined10:58
→ samhassell joined10:58
→ saysjonathan joined10:59
← deryl left10:59
→ deryl joined11:00
← linusoleander left11:01
→ Titosemi1 joined11:01
← harshu left11:01
→ urkud_ joined11:02
→ marcob joined11:02
← saysjonathan left11:03
→ harshu joined11:03
← Titosemi left11:04
→ petrux joined11:04
→ saysjonathan joined11:04
→ harshu___ joined11:04
→ alfadir joined11:05
→ koltroll joined11:08
← deryl left11:08
→ pulse00 joined11:08
← saysjonathan left11:09
pulse00 hi there. is there a way i can make git overwrite untracked files when i pull a remote branch? we're getting "untracked working tree files would be overwritten by merge"11:09
→ saysjonathan joined11:09
→ juangiordana joined11:10
rane_ try -f11:11
pulse00 tried that already, getting the same error11:11
rane_ can't you just remove the files?11:11
bremner pulse00: you could see man git-clean11:11
jast pulse00: the 'git-clean' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clean.html [automatic message]11:11
→ deryl joined11:12
pulse00 it's happening too often, and manually removing them would take too long. we're looking for some automated to get rid of the problem11:12
rane_ git clean -f11:12
removes untracked files11:12
pulse00 the problem with that is that it also seems to remove files from .gitignore11:12
bremner no11:12
not without -x11:13
pulse00 ok, when i run clean -f i get "not removing path/to/file"11:13
Nickeeh clean -fd11:13
"directories"11:13
it won't remove directories bij default11:13
pulse00 well, that definitifely removes files from gitignore in our setup11:14
← saysjonathan left11:14
Nickeeh it won't11:14
→ saysjonathan joined11:14
bremner pulse00: that would be a serious bug11:14
fr0sty|android indeed11:14
→ RaceCondition joined11:15
→ Determinist joined11:16
pulse00 well, then it has a bug, because it just deleted the files from .gitignore11:17
→ serard joined11:18
pulse00 our gitignore has the following line: lib/model/doctrine/base/* and "git clean -f -d" just deleted the directory11:18
bremner if you can duplicate this, you should send a message to the git mailing list.11:18
← saysjonathan left11:18
fr0sty|android was the directory empty?11:19
pulse00 no11:19
it contains autogenerated files11:19
→ saysjonathan joined11:19
fr0sty|android and there are no other .gitignore files in that path?11:20
pulse00 there are a lot more11:20
← Silowyi left11:20
pulse00 ah sorry, not in that path11:20
← bradleyayers left11:20
pulse00 but there are other paths in .gitignore11:20
fr0sty|android any negating patterns? (start with !)11:21
pulse00 no11:21
fr0sty|android can you paste an example to gist.github.com?11:21
→ d0k joined11:22
→ bentob0x joined11:22
fr0sty|android include git --version11:22
pulse00 an example of the .gitignore?11:23
fr0sty|android create a file, demonstrate that it is being ignored, run git clean and show that it was removed.11:23
← saysjonathan left11:23
← gretch left11:24
fr0sty|android including your .gitignore might help too,11:24
your info/excludes as well.11:25
← deryl left11:25
→ saysjonathan joined11:25
← josue left11:26
→ winterweaver joined11:29
← saysjonathan left11:29
← Ruudjah left11:30
→ Halerazematt_ joined11:30
→ saysjonathan joined11:30
→ linusoleander joined11:32
← Halerazematt left11:32
→ consolers joined11:32
pulse00 https://gist.github.com/82923711:33
→ tvw joined11:33
← saysjonathan left11:35
← davzie left11:35
→ deryl joined11:35
pulse00 update https://gist.github.com/829237 which shows that the file existed11:35
→ davzie joined11:36
→ j0su3 joined11:36
← error_developer_ left11:37
fr0sty|android can you humor me and do git status before git clean?11:39
← psankar left11:39
← consolers left11:41
→ workmad3 joined11:41
fr0sty|android man git-clean11:41
jast the 'git-clean' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clean.html [automatic message]11:41
→ saysjonathan joined11:42
← normanrichards left11:42
→ fr0sty|mobile joined11:43
← fr0sty|android left11:46
← saysjonathan left11:46
fr0sty|mobile pulse00: you have me intrigued, but i have to run. I'll be back later.11:47
→ saysjonathan joined11:47
pulse00 updated version with git status before clean: https://gist.github.com/82923711:48
← BobJelly left11:50
← saysjonathan left11:51
→ dreiss joined11:51
→ saysjonathan joined11:52
→ EugenA joined11:53
→ Dave^| joined11:53
EugenA my .gitignor file is being ignored by git, i don't want to commint some files11:53
← smuf left11:53
EugenA why?11:53
rane_ you have those files tracked11:54
EugenA rane_: git rm ?11:55
→ sjokkis_ joined11:55
→ BobJelly joined11:55
← sjokkis_ left11:56
← saysjonathan left11:56
→ sjokkis joined11:57
sjokkis i know i can partially add a file with git add -p [file], however, this only works if the file is already under version control. how do i do this with a file that isn't already in git?11:57
→ saysjonathan joined11:57
rane_ you don't because the whole file is new to git11:57
→ Error404NotFound joined11:58
rane_ there are no hunks11:58
sjokkis rane_: but i only want to stage part of it. is that impossible?11:58
rane_ edit the file to match what you want to stage11:58
← samhassell left12:00
sjokkis thanks12:01
← saysjonathan left12:02
→ hyperair joined12:02
← workmad3 left12:02
→ workmad3 joined12:03
→ marcosiegl joined12:04
← koltroll left12:04
← sechrist_ left12:05
→ Yuuhi joined12:06
→ MacGyverNL_ joined12:06
← MacGyverNL_ left12:06
→ MacGyverNL_ joined12:06
← workmad3 left12:06
→ saysjonathan joined12:09
← project2501a left12:09
← MacGyverNL left12:09
→ jcromartie joined12:09
→ djkk joined12:10
← djkk left12:11
→ _ollie joined12:11
← dreiss left12:11
← jdav_gone left12:11
← goshawk left12:11
→ dreiss joined12:12
← _ollie left12:13
← saysjonathan left12:13
fr0sty|mobile pulse00: 1.7.3.2 has the same issue, it seems. I will look into this more later...12:13
← hyperair left12:14
→ saysjonathan joined12:14
→ Saur joined12:15
→ samhassell joined12:15
← Saur left12:15
→ jdav_gone joined12:16
EugenA git commit --amend --author='Your Name [email@hidden.address]12:16
what means --amend?12:16
→ hyperair joined12:17
← radarek left12:17
← saysjonathan left12:18
→ cwillu_at_work joined12:18
pulse00 fr0sty|mobile: thanks12:18
LoRe EugenA: "to amend" is a synonym for "to correct"12:19
→ saysjonathan joined12:19
← _Vi left12:20
← Determinist left12:22
EugenA oh.. ok, thanks :-)12:22
→ Determinist joined12:22
wereHamster EugenA: see man git-commit12:23
jast EugenA: the 'git-commit' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-commit.html [automatic message]12:23
← saysjonathan left12:23
wereHamster sjokkis: git add -N <file> && git add -p <file>12:24
rane_: ^^^12:24
EugenA visit linus torvalds sometimes this irc channel?12:24
wereHamster no12:24
→ saysjonathan joined12:25
← tk8 left12:28
sjokkis wereHamster: i already fixed it, but that is more elegant12:29
← saysjonathan left12:30
→ Utkarsh joined12:30
→ saysjonathan joined12:30
→ josue joined12:32
← j0su3 left12:33
← saysjonathan left12:35
→ saysjonathan joined12:35
← deuterium left12:38
← saysjonathan left12:40
→ saysjonathan joined12:40
← stamina__ left12:40
→ stamina_ joined12:41
← EugenA left12:42
← dreiss left12:43
← jcromartie left12:44
← j416 left12:44
← saysjonathan left12:44
sgronblo Linus probably only uses mailing lists12:45
→ saysjonathan joined12:45
alfadir need some advice how to set up two git project that depend on eachother. www-templates (general templates) www-site (site specific implementation). i would like to "vendor drop" in changes when the www-template gets updated. directory structure is different but www-template:/style/x would be placed in www-site:path/style/y. is this at all possible or is there another recepie availible ?12:46
→ tb_1 joined12:46
← teuf left12:46
→ ymasory joined12:46
← ymasory left12:46
cebewee alfadir: have you already looked into submodules?12:47
alfadir yes.. but.. have not really been able to map the method to how i should do it.. but will read more12:48
reading http://book.git-scm.com/5_submodules.html12:48
← jony left12:48
alfadir and http://stackoverflow.com/questions/769786/vendor-branches-in-git12:49
← t_b left12:49
alfadir a submodule is a full git ? i only need a part.. of the git .. the template dir.12:49
← saysjonathan left12:50
→ opalka joined12:50
← opalka left12:50
→ opalka joined12:50
← _Ants_ left12:50
→ saysjonathan joined12:51
alfadir cebewee: answer 2 from Paul at http://stackoverflow.com/questions/769786/vendor-branches-in-git seems to make most sense. but not sure how to set that up12:52
→ EugenA joined12:52
alfadir well will read more about subtree merge12:53
EugenA i'd like to commint only subdir from my project dir12:53
← saysjonathan left12:55
alfadir kind of this unanswered question http://stackoverflow.com/questions/4994018/how-to-replace-svns-sub-directory-vendor-branch-with-gits-subtree-merge12:55
well will read more pointers welcomed.12:56
→ orm_ joined12:56
orm_ heyo, more repo problems. For a while now, I have been asked to enter my password multiple times only to have git spit back at me that I failed to authenticate too many times. then after that initial round, i try again and it works. what is the deal?12:56
cebewee alfadir: Sorry, I have to leave now12:56
→ deuterium joined12:59
→ goshawk joined12:59
← hobodave left12:59
→ saysjonathan joined13:02
→ EnginA joined13:02
alfadir cebewee: thanks anyway13:03
found this, which at least formulates my thinking (used svn for quite some time) will read up on submodules13:03
http://stackoverflow.com/questions/677393/tracking-3rd-party-code-with-git13:03
→ bad_mantis joined13:04
← bad_mantis left13:04
← Engin left13:05
→ natek joined13:06
← saysjonathan left13:06
← teweWork left13:07
q0rban|afk → q0rban13:07
← hyperair left13:07
→ saysjonathan joined13:08
→ radarek joined13:08
→ teuf joined13:09
→ mattayers joined13:09
→ sternocera joined13:10
← freimatz left13:10
← nullvariable left13:10
→ nullvariable joined13:11
← saysjonathan left13:12
→ saysjonathan joined13:13
← MetaCosm left13:14
← a7ndrew left13:14
→ vitor-br joined13:14
→ MetaCosm joined13:14
← MetaCosm left13:14
→ MetaCosm joined13:14
→ Metathink joined13:15
Metathink hi13:15
→ MetaCosm_ joined13:15
← rchavik left13:17
← saysjonathan left13:17
← RobertMeta left13:17
← samhassell left13:18
→ holsta joined13:18
← dpino_ left13:18
→ saysjonathan joined13:18
← pasky left13:18
Metathink I'm trying to understand with gitolite, when and how the conf/gitolite.conf is analysed after a push to perform some actions like create new repos, updating permissions, etc. Any help please ?13:18
← tek0 left13:19
cbreak_work Metathink: do you know what a git hook is?13:19
Metathink yes, but i don't find where is the script who perform the parsing of gitolite.conf13:20
→ _Vi joined13:21
cbreak_work just look in the hooks of the admin repo on the server13:21
it should be called from there13:21
→ freimatz joined13:21
cbreak_work or wait for sitaram to be around13:21
→ gebi joined13:23
← gebi left13:23
→ gebi joined13:23
→ pasky joined13:23
← marcosiegl left13:23
← saysjonathan left13:23
→ mdel joined13:23
→ marcosiegl joined13:23
→ saysjonathan joined13:23
← marcosiegl left13:23
→ soc42 joined13:24
→ jfkw joined13:25
← saysjonathan left13:28
← rns left13:28
→ tek0 joined13:29
→ g0bl1n joined13:29
Metathink I think i found it, it's a perl script named gl-compile-conf13:29
→ girishr joined13:29
EugenA git branch -a13:30
i see only one line: remotes/origin/master13:30
what does it mean?13:30
← girishr left13:30
EugenA "remotes/origin/master" is the name of the branch?13:30
cbreak_work EugenA: it's very likely that that's a remote tracking branch13:31
so you have a remote called origin, and that has a master branch13:32
→ lol_h joined13:32
EugenA i've done git remote add origin ssh://.... and afte that git fetch13:33
← mdel left13:33
cbreak_work yes. sounds like it worked13:33
← RaceCondition left13:33
→ Eelis joined13:33
EugenA now i still don't see any files in my dir, i should do "git checkout", right?13:33
cbreak_work git checkout -t origin/master13:33
that will create a new branch, master, just for you13:34
which tracks origin's master13:34
→ RaceCondition joined13:34
EugenA why should i do -t origin/master and not -t remotes/origin/master ?13:34
← Eelis left13:35
→ Eelis joined13:35
Eelis when i do "git show HEAD" i get a complete diff of all the changes made by that commit. how do i get something resembling the output of "git status -s" instead, where i can just see the number of additions/removals per file?13:35
← bentob0x left13:35
→ saysjonathan joined13:35
wereHamster Eelis: git diff --stat HEAD^!13:35
← harshu___ left13:35
Eelis perfect! thanks :)13:35
wereHamster Eelis: or even show --stat13:35
← mattayers left13:35
Eelis ah, even better. thanks!13:35
← EricInBNE left13:36
→ mattayers joined13:36
← RaceCondition left13:36
cbx33 hey all13:36
cbreak_work: you got a sec for a pm?13:37
→ IslandUsurper joined13:37
← harshu left13:37
silex → Silex13:39
→ RaceCondition joined13:39
EugenA cbreak_work: how should i commit?13:39
← saysjonathan left13:39
sitaram Metathink: ping13:39
EugenA if i try git commit, i get info: Changed but not updated: myfile.txt13:40
wereHamster EugenA: looks like you haven't staged anything yet13:40
→ saysjonathan joined13:40
→ LouisJB joined13:41
EugenA wereHamster: what should i do? how do i stage? i'm new to git13:41
wereHamster EugenA: http://progit.org13:41
→ ben_alman_ joined13:42
← saysjonathan left13:44
EugenA wereHamster: should i do git add?13:45
→ flaguy48 joined13:45
→ saysjonathan joined13:46
← dyer left13:46
← ben_alman left13:46
→ teweWork joined13:47
cbreak_work EugenA: read a basic tutorial13:47
← ben_alman_ left13:47
→ xiangfu joined13:49
EugenA cbreak_work: i don't have basic config here...13:49
→ Ryback_ joined13:49
← Ryback_ left13:49
→ Ryback_ joined13:49
EugenA remote origin... and so on..13:50
← saysjonathan left13:51
← sh4rm4 left13:51
→ sh4rm4 joined13:51
→ saysjonathan joined13:51
→ Ruudjah joined13:52
← soc42 left13:52
Davey what happens when I merge two branches which have cherry-picked commits from one to the other already committed? conflicts?13:55
← saysjonathan left13:55
→ tstclair joined13:56
teuf Davey: git should be smart enough to detect they are the same patches13:56
→ saysjonathan joined13:56
Davey teuf: OK, so if I rebase squash the manual merges, and cherry-pick instead, that might be a way to fix up my (from svn) history?13:56
→ normanrichards joined13:57
→ smolyn joined13:57
wereHamster EugenA: man gittutorial13:57
jast EugenA: the 'gittutorial' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/gittutorial.html [automatic message]13:57
Silex → silex13:58
teuf Davey: no idea what you are trying to achieve :)13:58
→ Maletor joined13:58
← BobJelly left13:59
→ coderdad joined13:59
← saysjonathan left14:00
→ ch077179 joined14:00
→ saysjonathan joined14:01
silex → Silex14:02
→ BobJelly joined14:02
← GaVrA left14:04
→ globik joined14:04
alfadir cebewee: fyi, i think i have found my solution here (subtree merge) : http://support.github.com/discussions/repos/4733-merging-a-subdirectory-of-remote-branch-into-a-subdirectory-on-my-fork14:04
SethRobertson alfadir: Some workflows work better with gitslave than git-submodules or subtree merge. But those are your options14:05
cebewee alfadir: ah, fine14:05
alfadir: have you seen http://www.kernel.org/pub/software/scm/git/docs/howto/using-merge-subtree.html?14:06
← normanrichards left14:06
→ zivester_ joined14:06
← kreemer left14:06
← saysjonathan left14:06
→ emptyflask joined14:06
← emptyflask left14:06
alfadir yes, but http://progit.org/book/ch6-7.html was the first reference that really rung a bell :)14:06
→ emptyflask joined14:06
cebewee ok :)14:06
alfadir then the subdir trick i found via google :)14:06
SethRobertson: ok, have not yet read about gitslave.. will see how that fits my model14:07
SethRobertson http://gitslave.sf.net14:07
alfadir SethRobertson: thanks14:07
→ nonnooo joined14:08
SethRobertson Basically to support third parties (not the normal use case) you might have local masters for those third parties. But the purpose of gitslave is to keep all repos at the tip of their branches, and eliminates the tracking of what subrepo revisions you were at when a superrepo commit was made14:09
In exchange for losing that, it is much more intuitive to use.14:10
→ Pewpewarrows joined14:12
→ engla joined14:12
← engla left14:12
→ engla joined14:12
→ saysjonathan joined14:13
← vmiklos left14:14
← linusoleander left14:15
→ vmiklos joined14:15
→ lessthanzero joined14:15
lessthanzero I have a git repo and I'd like to "add" someone's fork. do I simply git remote add and pull forkedrepo master ?14:15
→ linusoleander joined14:16
← vmiklos left14:16
← ph^ left14:16
← LouisJB left14:17
← saysjonathan left14:17
→ saysjonathan joined14:18
← linusoleander left14:18
→ koltroll joined14:18
→ linusoleander joined14:19
→ vanmik joined14:19
engla lessthanzero: add the fork as a remote, use git fetch to update the remote tracking branches. use pull only if you want to merge14:19
→ agile joined14:19
→ vmiklos joined14:20
→ zombor_ joined14:20
→ dyer joined14:20
← dyer left14:20
→ dyer joined14:20
zombor_ → zombor14:21
→ normanrichards joined14:21
← tek0 left14:21
→ GaVrA joined14:21
→ tek0 joined14:22
← saysjonathan left14:22
→ saysjonathan joined14:23
← zorzar_ left14:23
→ sivy joined14:23
→ wysek joined14:23
→ mdel joined14:23
ChanServ set mode: +v14:25
wysek Hello, it seems like I do (git checkout some_branch; git merge master; git checkout master) a lot. Is there any other way to "push" changes to local branch some_branch?14:25
lessthanzero engla: ahh.. never really figured out the fetch vs. pull, but makes complete sense now. I'll read up on the fetch cmd and might ping back here if I have any problems - thanks!14:25
← GaVrA left14:26
→ cstephens4 joined14:27
→ jmil joined14:27
← saysjonathan left14:27
← cstephens4 left14:28
→ cstephens4 joined14:28
lteo is there a way for me to just view a file from another branch without checking it out?14:29
→ filipe_ joined14:30
bluenovember wyek, if that branch is a feature that will later be merged into mater, you should rebase it. (git checkout some_branch; git rebase master)14:31
its parent is then master, and will have any changes made on master as part of it14:32
engla lessthanzero: great :-)14:32
bluenovember oops, sorry, @wysek. similar names14:33
lessthanzero engla: quick question? I named my remote "johndoe" and although my "git fetch johndoe" and "git fetch johndoe master" both return success style messages (no errors) I don't think it's actually pulling code - am I missing something basic?14:34
→ saysjonathan joined14:35
lessthanzero maybe I'm just fetching the wrong branch14:35
→ jherdman joined14:35
→ Mannequin joined14:35
→ wmader joined14:35
ddilinger does git diff -w consider tabs to be whitespace? someone re-indented a file, i'm trying to merge my changes in, but git diff -w is claiming every line in the file changed so its hard to figure out what the merge is14:36
Davey teuf: perhaps this is a better phrasing for my question? given two timelines in git, that *did* branch from the same source; that I have /manually/ merged patches between (because it was svn), is there anyway to like, interactively merge and squash commits that were the same?14:36
lessthanzero nope, looks like I'm trying to fetch the right branch14:37
engla lessthanzero: git fetch johndoe will fetch all its branches. johndoe/master is not where you expected it to be?14:38
← linoj left14:38
teuf Davey: dunno :)14:39
lessthanzero engla: yea, I noticed all of the branches got pulled in. I bet it's setup right, I just need to checkout (or merge :/) the johndoe/master branch? (sorry, first times are always tense heh)14:39
teuf Davey: I generally use git rebase to "merge" git-svn branches14:39
→ booginga joined14:39
→ p3rror joined14:39
Davey teuf: that's... possible?14:39
← saysjonathan left14:39
lessthanzero engla: I'm looking at a github forking guide, which is designed around creating your own fork, but it fetches upstream repo, and then merges14:39
teuf Davey: for 2 git branches created from the same git-svn branch, and as long as you only rebase git commits, yep14:40
→ saysjonathan joined14:40
→ scelis joined14:41
engla lessthanzero: I can't say what you want to do of course, you can merge right away if wanted. or, how to checkout: git checkout -t -b doe-master johndoe/master (gave the checked-out branch a new name)14:41
→ Swimming_bird joined14:41
bluenovember teuf do you merge then rebase the result? (If not how does one use rebase with multiple branches?)14:42
← natek left14:42
Davey bluenovember: that's my question too :)14:43
ddilinger how can i get 'git diff' to display a patch without considereing the difference between tabs and spaces? i thought 'git diff -w' wsa suppose to but its now14:43
s/now/not/14:43
lessthanzero engla: ok, so I can checkout to keep johndoe's work separate from mine, or merge to bring his work into my local master repo - I think I'll merge. Although, I think I'll be pulling in a few different forks of this same codebase - does that affect my process? (should I checkout the specific files I want from each upstream repo rather then merging them all?)14:43
engla: p.s. thanks for this help - I know time doesn't come w/out a price :)14:43
← saysjonathan left14:44
← mcjb left14:45
engla lessthanzero: to reconcile branches you need to use merge, "checkout" is only used to take different viewpoints. I would merge after first reviewing the changes and seeing that they are good14:45
teuf bluenovember: no merge at all with git-svn, I'd be too afraid to break things ;)14:45
← normanrichards left14:46
teuf I only use git rebase when I want some poor man's merge with git svn14:46
→ bhosie joined14:46
ddilinger am i basically SOL if someone re-spaced a file?14:46
bluenovember ddilinger, perhaps some magic with auto-crlf14:47
wysek bluenovember: that branch is a placeholder for other repos, so that other repos have a place to push :), does your suggestion still apply then?14:47
bluenovember teuf, interesting. can you expand on that? How do you perform the merge?14:47
ddilinger bluenovember: whats auto-crlf?14:47
→ selckin joined14:47
→ panike joined14:48
← cstephens4 left14:48
teuf bluenovember: well, I don't merge at all, I just rebase stuff on top of other stuff14:48
bluenovember ddilinger, no I'm talking rubbish, that's line endings14:48
teuf, ah I see. tyvm14:48
Davey, ^14:48
wysek, yes, I think so. So you have one remote branch that everyone pushes to? like an integration/master?14:49
wysek bluenovember: yes14:49
bluenovember then you want to take in commits into branch you are working on that was branched from earlier in integration?14:49
yes, rebase is exactly what you want14:49
→ feindbild joined14:49
wysek bluenovember: actually, only I push there, but from another machine14:50
feindbild hi :)14:50
bluenovember that's a little confusing14:50
how does your local branch relate to this remote?14:50
→ coachz joined14:50
coachz hi14:50
→ Bass10 joined14:50
→ girishr joined14:51
→ sonink joined14:51
→ MacGyverNL joined14:51
← MacGyverNL left14:51
→ MacGyverNL joined14:51
wysek bluenovember: on localhost I develop on master, and I have branch win7 so that other machine can push there, and I can merge win7 <-> master14:51
← MacGyverNL left14:51
feindbild I'm trying to work out a cmd that lists all ingored files and directories, but 'git ls-files --others -i --exclude-standard' doesn't list ingored directories or files contained in them. any hint maybe? ^^14:52
coachz i checked out an earlier commit i'll call commit 1, my head is commit 2, if i commit changes from commit 1 (which becomes commit 3) will i still be able to get to commit 2 ?14:52
← csanz left14:52
wysek bluenovember: the only reason win7 branch exists is that from other machine I cannot push into master if master is checked out locally14:52
→ normanrichards joined14:54
← henrikbjorn left14:54
← tb_1 left14:54
← coachz left14:54
cbreak_work wysek: do NOT push to non-bare repos (unless you know what you're doing)14:54
you should use a bare repository for syncing14:54
← gebi left14:54
wysek yeah, but it will double my repository count14:55
so I just doubled branch count14:55
→ pgib joined14:56
→ psundaram joined14:56
← psundaram left14:56
← zivester_ left14:56
→ poseid joined14:57
ddilinger do i cancel a merge with git reset HEAD, or something diff?14:57
→ mando joined14:57
ddilinger basically i have a merge in progress from master to my branch, but its all wack and i want to cancel14:57
poseid hello... it has been a while... how can I make git ignore: # new file: UserIF/ATMMXLDigitalModule.rc14:57
→ nevyn^_ joined14:57
poseid command /*.rc does not work14:58
neither does .rc14:58
→ saysjonathan joined14:58
wysek ddilinger: if the merge succeeded in a commit, just git reset --hard HEAD~114:59
← EugenA left14:59
← p3rror left14:59
ddilinger wysek: it didn't even succeede, someone changed the spacing in files and its causing the merge to think entire files need to be replace14:59
→ tangledhelix joined14:59
ddilinger wysek: i'm thinking to undo the merge, apply the tab->space conversion that the other guy did, then try the merge over14:59
← tangledhelix left15:00
wysek ddilinger: so you can just reset --hard HEAD15:00
ddilinger wysek: ok, thanks!15:01
← nevyn^ left15:01
nevyn^_ → nevyn^15:01
wysek (I strongly suggest you read git-reset manual, though ;))15:01
← ayanes left15:01
cbreak_work poseid: you already git added it?15:01
git rm --cached it to stop tracking.15:01
→ alberto56 joined15:02
poseid hm.. and how do I exclude all .rc filetypes ?15:02
← saysjonathan left15:02
← freimatz left15:02
poseid app /*.rc and .rc do not have the wanted effect15:02
→ saysjonathan joined15:03
wysek *.rc15:03
ddilinger poseid: try it in .git/info/exclude15:03
poseid: i think .gitignore needs full path, exclude just takes a pattern(like .*.un~ for mine)15:04
→ rayners joined15:04
cbreak_work poseid: /*.rc ignores all files in the root folder of the project ending with .rc15:04
poseid: .rc ignores all files with the name .rc anywhere15:04
← winterweaver left15:05
cbreak_work you want *.rc, to ignore all files ending in .rc, anywhere in the project?15:05
poseid yes15:05
I have tried .rc first15:05
wysek try *.rc15:05
← mattayers left15:05
cbreak_work as I said, .rc isn't what you thought it'd be15:05
← OOPMan left15:06
→ mattayers joined15:06
poseid hmm... i was thinking that the * is added automagically, if I understand this right: http://www.kernel.org/pub/software/scm/git/docs/gitignore.html15:06
← angelsl left15:07
cbreak_work no15:07
why?15:07
← m_3 left15:07
← saysjonathan left15:07
poseid hmm "git treats the pattern as a shell glob suitable for consumption"15:08
→ psankar joined15:08
← psankar left15:08
→ psankar joined15:08
poseid pattern includes wildcards ?15:08
cbreak_work poseid: it does15:08
← mabrand left15:08
cbreak_work it's like when you type rm .rc15:08
← tannern left15:08
cbreak_work .rc is a shell glob pattern15:09
so rm deletes all files that match15:09
→ saysjonathan joined15:09
poseid ok.. so glob is asking for wildcards too... i was thinking it was regex style...15:09
cbreak_work it's shell glob pattern style :)15:09
poseid ok.. now .gitignore does its job perfect15:09
← kpreid left15:10
← mando left15:10
Davey teuf: we're not using git-svn, it's just a converted repo (so the commits happened in SVN, and therefore it wasn't a try cherry-pick, but a manual merge)15:11
→ ISF joined15:12
Davey teuf: I'm having trouble figuring out which commit I should be rebasing on... I'm in the branch I want to contain the history of both branches; I'm doing: git rebase -i <first commit on the other branch> and all I get in my editor, are the commits that happens in the current branch :/15:12
→ ceej joined15:12
cbreak_work Davey: git rebase otherbranchname15:13
→ beatak joined15:13
Davey cbreak_work: you're kidding right? heh15:13
← saysjonathan left15:13
→ m_3 joined15:13
← jfkw left15:13
→ Halerazematt joined15:14
→ p3rror joined15:14
→ saysjonathan joined15:14
← fr0sty|mobile left15:14
teuf Davey: ah, if you haven't used git cherry-pick, dunno how well git will behave at detecting the patches are the same15:14
Davey teuf: well, it breaks on merge (conflict)15:15
→ fr0sty|mobile joined15:15
Davey cbreak_work: am I rebasing <current> *onto* <otherbranchname> or the other way?15:15
teuf Davey: can't you remove the problematic commits from one of the branches and then merge ?15:16
Davey teuf: the problem is, most of them are tagged :/15:16
← Halerazematt_ left15:17
← girishr left15:18
← saysjonathan left15:19
→ saysjonathan joined15:19
Silex → silex15:19
← fr0sty|mobile left15:20
→ jnwhiteh joined15:21
→ Fullmoon joined15:21
→ DiGitalX joined15:22
→ fr0sty|mobile joined15:23
← ch077179 left15:23
→ jstemmer joined15:24
← saysjonathan left15:24
→ saysjonathan joined15:24
→ kpreid joined15:25
→ cannonball joined15:26
← cannonball left15:26
→ cannonball joined15:26
← tatsuya left15:26
rane_ wereHamster: thanks for the correction15:28
→ mrconnerton joined15:28
→ gebi joined15:28
← gebi left15:28
→ gebi joined15:28
sternocera How can I check out a revision of a file associated with a particular commit/sha1 hash?15:29
← saysjonathan left15:29
frerich sternocera: Using "git checkout <commit> <file>"15:30
mrconnerton How do I revert an entire repository back 10 or so pulls to a specific comit?15:30
→ saysjonathan joined15:30
sternocera frerich: Great, thanks15:30
→ tangledhelix joined15:30
← tangledhelix left15:31
rane_ mrconnerton: what kind of revert?15:32
frerich mrconnerton: It depends on what effect the 'revert' should have. If you want to revert the effect of the last 10 commits you can do 'git revert -n <commit> <commit>...'. If you want to have it look as if the last ten commits have never been done, you could use 'git rebase' or 'git reset'.15:33
rane_ you could reset to HEAD~1015:33
if that works for you15:33
→ obrienmd joined15:33
shruggar how can I detect whether a particular remote is defined, in a script? I want to loop through a list of repos, and "git remote update upstream" all those which have that remote defined15:33
mrconnerton frerich, I have a staging server and a prod server. I want to be able to (I believe) rebase the staging server to the same commit as the prod server15:33
→ rgr joined15:33
shruggar git remote | grep -q '^upstream$' feels clunky15:34
← saysjonathan left15:34
→ saysjonathan joined15:35
→ AzaToth_work joined15:36
→ ch077179 joined15:36
← Pewpewarrows left15:37
← linusoleander left15:37
← dc5ala left15:37
← c00kiemon5ter left15:37
→ c00kiemon5ter joined15:37
← c00kiemon5ter left15:37
→ c00kiemon5ter joined15:37
→ jmil_ joined15:38
→ hyperair joined15:38
→ linusoleander joined15:38
→ mando joined15:38
← mando left15:39
fr0sty shruggar: are you looking for a particular branch in the 'upstream'?15:39
→ soc42 joined15:39
shruggar fr0sty, the end goal is to send me an e-mail about any new tags15:40
← saysjonathan left15:40
← hunterloftis left15:40
shruggar (other than that, to make pushes of our own changes, which would be based on those tags, a bit faster)15:41
fr0sty new tags from whose perspective?15:41
→ saysjonathan joined15:41
shruggar as in: "here are the current tags" -> "git remote update" -> "here are now the current tags" -> email the difference between "current" and "now the current"15:42
failing anything else, I'll use "git for-each-ref" and "diff -u|grep '^+'"15:42
→ miki joined15:43
→ LemensTS joined15:43
LemensTS In GIT, how can I ignore a modified file so it is not commited at this time, but I can at a later time?15:44
fr0sty shruggar: sounds like a reasonable approach.15:44
← miki left15:44
shruggar LemensTS, just don't add it to the commit, and it won't be committed?15:44
teuf LemensTS: just don't add it to the index/don't use git commit -a15:44
SethRobertson LemensTS: You can just not `git add` it, or you can put it in .gitignore (and later remove it)15:44
fr0sty shruggar: btw, the output of git remote update should notify you of new tags15:44
LemensTS The file has already been added to git before15:44
SethRobertson git commit -a will not add a file not stored in git15:45
← EnginA left15:45
→ fission6 joined15:45
shruggar fr0sty, yeah, but I'd prefer not to get that in my e-mail unless there are new tags :)15:45
SethRobertson LemensTS: Ah, then you need to manually add files to commit. Also, see http://thomasrast.ch/git/local-config.html15:45
fission6 is there a way to view the "updates" on the upstream version before pulling or better yet check the log to see what changes are there15:45
shruggar (and grepping porcelain doesn't sound fun to me)15:46
→ tcv_br joined15:46
← tcv_br left15:46
shruggar fission6, what are you actually trying to do?15:46
← saysjonathan left15:46
SethRobertson fission6: Yes, very easy. git fetch and then `git log` or `git diff` to see the changes. I have a progam to automate this15:46
fission6 review the changes upstream, not nessarily the code but perhaps the log15:46
SethRobertson (Called gitin, and then the corresponding gitout)15:46
→ ben_alman joined15:47
→ saysjonathan joined15:47
← _Vi left15:47
shruggar fission6, but why? to save bandwidth, or because you don't want to merge and don't know about "git remote update" ?15:47
← LemensTS left15:47
SethRobertson fission6: http://gitslave.git.sourceforge.net/git/gitweb.cgi?p=gitslave/gitslave;a=blob_plain;f=contrib/gitin;hb=HEAD15:48
← opalka left15:48
→ h32Lg joined15:48
← sonink left15:48
← h32Lg left15:48
→ h32Lg joined15:48
→ PigeonFriend joined15:49
← nd__ left15:49
→ leavittx_ joined15:49
← xiong left15:49
→ Tobias__ joined15:50
→ nd__ joined15:50
← saysjonathan left15:51
→ saysjonathan joined15:52
→ hasrb joined15:52
→ Yuffster joined15:52
← smolyn left15:53
← NickyP left15:53
ChanServ set mode: +v15:54
→ zerep joined15:54
→ daviddoria joined15:54
Tobias__ I have a strange problem: I have a git repository (".../git-repro-dir.git") on a webserver (w/ bare = true). However, if one clones the repository only the original commit is show. If one goes into the directory on the webserver, "git log" shows all - also gitweb which accesses the directory shows all commits. Any idea?15:54
← ormaaj left15:55
← zerep left15:55
fr0sty Tobias__: does 'git log --graph --all --oneline --color --decorate' on a clone show more than one commit?15:55
→ ormaaj joined15:55
← ormaaj left15:56
← daviddoria left15:56
Tobias__ fr0sty: No15:56
← saysjonathan left15:56
mrconnerton I have a staging server and a prod server which both have a clone of my repository. What is the best way to reset my staging repository to match the prod repository? As the staging repository might be ahead of the prod15:57
SethRobertson Tobias__: `git branch` says15:57
← Araxia_ left15:57
SethRobertson And `git status`15:57
← leavittx_ left15:57
→ ph^ joined15:58
Tobias__ git branch -a on the server says "* master" and on the clone "* master" / "remotes/origin/HEAD -> origin/master" / "remotes/origin/master"15:58
← nevyn^ left15:58
→ nevyn^ joined15:58
→ emptyflask_work joined15:58
fr0sty Tobias__: does 'git remote update' on the clone change anything?15:59
Tobias__ Regarding status: client "nothing to commit (working directory clean)" / server: "fatal: This operation must be run in a work tree"15:59
cbreak_work mrconnerton: bare or non-bare?16:00
Tobias__ Regarding 'git remote update': That just says "Fetching origin"16:00
leo2007 is it normal for git log to take longer in print 100 oneline logs in a big repo than a small one?16:00
→ jfkw joined16:00
mrconnerton cbreak_work, I honestly don't know what that means16:00
→ refund joined16:00
→ mattly joined16:01
refund I merged two local branches and did git pull. How can I revert back to how it was before the merge? git reset --hard did not undo the merge16:01
fission6 so your suggest git fetch, git log to see the latest logs upstream?16:01
fr0sty Tobias__: what command did you use to clone?16:02
→ kukks joined16:02
fr0sty fission6: yes.16:02
cbreak_work mrconnerton: are they repositories you work on, or repositories that don't have any files themseles and you push to?16:02
fission6 what if i dont want the changes?16:02
← bhosie left16:03
mrconnerton cbreak_work, (I come from an svn background) they aren't edited directly, they exist to pull update in. prod = www.mysite.com and staging = staging.mysite.com16:03
→ saysjonathan joined16:03
← skoop left16:04
fission6 git fetch origin master did no show anything new in git log16:04
git pull origin master did16:04
whats up with that16:05
Tobias__ fr0sty: git clone --progress http://.../git-repro-dir.git16:05
→ j416 joined16:05
mrconnerton cbreak_work, We develop locally push/pull to a DEV server, when things look good we pull to STAGING for client to approve, when things look good we pull to PROD for live. When we pull to staging though, we want to reset the site to match prod first before we pull in.16:05
← BobJelly left16:06
→ bhosie joined16:06
→ emdete joined16:06
→ v0n joined16:07
emdete hi, wasnt there a shortcut to checkout the previous state after a checkout? i left a 'detached HEAD' without noting the sha and need to get back there...16:07
← v0n left16:07
→ tjholowaychuk joined16:07
← saysjonathan left16:08
← tek0 left16:08
fr0sty fission6: because 'git fetch origin master' puts the changes in FETCH_HEAD, and you would need 'git log FETCH_HEAD' to see them.16:08
→ v0n joined16:08
fr0sty "git pull origin master" fetched the changes and then merged them into your current branch.16:08
shruggar fr0sty, git fetch --tags appears to produce the desired output16:09
← jhoepken left16:09
→ saysjonathan joined16:09
← petrux left16:09
→ EnginA joined16:10
→ nadavoid joined16:10
← mattayers left16:10
→ cwebber joined16:10
→ mattayers joined16:10
← mattayers left16:10
→ BobJelly joined16:10
← warthog9 left16:11
← jnwhiteh left16:11
→ jrmuizel joined16:11
shruggar GIT_DIR="$repo" { git remote | grep -q '^upstream$' && git fetch --quiet --no-tags upstream && git fetch --tags upstream; }16:11
fission6 thanks fr0sty16:12
fr0sty Tobias__: and to be clear: the repository you cloned lists one commit with 'git log --all', right?16:12
shruggar grep -q -x upstream, a bit nicer16:13
→ Araxia_ joined16:13
fr0sty fission6: np16:13
← saysjonathan left16:13
→ leavittx_ joined16:13
Tobias__ Yes. What I just saw is (from the logs) that "git clone http://" reads "info/refs" which contains "refs/heads/master" which points to the first commit; a file-system based "git clone /web/server" works.16:14
← radarek left16:14
→ _psychic_ joined16:14
→ saysjonathan joined16:14
← saschpe left16:15
← j416 left16:15
← cbreak_work left16:15
fr0sty Tobias__: there is an update-server-info script that you need to run: http://www.kernel.org/pub/software/scm/git/docs/git-update-server-info.html16:15
→ j416 joined16:16
← Halerazematt left16:17
← hasrb left16:18
→ LouisJB joined16:18
→ smolyn joined16:18
→ pheaver joined16:19
← gebi left16:19
Tobias__ fr0sty: Thanks!16:19
← saysjonathan left16:19
→ gebi joined16:19
← gebi left16:19
→ gebi joined16:19
→ saysjonathan joined16:20
← pheaver left16:20
← pgib left16:20
← EnginA left16:21
→ Halerazematt joined16:22
→ alvaro_o joined16:23
← saysjonathan left16:24
← emptyflask_work left16:24
← Ruudjah left16:24
← pulse00 left16:24
→ pandur joined16:25
→ saysjonathan joined16:25
pandur hello channel16:25
→ Silowyi joined16:26
mrconnerton If there any git guru freelancers that can consult with me on something right now, please pm me and let me know what 30 minutes of your time costs16:27
NfNitLoop mrconnerton: or you could just ask your question in the channel for free.16:28
→ thoolihan joined16:28
← friskd left16:28
mrconnerton NfNitLoop, I have asked but haven't got help. Usually I'm a very patient person by my sprint ends today and I need to get it resolved16:28
← saysjonathan left16:29
→ pulse00 joined16:29
← thoolihan left16:29
pandur mrconnerton: ask again :)16:29
← Araxia_ left16:30
← jrmuizel left16:30
→ saysjonathan joined16:30
← gebi left16:30
mrconnerton pandur, It would be the third time within an hour. Thats spamming in the channels I come from :-)16:31
Are people not motivated by money for some quick consulting here?16:31
NfNitLoop the overhead of getting paid for 30 minutes is not worth what you would likely pay for such information.16:32
→ ISF_ joined16:32
NfNitLoop anyway, people are in here to give free help.16:32
I see your question in scrollback.16:32
You want to pull and check out a particular version of your repository to a staging server.16:32
mrconnerton NfNitLoop, but people haven't told me their rates. I make $100 for the work I do, I just assume that you guys are in the same boat16:32
NfNitLoop that's a very trivial operation. "git fetch, git checkout commitID" possibly with an -f flag.16:33
mrconnerton I want to reset my staging server to match my prod server so I can then git pull to test pushing live16:33
NfNitLoop same operation.16:33
git fetch, git checkout productionCommitID.16:33
Davey teuf: cbreak : what am I *seeing* in vim when I git rebase <other branch> ?16:34
← bhosie left16:34
Davey the commits that conflict?16:34
→ bhosie joined16:34
← ISF left16:34
← j416 left16:34
fr0sty Davey: 'git rebase -i', you mean? those are the commits you are rebasing16:34
← Halerazematt left16:35
→ sonink joined16:35
← saysjonathan left16:35
← ChUrP left16:35
Davey fr0sty: er, yes, and I'm confused then :/16:35
→ _Vi joined16:35
NfNitLoop mrconnerton: you can even simplify the process by using tags. like production-release-yy-mm-dd or whatever.16:35
→ Anubisss joined16:35
→ Schmallon joined16:36
→ saysjonathan joined16:36
← Tobias__ left16:36
NfNitLoop mrconnerton: but reverting to a particular version and checking out a particular version should be covered in any git tutorial. That's what SCMs are for.16:36
→ pgib joined16:36
Davey fr0sty: I'm in branch "hera" and I'm git rebase -i "hera-working"; what I see in vim, are the commits in "hera"... this isn't what I expect16:36
← poseid left16:37
NfNitLoop mrconnerton: I don't want to get paid, and please ask questions in here so that others can benefit from our conversation.16:37
mrconnerton NfNitLoop, tutorials are great and all but getting one on one attention for a bit is far more beneficial.16:37
fr0sty mrconnerton: and if you want to just 'match' the other repository's state you shouldn't be using 'pull'...16:37
teuf Davey: I recommend a careful reading of man git-rebase, it's all explained in there16:37
jast Davey: the 'git-rebase' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-rebase.html [automatic message]16:37
pandur my git status hangs. strace reveals that it is trying to read from STDIN. GIT_TRACE gives no details. any ideas?16:38
mrconnerton Ok. I need to hire a git consultant for some personal attention that is more indepth then general tutorials and specific to my project. If anyone has some availability today, please let me know.16:39
→ rajeshsr joined16:40
← saysjonathan left16:40
→ saysjonathan joined16:41
← thierryp left16:41
→ MattDiPasquale joined16:42
NfNitLoop scratches his head.16:43
← goshawk left16:43
NfNitLoop pandur: what command were you running?16:44
pandur NfNitLoop: "git status"16:44
NfNitLoop just 'git status'? hmmm...16:44
← mkramer left16:44
→ robotarmy joined16:44
← tewecske left16:44
NfNitLoop pandur: what version of git?16:44
→ hasrb joined16:44
fr0sty you have N-on-one attention currently, the solution to your problem is 'git fetch' followed by either 'git checkout' or 'git reset'.16:44
pandur NfNitLoop: it all began yesterday while committing. the commit wnt through, but I want to rebase -i it. it hangs on step 2/4 and when I ^C I get in this situation16:45
→ cylence joined16:45
pandur NfNitLoop: 1.7.116:45
← saysjonathan left16:45
→ saysjonathan joined16:46
NfNitLoop fr0sty: I assume that was for mrconnerton. *shrug* If he would rather wait and pay someone than get free help in here, not much we can do about it. :)16:46
→ friskd joined16:47
← mattly left16:48
→ ShooterMG joined16:48
→ TeckniX joined16:48
← jmil_ left16:48
NfNitLoop pandur: have you tried resetting and doing a 'git fsck'?16:49
pandur NfNitLoop: status -u no works O.o16:49
NfNitLoop: fsck --full several times, will try reset16:49
→ Heimidal joined16:49
mrconnerton fr0sty, NfNitLoop, the help in here is great and free and I really do appreciate it. and will attempt those commands in the mean time. But I prefer to pay people for consultation that is unique and special to me. I do what I do best and pay other people to do what they do best.16:50
← Behi left16:50
← saysjonathan left16:51
← serard left16:51
→ rgr_ joined16:51
→ saysjonathan joined16:51
← rgr left16:52
sonink Good morning all. I'm new to git and have read quite a bit of the manual and book.git-scm.com tutorials but am still unsure how pushes work for syncing repositories. Does the bare repository sync all of the other repositories solely from its history in the .git directory? If remote repositories pull info from .git directories, why are bare repositories better / why not sync on a main repository that does have the working files? Just looking for clarity in ho16:52
push/pull works and when to use it.16:52
shruggar is there something like a "git fsck --actual-problems-only-please", ie: something suitable for running via cron? (ignoring "dangling" objects, for example)16:53
pandur NfNitLoop: no difference. I restarted the rebase with GIT_TRACE=1 and the last line is "[detached HEAD d06bc78] <my new ci msg"16:53
→ khaije|minder joined16:53
← priidu left16:53
→ nthompson joined16:53
shruggar sonink, "sync" is not a git term, afaik16:53
→ serard joined16:53
← pgib left16:53
→ dekebalus joined16:54
khaije|minder i was thinking about adding an alias for "git add --all && git commit -am 'update' && git pull && git push" to my config. any recommendations or suggestions on the alias name? the best i could think of so far is alias.syncup16:54
NfNitLoop sonink: Because if you do a merge, you *require* a working copy in which to do the work. It's best to always do a amerge on your machine. Your server shouldn't ever do merges, so it doesn't need a working copy, hence bare repositories.16:54
sonink shruggar: well I'm thinking of the bare repository as the stable, public repository. Does that make sense?16:55
→ mando joined16:55
ChanServ set mode: +v16:55
NfNitLoop sonink: that is a good idea. :)16:55
→ offby1 joined16:55
nthompson Hello, I was hoping I could ask somebody about a git workflow I'm trying to implement and see what others thought about it.16:55
← rgr_ left16:55
→ brntbeer joined16:56
nthompson We use an open source software package that is hosted by Github, and we want to host our own cloned version of that to make modifications. When the source does a new release, we want to merge that into our own hosted version.16:56
shruggar sonink, the bare repository is the one whose only purpose is to hold changes for other people. "stable" would be a bad thing to pretend it represents, as public shared developer branches are not uncommon16:56
← saysjonathan left16:56
← pulse00 left16:57
nthompson I was thinking of using master to be a remote tracking branch of the main repo, and setup DEV, TEST, STAGE, and PROD branches of master, that is where we'd do coding.16:57
NfNitLoop nthompson: That will be very simple in git. Create a bare 'clone' of the upstream. Make your own branches in there to do work.16:57
shruggar sonink, "shared" meaning "accessible" in this case, not necessarily "used by multiple people"16:57
← alvaro_o left16:57
← brntbeer left16:57
→ alvaro_o joined16:57
→ LiamH joined16:58
← feindbild left16:58
sonink shruggar, NfNitLoop: Maybe if I just explain the environment in which I want to implement git you can advise me better :-)16:58
nthompson NfNitLoop, so would the bare clone of the upstream be master?16:59
→ t_b joined16:59
← yhager_ left17:01
→ sftp joined17:01
→ yhager_ joined17:01
pandur NfNitLoop: got it. I accidently added an empty file called '-' *cough*17:01
← alvaro_o left17:02
→ pulse00 joined17:02
→ alester joined17:02
→ alvaro_o joined17:02
→ slonopotamus joined17:03
← iravid left17:03
→ iravid joined17:03
shruggar better than '-rf /' :(17:03
*:)17:03
→ justin-george joined17:03
→ saysjonathan joined17:03
pandur shruggar: hrr17:03
← nonnooo left17:03
pandur should I file a bug for this or did I just used all the rope that was given to me?17:04
sonink I have a test web server that myself and another programmer are working on (will be more programmers in the future, one of many reasons I want to get git setup). There are multiple projects on it that I want to setup as separate git repositories. We want to be able to work on separate changes on the server and see them live as it will serve the pages.17:04
I'm thinking have one copy of the projects left as the stable versions and then we clone whatever we want to work on into a separate test directory so we can see our changes live on the server. Does this make sense? Will we be able to just push commits from our clones in separate directories back to the stable repos?17:05
→ noahcampbell joined17:05
→ necromancer joined17:06
← koltroll left17:07
← rajeshsr left17:07
← saysjonathan left17:08
← Badiss left17:08
→ steph021 joined17:08
← steph021 left17:08
→ steph021 joined17:08
← davzie left17:09
→ pgib joined17:09
→ Araxia_ joined17:09
← slonopotamus left17:11
← alvaro_o left17:11
→ project2501a joined17:11
→ alvaro_o joined17:11
→ radarek joined17:12
→ Buglouse joined17:12
← psoo left17:13
→ pheaver joined17:14
← pheaver left17:14
← avinashhm left17:14
→ pheaver joined17:14
← steph021 left17:15
→ hobodave joined17:15
→ gebi joined17:16
← gebi left17:16
→ gebi joined17:16
← pfrenssen left17:16
→ aah_ joined17:17
fr0sty sonink: sounds reasonable to me.17:18
khaije|minder likes git, it's not simple, but it's not more complicated than it appears17:20
← Fullmoon left17:20
← cilly left17:20
→ mattly joined17:20
← soc42 left17:20
→ __name__ joined17:21
← __name__ left17:21
→ __name__ joined17:21
Nugget <nDuff> git is the biggest, most beautiful footgun around!17:21
← leavittx left17:21
→ cilly joined17:22
← p3rror left17:22
sonink fr0sty: Do I need to setup a bare repository in this case because of the merge problem that NfNitLoop described earlier? If so, how do I keep a clean copy of the files separate from the test directories to which we will copy while modifying code?17:23
← jmil left17:23
← h0st1le_ left17:23
← kpreid left17:24
→ h0st1le joined17:24
← multitallented left17:24
→ gusnan joined17:24
← alvaro_o left17:25
← joel_falcou left17:25
fr0sty if you need to coordinate work between multiple devs having a bare repository to push/pull is helpful. what do you mean 'clean copy'?17:25
→ alvaro_o joined17:26
← khaije|minder left17:26
← pgib left17:26
→ artefon joined17:26
→ steph021 joined17:28
← steph021 left17:28
→ steph021 joined17:28
← hasrb left17:28
→ amerine joined17:28
← teweWork left17:28
→ triptec joined17:28
← leavittx_ left17:30
← PigeonFriend left17:30
sonink Just copies of the project that update to and stay at each commit, not reflecting modifications to the code in our test directories. Copies that we can show to manager/co-workers/clients for feedback. Would I just need to clone into a different non-test folder from the bare repository?17:30
→ henrikbjorn joined17:31
→ maattd joined17:31
→ cirwin joined17:31
← shruggar left17:31
fr0sty that would work. you would need to either update the manually or write some hooks to keep them up to date.17:32
→ saysjonathan joined17:33
← artefon left17:33
sonink Okay. Do you know of any tutorials that are similar to lead me in the right direction on setting up hooks scripts to automate that?17:33
← bhosie left17:34
← mando left17:34
← saysjonathan left17:34
← refund left17:34
← mrconnerton left17:34
→ bhosie joined17:34
→ saysjonathan joined17:35
→ p3rror joined17:35
DrNick2 → DrNick17:35
→ priidu joined17:36
→ leavittx joined17:36
← Schmallon left17:36
fr0sty http://toroid.org/ams/git-website-howto17:37
sonink: that page ^ has been recommended in the past.17:38
→ rajeshsr joined17:38
← hyperair left17:41
← pulse00 left17:41
← deuterium left17:42
→ hyperair joined17:42
← Determinist left17:44
sonink fr0sty: Thanks. I'll give that a look and let you know if I have any questions. I think I had read some pages about setting up hooks while researching git but I don't remember where they were at the moment.17:46
← mattly left17:46
→ artefon joined17:46
← henrikbjorn left17:47
← rajeshsr left17:48
← alvaro_o left17:48
→ mattly joined17:48
→ alvaro_o joined17:48
← urkud_ left17:48
← linusoleander left17:49
adamholt → adamholt_away17:49
→ ChUrP joined17:49
adamholt_away → adamholt17:50
→ urkud_ joined17:50
← mutex left17:50
← LouisJB left17:50
→ Determinist joined17:51
← alvaro_o left17:51
← Buglouse left17:52
→ alvaro_o joined17:52
→ Utkarsh_ joined17:52
→ kakos joined17:52
→ _ollie joined17:53
← bronson left17:53
← tjholowaychuk left17:53
← Utkarsh left17:55
→ hornswoggles joined17:55
← alvaro_o left17:56
→ alvaro_o joined17:56
← Yuuhi left17:56
→ Yuuhi joined17:56
→ eft_ joined17:57
← darkredandyellow left17:57
← bhosie left17:58
← alvaro_o left17:59
→ alvaro_o joined17:59
→ bhosie joined17:59
cbx33 → Guest9498418:00
→ cbx33 joined18:01
cbx33 hey peeps18:01
← hyperair left18:01
→ shyam_iyer joined18:02
cirwin hi18:03
← serard left18:03
Davey How can I compare two branches? (i.e. diff -ur <branch> <otherbranch>)18:03
cbreak Davey: just git diff brancha branchb18:03
Davey cbreak: hrm... then my merge didn't work (despite every commit that I manually check the patch for being there)18:04
← marcob left18:04
cbreak just look at the history with git log --graph --decorate18:05
← alvaro_o left18:05
→ alvaro_o joined18:05
← urkud_ left18:06
Davey cbreak: that looks fine18:06
→ linusoleander joined18:06
Davey but git diff still shows (huge) differences between them18:06
cbreak between what?18:06
Davey the two branches that I merged18:06
cbreak of course18:06
that's normal18:07
a merge doesn't make your branch like the one you merge18:07
it just makes the changes that branch did be in your branch18:07
the merge commit will be different than both the parents18:07
Davey oh, I get /that/18:07
but it shouldn't be in this case18:08
lemme splain18:08
two branches from the same common ancestor, the *only* changes to one of them, was cherry-pick'ed stuff from the other one18:08
therefore git merge should make them identical18:08
← alvaro_o left18:10
→ Raging_Hog joined18:10
→ mezgani_ joined18:10
→ alvaro_o joined18:10
cbreak no18:10
a merge does not change the other branch18:10
only yours18:10
→ Zephyrus joined18:10
fr0sty Davey: the only changes are what was done in A that wasn't cherry-picked to B18:10
Davey right...18:11
← vsayer left18:11
Davey draws18:11
cbreak just make a screenshot of your git log --graph --decorate --all18:11
or use gitx/gitk18:11
← p3rror left18:11
→ vsayer joined18:11
→ tewecske joined18:12
fr0sty + --oneline18:12
fr0sty doesn't need to read the full commit messages :-)18:13
Raging_Hog How do you conveniently review incoming changes, edit them, possibly leave some of them out and resolve conflicts? I've been testing out rebase and then difftool against ORIG_HEAD. Another option is merge with no-commit and no-ff and then diff --staged. Are these methods of evil?18:14
← _ollie left18:14
← BobJelly left18:14
cbreak why not just merge them?18:15
and afterwards read the log?18:15
Davey http://screencast.com/t/6Hjy6H8NA <--- cbreak, fr0sty — this is what I did. the dashed lines are cherry picks. I'm merging B, into A. Shouldn't the result files on disk, be identical?18:15
← linusoleander left18:15
cbreak cherry picks from where to where?18:16
Davey sorry, from B -> A18:16
cbreak you are on A?18:16
cirwin what did you use to draw that, out of interest?18:16
Davey google docs18:17
cbreak: when I did the merge? yes18:17
→ Wooga joined18:17
cbreak Davey: I mean while merging and cherry picking18:17
Davey yes18:17
Wooga hello, is email is only way to submit pattch for git priject?18:17
cbreak and there's no commit in A that is not also in B?18:17
← sudonotplease left18:17
Davey cbreak: correct18:17
cbreak then the result should be that A is like B18:17
Wooga i mean, isn't there any specific git mechanism for sending patches?18:17
Davey (actually, incorrect, but I'm seeing more in my diff than just those changes)18:18
cbreak Wooga: git send-email18:18
Raging_Hog cbreak: I'd like to see the actual changes in the code. I've been reasoning that if I do no-commit no-ff merge, I can more easily diff the changes, otherwise I need to look up my own last commit from the log and diff against that.18:18
→ sebersole joined18:18
Davey cbreak: my git diff A B, shows all the differences between A and B as they were before the merge, near as I can tell18:18
fr0sty Davey: you did 'git checkout A; git merge B'?18:18
Wooga cbreak: so, email is only way?18:18
cbreak Raging_Hog: not really18:18
Raging_Hog: you can just git show somecommit18:18
→ BobJelly joined18:18
Davey fr0sty: I did: git checkout A; git merge -s recursive -X ours B18:18
cirwin Wooga: you can do whatever you like — email is what the git project itself uses18:19
if you're on github, they have a much better mechanism of pull requests18:19
→ jrmuizel_ joined18:19
bremner Wooga: do you mean _a_ git project or _the_ git project?18:19
fr0sty Davey: why --ours?18:19
cbreak Wooga: no, normal people just push to some repo and let others pull18:19
bremner well, sometimes email is used for review, even by normal people18:19
Davey fr0sty: they weren't cherry-picks, manual patch/commit, so git conflicted on them18:19
→ chalkbag joined18:19
Davey which, as I understand the recursive strategy, would pick "ours" over 'theirs", whenever a conflict occurred18:19
cirwin bremner: they're not normal people if they think that diff-over-email is a good thing :p18:20
Wooga bremner: sorry, can't tell the difference due to weak knowledge of english :D18:20
Davey fr0sty: I'm trying to simplify the problem because I know my repo is a mess from our svn conversion18:20
bremner Wooga: what project do you want to submit patches for?18:20
Raging_Hog cbreak: I'd like to be able to use difftool with the diff18:20
cbreak Davey: there should not be any conflicts18:20
fr0sty Davey: do you prefer the commits in A over B (are the manual picks 'better'?)18:21
→ TheFox21 joined18:21
← steph021 left18:21
cbreak Raging_Hog: git difftool somecommit~1 somecommit?18:21
Davey fr0sty: I don't understand the question; the changes are the same, hence the conflict, the reason I prefer A, is because there are tags on each of those commits (it's a release branch, with hotfixes being applied for releasing)18:21
→ steph021 joined18:21
← steph021 left18:22
→ steph021 joined18:22
Wooga bremner: well, now i am understanding that it is worth to note that i am using github and want to test it's patch sending features. but apperently, this is not a git question, but a github one, so i'd better to read some docs on github. thanks for helping!18:22
bremner Wooga: there is also #github18:22
cirwin http://help.github.com/pull-requests/18:22
Wooga oh, thanks!18:22
cbreak Davey: a merge doesn't change tags, no matter how you do it18:22
since a merge does not change history18:22
Davey cbreak: right, but the only way I figured out how to fix the conflict, was to rebase and remove the conflicting commits...18:23
I *suppose* I could /manually/ resolve the conflicts :P18:23
Fuu` → Fuu18:23
← justin-george left18:23
→ linusoleander joined18:23
← marc_kdab_ left18:23
fr0sty Davey: if you want the result of the merge to look like B use --theirs, not ours.18:24
Davey fr0sty: I /could/ do that, now I know it wont affect the tags18:24
fr0sty thinks the picked commits did not apply cleanly and the resolutions conflicted with subsequent commits on the other branch.18:25
Davey they weren't cherry-pick'ed18:25
there were manual diff | patch; commit18:25
(because there were done in subversion, before conversion to git)18:25
→ n3m joined18:25
← ptl left18:26
fr0sty doesn't matter how they were done, just whether they applied cleanly...18:27
← rayners left18:27
fr0sty if the patch results in a sort-of-similar-but-not-identical change the merge will conflict.18:28
Davey fr0sty: I see, git uses the /actual/ text changes to compare? I figured it was meta-data18:28
fr0sty but this discussion is academic if you are going to use --theirs.18:28
← wmader left18:28
← globik left18:28
→ LouisJB joined18:28
→ ptl joined18:28
→ rayners joined18:28
← ptl left18:28
→ ptl joined18:28
fr0sty *chuckles* ... meta-data18:28
Davey oh hush; I'm an svn luddite :(18:29
fr0sty in git there is only content.18:29
← ningia left18:29
→ nona joined18:29
→ captproton joined18:29
fr0sty metadata is for those who can't post-process, or who throw away too much information.18:31
← emptyflask left18:31
→ mikepack joined18:32
← werdan7 left18:32
← LouisJB left18:33
→ joel_falcou joined18:33
← webbananaboy1 left18:33
cbreak fr0sty: commit objects are almost purely metadata18:33
← leavittx left18:33
cbreak and trees are also filled with rather a lot of meta data18:33
fr0sty quit being so correct...18:34
cbreak like file names... and executability bits...18:34
fr0sty ;-)18:34
← guardian left18:34
fr0sty and rename support...18:34
cbreak rename support?18:35
→ emptyflask joined18:35
← MattDiPasquale left18:35
fr0sty just kidding...18:35
→ kpreid joined18:35
← captproton left18:35
fr0sty it was that sort of metadata I was thinking of.18:35
cbreak that's like saying "this text file supports changing words"18:36
→ msierks joined18:36
→ dreiss joined18:36
fr0sty not if you are using CVS...18:37
msierks I am using git over http and I have to put my password in a .netrc file in plain text, is there some way to use a hashed password in that file or avoid this ?18:37
fr0sty explicit rename tracking is a feature in some VCS18:37
and they tend to use metadata to support it, and you have to do it right at the time or it's forever broken.18:38
cbreak msierks: can you send the hashed password to the server?18:38
← agile left18:38
→ kumbayo joined18:38
msierks I am not sure, I don't think so if it is http18:39
→ Nayruden joined18:39
→ agile joined18:40
→ ashc joined18:40
← jefferai left18:40
← Teo` left18:40
→ jefferai joined18:41
→ Teo` joined18:41
← Teo` left18:41
→ Teo` joined18:41
PerlJam msierks: then ... don't use git over http? :)18:41
cbreak if you use https, you might be able to set up something with real public key authentication18:41
and client certs18:41
SethRobertson Does curl support that? I don't think so.18:42
cbreak I wouldn't know18:42
but https does18:42
SethRobertson git uses curl18:42
cirwin curl supports client certificates18:42
→ kenichi joined18:42
SethRobertson Then that is how you would have to do it. Good luck, it is painful.18:43
→ MikhX joined18:43
← cirwin left18:43
msierks well i was attempting to use ssh with rsa keys but was unsure of a way to add authorized_keys using php18:43
ashc any suggestions on software or scripts that can generate daily changelogs for code reviews? I looked at http://codemonkeytips.blogspot.com/2010/01/daily-commit-logs-batching-git-commit.html but the output is horrible.18:43
cbreak it's easy to do with gitolite over ssh18:43
so never forget: there is salvation18:44
→ mando joined18:44
cbreak msierks: can you do perl?18:44
← priidu left18:44
msierks hmm I will have to look at gitolites code, I am trying to integrate git with my database18:44
SethRobertson ashc: Isn't that what gerrit is supposed to do for you? I've never used it18:44
msierks cbreak, yes18:44
cbreak gitolite is perl, and it does authorized-key management18:45
ashc SethRobertson: never heard of it... i will check it out though18:45
→ Schmallon joined18:45
msierks cbreak, okay I will look into that thanks18:45
what are the chances of the git protocol getting authentication ?18:45
→ slonopotamus joined18:45
→ jmil joined18:45
rane_ isn't ssh there for that…18:46
SethRobertson ashc: Another way is to have a review branch or review repo which people pull from. Some people use git-notes for code-review, but you have to arrange for those notes to be pushed and pulled.18:46
msierks rane_, I think it is, but it would be easier if git:// already had it18:47
cbreak msierks: I don't think there's a chance18:47
not like it's useful18:47
← nona left18:47
cbreak because it'd either be too much work (rewriting ssh), orinsecure18:47
rane_ basically reinventing the wheel, and for what18:48
msierks hmmm yeah I get your point okay thanks18:48
→ Cvbge joined18:48
→ nona joined18:49
← tvw left18:49
Cvbge Hi18:50
← ph^ left18:51
→ ph^ joined18:51
ashc SethRobertson: that seems to *heavy* and people would be less likely to do it...18:52
looking at a Redmine plugin... maybe that will work18:52
SethRobertson ashc: Another option is for users to push to a pre-review repo and have a review process /approver be the way to get from that repo to the master repo18:53
cbreak or branch18:53
SethRobertson But if you have a workflow in mind, you should say what it is18:53
cbreak or personal branches18:53
with integrator workflow18:53
→ LouisJB joined18:54
← nevyn^ left18:55
← ISF_ left18:55
→ WebDragon joined18:55
← ptl left18:56
← kpreid left18:57
→ bronson joined18:59
→ MattDiPasquale joined18:59
→ mrconnerton joined19:00
→ bitkiller joined19:00
← psankar left19:00
fission6 how do i set up a gig project where one folder inside the project is another gig repo19:00
like 3rd party source19:00
cbreak fission6: man git-submodule19:01
jast fission6: the 'git-submodule' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-submodule.html [automatic message]19:01
fission6 ok19:01
cbreak there are others19:01
google for subtree merging and git-slave19:02
although I think submodules are best for external deps19:02
Raging_Hog cbreak: thanks for your advice a while ago. I was delayed because I was forced into torture chambers because of my lack of git skills19:03
→ kpreid joined19:05
← Cvbge left19:05
→ Pewpewarrows joined19:05
→ spearce joined19:06
→ marc_kdab_ joined19:06
→ Cvbge joined19:06
→ goshawk joined19:06
unreal-dude ok, so, i got git up and running locally, but remotely i cant seem to find out how/why/w/e this key stuff works with nothing that actually tells me what is happening with it19:09
it just continues to ask for a password19:09
← AzaToth_work left19:10
→ guardian joined19:10
→ Deesl joined19:12
Metathink hi, is there a command to display an entire file's content from a specific commit ? without checkout it19:12
rane_ git show SHA:path/to/file19:13
Metathink I would like to display the original file, not the diff19:13
rane_ indeed19:14
PerlJam Metathink: did you really want to see the diff?19:14
oh,19:14
PerlJam misread19:14
← Mkaysi left19:15
← MattDiPasquale left19:15
→ pgib joined19:15
→ opalka joined19:16
← opalka left19:16
→ opalka joined19:16
→ bcardarella joined19:17
→ pfrenssen joined19:17
→ avinashhm joined19:17
→ xerora joined19:18
← Exposure left19:18
Metathink all the tools i'v found was to display object's content, no file's content19:18
rane_ git show SHA:path/to/file19:19
→ bkosborne joined19:19
Metathink Oh, i'v just understand what i'm supposed to do x) thank you rane_19:20
rane_ np19:20
← DiGitalX left19:21
← vsayer left19:21
← sebersole left19:21
→ vsayer joined19:22
→ MattDiPasquale joined19:22
← sonink left19:22
← LouisJB left19:22
← panike left19:22
→ rcu joined19:23
opalka → opalka_afk19:24
WebDragon → WebDragon|coding19:25
← Zephyrus left19:26
← saysjonathan left19:27
→ saysjonathan joined19:27
← SethRobertson left19:28
← lol_h left19:28
→ zorzar joined19:31
← alvaro_o left19:31
→ nchaimov joined19:31
→ alvaro_o joined19:31
→ j416 joined19:32
opalka_afk → opalka19:32
← nchaimov left19:32
→ priidu joined19:33
→ nchaimov joined19:33
← pgib left19:34
→ tjholowaychuk joined19:35
→ Eridius joined19:36
→ ptl joined19:37
← ashc left19:38
← saysjonathan left19:39
→ Textmode joined19:40
→ saysjonathan joined19:40
← rcu left19:41
→ rcu joined19:41
← simplechat left19:41
→ SethRobertson joined19:42
WebDragon|coding → WebDragon19:42
→ LouisJB joined19:42
→ trupheenix joined19:43
trupheenix hi. i made a commit few days back. someone else made commits on top of it. but i want to make some minor change in my old commit. what do i do?19:44
SethRobertson If you pushed, you should make a second commit19:44
rane_ not a good idea to change old commit like that if it's already public19:45
→ tvw joined19:45
stew yeah, because if it is public, someone might have made commits on top of it, doh!19:45
→ psoo joined19:45
NfNitLoop if you want to show that you are making a fix to that commit, you could: 1) Create a branch at that commit. 2) put your fix on top of that. 3) Merge your branch back into mainline.19:46
→ lazzer joined19:46
NfNitLoop but it's probably easier to just put your fix in your mainline branch. :)19:46
ChanServ set mode: +v19:47
← bronson left19:47
lazzer Hi! Can i configure git on the server so it is impossible to remove tags on it?19:48
→ biodegabriel joined19:48
biodegabriel When I install git on my webserver, should I avoid using git as root? Is it better to create another user with sudo powers and do "sudo git pull...." etc...??19:49
→ mutex joined19:50
fr0sty lazzer: http://stackoverflow.com/questions/2022438/dissallow-deletion-of-master-branch-in-git19:50
adapt that ^19:51
biodegabriel: don't even sudo.19:51
lazzer fr0sty: Thanks!19:51
→ mabrand joined19:51
fr0sty just give the user rights to a directory and leave it at that.19:51
← trupheenix left19:51
biodegabriel ok, thanks fr0sty19:51
← mabrand left19:52
← rcu left19:52
→ rcu joined19:52
← kpreid left19:53
→ thierryp joined19:53
→ urkud_ joined19:54
← SethRobertson left19:55
← msierks left19:55
← cwebber left19:55
← Exodist left19:55
← bkosborne left19:55
→ Exposure joined19:55
← joel_falcou left19:56
→ joel_falcou joined19:57
← Soliton left19:58
← fr0sty|mobile left19:59
← urkud_ left20:01
← Nickeeh left20:01
→ urkud_ joined20:02
← ptl left20:03
→ harcourtfentonmu joined20:03
→ kpreid joined20:03
← rcu left20:03
← thierryp left20:05
→ SethRobertson joined20:06
→ justin-george joined20:06
← _Vi left20:06
→ thierryp joined20:06
← mikepack left20:06
→ _Vi joined20:06
→ panike joined20:07
→ agentmagnetic joined20:07
→ Soliton joined20:07
← thierryp left20:07
← Soliton left20:07
← mrconnerton left20:08
→ tizbac joined20:08
← jrmuizel_ left20:09
← agentmagnetic left20:09
→ rcu joined20:09
→ Soliton joined20:10
← spaceonline left20:10
← justin-george left20:10
→ justin-george joined20:10
→ Oxf13 joined20:11
Oxf13 anybody around that hacks on the GitPython module?20:11
I think I've found a bug in it.20:11
→ werdan7 joined20:11
← tizbac left20:13
→ jrmuizel_ joined20:14
→ msierks joined20:14
→ cwebber joined20:14
→ Exodist joined20:14
→ VAAABOU joined20:15
→ flea joined20:15
→ emptyflask_work joined20:15
flea how can i use localhost filesystem path as REMOTES for push ?20:15
← bluenovember left20:15
flea i do not understand how the URI is specified for local20:15
fr0sty man git-clone, look at the 'GIT URLS' section20:16
jast the 'git-clone' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git-clone.html [automatic message]20:16
flea ty fr0sty20:16
← urkud_ left20:16
← doener left20:16
→ mrconnerton joined20:17
← Soliton left20:19
→ urkud_ joined20:20
→ silence joined20:20
silence → Guest3697120:20
→ mabrand joined20:20
← Guest36971 left20:21
← steph021 left20:21
← BobJelly left20:23
← curtana left20:23
→ montylounge joined20:24
← _Vi left20:24
→ leavittx_ joined20:24
→ _Vi joined20:24
← Mpenz_ left20:25
→ BobJelly joined20:26
→ Nickeeh joined20:27
→ Mpenz joined20:27
← linusoleander left20:27
← TheFox21 left20:28
→ Soliton joined20:28
→ EricInBNE joined20:29
Davey am I completely missing how to tell in github what branch a commit was made to?20:31
fr0sty Davey you can only know what branches contain that commit, not what branch it was originally made on. (that is true of git in general)20:32
Davey fr0sty: OK, fine, how do I get github to tell me that? :P20:32
wereHamster you can't20:32
← rcu left20:32
← hagabaka left20:32
wereHamster but if you clone the repo git will be able to tell you all kinds of informations20:33
→ ptl_ joined20:33
Davey I have a clone; how can I tell from there? :D20:33
wereHamster git branch --contains <commit>20:33
← robotarmy left20:35
← mando left20:35
← maattd left20:36
← opalka left20:37
← Soliton left20:37
← leo2007 left20:38
→ andromak joined20:38
→ Soliton joined20:39
← Oxf13 left20:39
← priidu left20:40
← andromak left20:40
← MikhX left20:41
← ptl_ left20:42
→ aesy joined20:42
msierks cbreak, I setup ssh access to git but I do not want to allow shell access similar to github, how would i accomplish this ?20:43
cbreak gitolite20:43
← unreal-dude left20:44
← vitor-br left20:45
→ rcu joined20:45
← emptyflask_work left20:45
→ steph021 joined20:46
← steph021 left20:46
→ steph021 joined20:46
← kpreid left20:48
← mabrand left20:50
← LouisJB left20:51
← ChrisPitzer left20:52
flea whats the order of branch:branch (local vs remote) when doing push?20:52
rane_ local:remote20:53
wereHamster <src>:<dst>20:53
.. is more accurate20:53
← steph021 left20:53
→ steph021 joined20:53
← steph021 left20:53
→ steph021 joined20:53
→ pantsman joined20:53
← pantsman left20:53
→ pantsman joined20:53
flea ty randomsort wereHamster20:54
err rane_20:54
→ steph021_mac joined20:54
← steph021_mac left20:54
→ steph021_mac joined20:54
← steph021 left20:54
steph021_mac → steph02120:54
flea for me to finally begin to understand scm feels like finding the answer to life.20:55
wereHamster no, that's 4220:55
flea hahaha20:55
← Yuuhi left20:56
flea understanding scm = understanding 4220:56
→ bradleyayers joined20:57
→ kpreid joined20:58
→ unreal-dude joined20:58
→ hasrb joined21:00
← dekebalus left21:00
cbx33 like HEAD21:01
→ sechrist joined21:01
cbx33 is there something which points to the start of the repo?21:01
← Mannequin left21:01
← avinashhm left21:02
← jrmuizel_ left21:02
wereHamster no21:03
because there can be multiple root commits21:03
← Raging_Hog left21:03
→ jrmuizel joined21:04
→ kipster joined21:05
← shyam_iyer left21:05
→ Yuuhi joined21:07
← pfrenssen left21:07
→ silverfix joined21:08
silverfix hello21:08
← javaanse_jongens left21:10
→ MikhX joined21:10
← stamina_ left21:10
← cilly left21:10
silverfix I create .gitignore in my local repo, I commit and push for make .gitignore available by all my team. But now I'd want to edit it not involving the remote repo, is there a way?21:11
← joel_falcou left21:11
→ coreyw joined21:11
fr0sty silverfix: use .git/info/excludes21:12
coreyw I made some updates to my project, then committed and pushed the changes to github.com, but it didn't include a directory. I must have forgetten to add it. But now when I try to do "commit -a" it says there is nothing to commit. So now my local repo is diff. from the remote one, and I don't know how to make them the same again.21:13
fr0sty coreyw: 1. get off the git commit -a habit, 2. understand that -a only updates previously added files.21:13
FauxFaux coreyw: -a doesn't make untracked files tracked, is the directory tracked?21:13
← mattly left21:13
coreyw FauxFaux: I think so. I used "git add directoryname"21:14
slonopotamus coreyw: git add -A && git commit21:14
← h32Lg left21:14
→ bentob0x joined21:14
coreyw slonopotamus: it returns "nothing to commit"21:14
FauxFaux Then the directory is committed unless add told you otherwise.21:15
fr0sty or it is ignored...21:15
FauxFaux add would tell you that.21:15
coreyw But when I try to push it, it doesn't update the remote though.21:15
fr0sty does 'git ls-files' list it?21:16
FauxFaux I care.21:16
FUU21:16
coreyw fr0sty: nope21:16
← tstclair left21:16
fr0sty is the directory empty?21:16
coreyw Oh. Duh. Forgot to add .gitignore files21:17
fr0sty :-)21:17
coreyw thanks!21:17
← artefon left21:17
→ ptl_ joined21:18
← kipster left21:21
→ bronson joined21:21
→ shyam_iyer joined21:22
ceej is this right? git submodule foreach git submodule update21:24
rchady git submodule update should do what you want I believe21:24
→ tstclair joined21:25
ceej it doesn't seem to because if I go into the submodule and do git pull origin master it gets updates21:25
git submodule foreach git pull origin master21:25
works21:25
cbreak ceej: update is not supposed to pull21:25
→ javaanse_jongens joined21:25
cbreak update is supposed to get the _correct_ commit21:25
NOT the newest21:26
→ thiago_home joined21:26
← aah_ left21:26
← msierks left21:26
SethRobertson ceej: If you are doing that all of the time (updating all sub repos to master), perhaps gitslave will be more to your liking.21:26
← icwiener left21:27
→ h32Lg joined21:27
ceej I'll have to take a look... but that command worked nicely21:27
← _Vi left21:27
cbreak git submodule is for strong binding of an exact commit in the submodule to a commit in the parent21:27
← ph^ left21:28
coreyw okay, so I just entere "git commit" and now its asking me for a commit message. How can I abort this command? I mistyped something21:28
SethRobertson gitslave is a loose binding (only tight at tags) but it is easier to use and think about, as long as you are not doing archeology21:28
rane_ coreyw: just quit without saving21:29
→ mattly joined21:29
rane_ how to do that depends on editor it launched…21:29
cbreak or save an empty file21:29
coreyw rane_: hmm, alright.21:29
→ silence joined21:29
→ mabrand joined21:30
silence → Guest181721:30
→ pfrenssen joined21:30
← Guest1817 left21:30
→ Mannequin joined21:30
→ priidu joined21:31
← silverfix left21:32
← coreyw left21:32
← Cvbge left21:32
← ptl_ left21:33
← rayners left21:35
→ robotarmy joined21:35
← robotarmy left21:35
→ robotarmy joined21:35
kpreid Is there a common convention for the message for commits that are 'broken code'?21:35
(obviously, don't put them on the master branch)21:36
in this case I have a big change which breaks stuff; I want to go back and redo it in little, tested pieces, and I'm committing the big change on another branch so I can diff with it21:36
← TrineLocke left21:37
cbreak kpreid: I usually make a tag for such stuff21:37
← robotarmy left21:37
cbreak dead/somename21:37
kpreid never used tags, but I can see that, but I don't think that especially matters.21:38
j416 I usually make a branch for it and I write "well this didn't work" for commit message21:38
.. :P21:38
kpreid my question is specifically, what to put in the commit message21:38
bremner "FAIL"21:38
→ TheFox21 joined21:38
rane_ I sometimes use "WIP"21:38
j416 I prefix my throw-away branches with tmp21:38
cbreak "If you can see this, then you should turn back"21:39
j416 that way I have them separate21:39
and I know which ones to kill later21:39
cbreak I use TMP: for commits that I want to later amend21:39
kpreid eh, I'll just label it WIP21:39
→ Zhonghua joined21:39
kpreid thanks for the suggestions21:39
j416 I use WIP for commit message for half-finished work21:39
then I amend and write a proper message21:39
→ skyl joined21:39
← goshawk left21:40
j416 if I'm not going to continue with it, but want to keep a reference of it21:40
→ ISF joined21:40
j416 I commit it explaining what doesn't work, and create a new branch for it21:40
(with tmp prefix)21:40
then I delete that branch when I fixed things21:40
cbreak I only create branches for stuff I work with21:40
for old stuff I just keep I make tags21:41
that keeps git branch clean21:41
j416 interesting.21:41
→ catalin joined21:41
j416 I like to have my tags clean21:41
I use tags for versions21:41
→ linusoleander joined21:41
cbreak I do that too21:41
but my version tags look like v-0.1-release21:41
and my keep-stuff-for-later look like dead/thishouldhaveworked21:42
← leavittx_ left21:43
→ OOPMan joined21:46
→ ptl joined21:48
→ kylehayes_ joined21:49
← g0bl1n left21:49
→ oracal joined21:49
← murr5y left21:50
→ dmilea joined21:50
← OOPMan left21:51
← rcu left21:51
← kylehayes left21:51
kylehayes_ → kylehayes21:51
→ Dieterbe joined21:53
Dieterbe Hi all! is this not the right syntax to enable the union merge driver for a file named 'todo' in the root of the repository?21:54
cat .gitattributes21:54
todo merge = union21:54
→ asoltys joined21:54
Dieterbe when i do a git merge like this, it will conflict the todo file, so it still seems to be using the text merge driver21:54
Textmode → angrymode21:54
← mezgani_ left21:55
→ p3rror joined21:55
→ dgao joined21:57
← dgao left21:57
→ Cynos joined21:57
→ leavittx_ joined21:58
← catalin left21:59
← Heimidal left21:59
Cynos Hi all, I'm trying to use a pre-commit hook to sanity check some things, and it doesn't fire when git merge does a commit? But if I use git merge --no-commit and then commit manually, it does... anyone know a way aroun dthis?21:59
← mrconnerton left21:59
← alvaro_o left22:00
→ alvaro_o joined22:00
← crash\ left22:01
← booginga left22:02
→ Heimidal joined22:02
← jstemmer left22:02
← JonasF_ left22:02
→ crash\ joined22:02
→ eletuchy joined22:03
ChanServ set mode: +v22:03
← ben_alman left22:03
→ benishor_ joined22:04
← IslandUsurper left22:04
← Pewpewarrows left22:04
→ _Ants_ joined22:04
angrymode → EGAMode22:05
← montylounge left22:05
← ISF left22:05
asoltys hi, i have an svn branch for mirroring to an svn repo using git-svn. it's way out of date so i want to take the current master branch and create a single commit that I can dcommit to the svn upstream. I was thinking i could use merge --squash but i'd rather just "patch" the svn branch. can i do this with git diff somehow?22:06
← bradleyayers left22:06
← Titosemi1 left22:06
context pretty sure you can do it with rebase22:06
cbreak asoltys: that's what git merge --squash does22:06
context or just rebase your changes on top of the svn branch and dcommit all of them22:07
← benishor left22:07
→ OOPMan joined22:07
→ rcu joined22:08
asoltys cbreak: but i'm getting conflicts... i almost want to "reset" it to master, but preserve the svn history22:08
← nona left22:08
→ khaije|minder joined22:09
← Ryback_ left22:09
khaije|minder heya, how do i reset my workingdir to my index?22:09
asoltys context: hmm, i might try that. would rebasing copy merge commits in master onto the svn branch as well though? cause i know git-svn doesn't like merge commits22:09
→ nona joined22:09
khaije|minder the index-stage (to be precise)22:09
← ShooterMG left22:09
→ Quilck joined22:10
← shyam_iyer left22:10
→ murr5y joined22:11
context asoltys: not sure, doubtful.22:11
i avoid git-svn now22:12
→ MikhX_ joined22:12
← TheFox21 left22:12
patrikf khaije|minder: checkout (-f), plus clean if you want files not in the index removed22:12
← jherdman left22:13
← MikhX left22:13
MikhX_ → MikhX22:13
→ maattd joined22:13
khaije|minder patrikf: thanks, i would expect this to co my last commit, the changes i want to restore were 'git add --all' but not committed... would what you said still work?22:15
← kumbayo left22:15
← ChUrP left22:15
← orm_ left22:16
→ mattly_ joined22:16
→ robotarmy joined22:17
← n3m left22:17
→ ormaaj joined22:17
→ vitor-br joined22:18
← hobodave left22:19
← mattly left22:19
← murr5y left22:19
→ mattly joined22:19
patrikf khaije|minder: yes, what you describe is what checkout HEAD would do22:19
checkout without a tree-ish refers to the index22:20
← rcu left22:20
→ rcu joined22:20
patrikf it's actually documented! yay. although not mentioned in the description22:20
how could I ever have doubted the git manpages22:21
khaije|minder ahhh! cool, i would have expected this to appear under git reset --restore-index , so i was looking in the wrong place22:21
→ JATCommand joined22:21
← JATCommand left22:21
→ murr5y joined22:21
→ _Vi joined22:22
→ shyam_iyer joined22:22
khaije|minder soo happy for that! I lost track on several days of work in the tangle of emacs buffers then couldn't find it, so i staged, then closed the app heheh22:22
← rcu left22:23
→ hobodave joined22:23
← mattly_ left22:23
→ rcu joined22:24
← fission6 left22:24
← Maletor left22:24
→ fission6 joined22:24
← ThomasLocke left22:24
← fission6 left22:25
← rcu left22:25
← shyam_iyer left22:25
→ shyam_iyer joined22:25
→ mando joined22:25
→ rcu joined22:26
→ frooh joined22:26
frooh is there a way to, with git add -p, add line 4 but not line 5?22:26
usually directly following lines won't split apart22:27
← rcu left22:27
← priidu left22:28
← nona left22:28
cbreak frooh: edit mode22:29
press e instead of y or n22:29
→ rcu joined22:29
frooh cbreak: I think I tried that but ended up creating a garbage diff22:29
→ helo joined22:29
cbreak well22:29
you have to know what you're doing :D22:30
frooh I don't :-)22:30
← rcu left22:30
kakos So, "git mv" appears to just be a convenience function and git detects renaming using some combination of black magic. This is causing an issue where I move a file, make some changes to the file in its new place, and then git suddenly no longer recognises it as being a renamed file and instead treats it as the old location being deleted and the new location being a new file. Is there a way to force git to treat a new file as a mo22:31
version of an tracked file?22:31
← ormaaj left22:31
← crash\ left22:31
→ rcu joined22:31
khaije|minder oh, i just realized i could have committed without changing the proper index then just reset to HEAD22:31
← Cynos left22:31
fr0sty kakos: no black magic, just comparing file contents. if they are similar enough it is treated as a copy/move22:32
→ darkredandyellow joined22:32
fr0sty git log takes some options to be more liberal with what it considers copying.22:32
→ crash\ joined22:32
← gusnan left22:32
patrikf khaije|minder: well, that way you learnt something new22:32
→ ormaaj joined22:32
fr0sty kakos: if you want to be really clear in your history make one commit that moves, and another that modifies...22:33
khaije|minder patrikf: fo'reals, it's def worth knowing that co takes the index by default22:33
← m_3 left22:33
← rcu left22:34
kakos fr0sty: How does "git log" help me when I commit?22:34
→ rcu joined22:34
khaije|minder still somewhat scared to do it...22:34
fr0sty kakos: nothing in the commit records file renames. it is all detected after the fact.22:34
→ cirwin joined22:34
kakos fr0sty: Sure, but it would be nice if it was in one commit, since the move and modification is one coherent change in functionality.22:34
cbreak you can do that too22:34
fr0sty how big is the modification?22:34
kakos It's only a couple of files, but the changes in each file are enough to trip up git into thinking it wasn't renamed22:35
→ alin-gym joined22:35
→ priidu joined22:35
darkredandyellow I have the exact same problem: http://stackoverflow.com/questions/1072171/how-do-you-remove-an-invalid-remote-branch-reference-from-git22:35
→ tydeas joined22:36
darkredandyellow unfortunately the accepted solution git gc --prune=now22:36
← rcu left22:36
darkredandyellow didnt do anything22:36
khaije|minder kakos: you can have an intermediate commit, then squash/amend it when you're done the heavy lifting22:36
tydeas sitaram: there?22:36
darkredandyellow git branch -avv still shows remotes/origin/staging22:36
khaije|minder i sometimes keep a rolling per-session commit22:36
darkredandyellow although "staging" doesnt exist anymore remote22:36
fr0sty kakos: try 'git commit --dry-run' and see what it says about the files?22:36
frooh darkredandyellow: did you try just rm .git/refs/remotes/origin/staging?22:36
darkredandyellow: kindav a giant hammer, but with a short22:37
shot*22:37
kakos fr0sty: Well, if yous ay the commit history doesn't actually store anything about renames and detects it after the fact, I guess there's really nothing to do.22:37
→ chrischambers_ joined22:37
panike You can always do git remote prune <remote>22:37
darkredandyellow frooh, .git/refs/remotes/origin/ just contains HEAD22:37
→ rcu joined22:37
← mattly left22:37
fr0sty kakos: try the --dry-run. rename detection in 'git status' is 'rename detection lite'.22:37
frooh darkredandyellow: ok, the it's in .git/packedrefs (I think) which is just a tiny text file22:38
← chrischambers left22:38
chrischambers_ → chrischambers22:38
kakos fr0sty: Okay. Thanks for the hlep22:38
fr0sty the result of 'git commit --dry-run' is the genuine article22:38
frooh darkredandyellow: editing that by hand works usually22:38
fr0sty might be different, might not.22:38
darkredandyellow panike, frooh git remote prune origin did the job, thanks for the help22:38
frooh I had problems with prune when I'd removed the remote and for some reason it didn't prune the refs for me22:39
but yeah, prune is really the right answer :-)22:39
← gaspar|work left22:39
darkredandyellow frooh, pandur although, I just realized that it removed *every* remote reference, even for tracking branches that still exists remote22:39
darkredandyellow sighs22:39
panike frooh: After you remove the remote, git does not have the url to check for what should be pruned. Then you have to do surgery22:39
→ tcv_br joined22:39
darkredandyellow git, why are doing this to me?22:40
frooh panike: right, but usually it removes all the refs for me22:40
helo i wrote this silly partial sha1 collision (only the first 7 characters) script, but it doesn't seem to ever actually collide: http://bash.pastebin.com/fdm0gbkT22:40
cirwin helo: how many trials did you do?22:40
panike darkredandyellow: just do git fetch origin. It will repopulate the remote branches22:40
← tcv_br left22:41
khaije|minder you mean git remote update origin ?22:41
← rcu left22:41
wereHamster darkredandyellow: git remote prune should not delete branchs which still exist in the remote. If it did it would be a bug22:41
helo based on a birthday attack on 28 bits, it should only take several thousand attempts to have a 50% chance, right?22:41
→ rcu joined22:41
→ m_3 joined22:41
← TomyLobo left22:41
helo cirwin: it's been running for weeks22:41
so millions? i thought it must be a bash scripting error, but i can't find it22:42
darkredandyellow wereHamster, panike, frooh hehe, I just realized that g b -avv probably wanted to display remote references, but got stopped by error: branch 'origin/HEAD' does not point at a commit22:42
like here: http://pastie.org/157252222:42
cirwin if you've done millions, then you should have seen a collision, so you have a bug22:42
← LongBeach left22:43
patrikf "--amend" sounds fishy22:43
darkredandyellow so, of to gmail to write an email that they fix my head reference...:-)22:43
patrikf do you actually keep the old commits?22:43
cirwin patrikf: yes, they're in the reflog22:43
wereHamster darkredandyellow: who do you want to fix it?22:43
cirwin helo: that said, you might have been "really" unlucky :p22:43
wereHamster darkredandyellow: the problem is in your local repo. So write an email to yourself...22:43
← bronson left22:43
helo heh22:43
← MikhX left22:43
cirwin hmm22:44
← nb left22:44
cirwin I don't even know what you're trying to do with that script though22:44
← lessthanzero left22:44
darkredandyellow wereHamster, the guys at sourcerepo. I see your point with the local repo BUT HEAD was pointing to branch "staging" remote as well. this branch doesnt exist anymore AND I can't change that remotely22:44
← jmil left22:44
darkredandyellow so I guess *they* need to change that for me22:44
fr0sty helo: are you saving and comparing each new sha against the millions you have already calculated?22:44
cirwin if you're just comparing the previous to the current, then you're not exploiting the birtday paradox22:44
darkredandyellow wereHamster, or why I would I be wrong with that?=22:45
helo the script takes one 7-character hex argument, and looks for collisions with that22:45
fr0sty if not you have a 1 in 268M chance of that working...22:45
that's not the birthday paradox.22:45
patrikf helo: that's not how the birthday paradox works, unless that's also a commit like the ones you generate22:45
← BobJelly left22:45
← t_b left22:46
patrikf well, it *is* the birthday paradox, but the wrong side of it :P22:46
cirwin you ought to compare every commit to every other commit — but there's got to be a quicker way of doing that than using git :p22:46
helo ahh, i see22:46
fr0sty it's more the lotter-ticket-buyers-paradox...22:46
helo so a 1 in 16**7 chance22:46
panike darkredandyellow: The problem is in your local repo. I think you want to do rm .git/refs/remotes/origin/HEAD and then try again22:46
→ gfto joined22:46
helo surely i've done more than that many iterations by now... oh well22:47
context git remote set-head origin -d works also22:47
← pandur left22:47
→ asmega joined22:48
context i always remove remote HEAD that bothers me when i git branch -v -a22:48
fr0sty helo: doing 10 per second (generous, I think) you only do 6M/wk22:48
panike Better is git branch -D -r origin/HEAD22:49
← rcu left22:49
fr0sty even 100/sec is only 60M so your 50% time is ~44wks22:49
darkredandyellow panike, ah, you were totally right, thanks a ton!22:49
→ t_b joined22:49
panike np.22:49
← panike left22:49
→ bronson joined22:49
patrikf helo: also, if your dupe-checking logic is the one in the script, it's flawed, as has been mentioned before22:49
→ BobJelly joined22:50
patrikf er, forget it, that's the same issue22:50
← ExtraSpice left22:50
← StyleWarz left22:50
→ rcu joined22:50
← LiamH left22:50
← starcoder left22:51
← asmega left22:51
cirwin helo: it's really cool that you actually thought to try that — even if it did go wrong. Top marks for effort :)22:51
patrikf thinks it would have been more spectacular if helo found a full sha1 collision :P but oh well...22:52
cirwin hehe — there's always a chance :D22:52
→ MikhX joined22:53
→ starcoder joined22:53
→ ldh joined22:54
← leavittx_ left22:54
→ McManiaC joined22:54
→ angelsl joined22:54
→ ki__ joined22:55
← h32Lg left22:55
← q0rban left22:55
uau seems that cherry-pick can trigger recursive merge code that ends up overwrite file contents on disk22:55
← pedrobelo left22:55
cirwin uau: git stash!22:55
uau even in cases where the cherry-pick doesn't actually do anything22:56
cirwin the cherry-pick doesn't commit if it has a conflicting merge22:56
← rcu left22:56
uau cirwin: i mean it's clearly buggy22:56
cirwin heh22:56
uau it ends up overwriting files on disk22:56
even if it (correctly) exits with nothing to do because the change was already applied22:57
→ xiong joined22:57
cirwin ah, interesting22:57
← Heimidal left22:57
fr0sty uau: write up a proof_of_evil and send it to the mailing list.22:57
← eletuchy left22:58
→ rcu joined22:58
uau 'git cherry-pick X; echo >foo.c; git cherry-pick X' and the change to foo.c will have been silently overwritten and it's back to HEAD content22:58
→ aah_ joined22:58
← McManiaC left22:58
→ pedrobelo joined22:58
wereHamster uau: did the cherry-pick succeed or fail due to a conflict?22:59
cirwin uau: it is at least documented that you should use a clean HEAD with no changes22:59
uau succeed22:59
helo cirwin: heh thanks22:59
uau the overwritten files are ones unrelated to any changes in the cherry-picked commit23:00
wereHamster uau: I can't reproduce. Can you make an example?23:00
uau i think it's related to what files had been moved/renamed in other changes since the branches diverged23:00
wereHamster by example I mean a transcirpt of commands we can run locally23:00
← alester left23:01
← oracal left23:01
uau wereHamster: i think it requires some complex enough history to trigger the recursive merge code23:01
← deryl left23:01
fr0sty wereHamster: random question: are you aware that 'git clean -df' will remove directories filled with .gitignored files?23:02
wereHamster fr0sty: are those files tracked?23:02
→ gnat42 joined23:02
→ deryl joined23:02
fr0sty no, just ignored.23:02
→ Heimidal joined23:02
khaije|minder patrikf: git co -f doesn't seem to have worked, was there something i missed in your advice?23:02
← tvw left23:03
fr0sty echo "foo/*" >> .gitignore; touch foo/bar; git clean -df => removing foo/23:03
gnat42 so git newbie here, I have a project, it has a master 'branch' and a couple other branches... I was working in a branch, checked-out master and ran git merge branchname... but in reality I only wanted to cherry pick a couple commits from that branch23:03
is there an undo?23:03
cirwin if you've just merged, and done no more commits, git reset HEAD^23:04
wereHamster fr0sty: no, I was not. Ask the mailing list. For extra credits write a test23:04
cirwin: ORIG_HEAD or HEAD@{1}23:04
← ldh left23:04
cirwin fr0sty: there's a patch set in progress about "treat ignored files as special" or something like that23:04
I think it's known about23:04
fr0sty cirwin: good to know.23:05
the _real_ solution is echo foo/ >>.gitignore instead of foo/*23:05
cirwin wereHamster: what is ORIG_HEAD?23:05
wereHamster it's the HEAD before the last operation23:06
cirwin does it ever differ from HEAD@{1} ?23:06
wereHamster I don't think so, but don't take my word for it23:06
cirwin ok23:06
thanks23:06
context cirwin: ORIG_HEAD im pretty sure is used during merge/rebase/bisect23:06
wereHamster ORIG_HEAD is older, the reflog came much later23:06
gnat42 cirwin: that didn't do anything23:06
cirwin ah, maybe in a rebase it's the head before that started23:07
cirwin will play23:07
← deryl left23:07
← cbx33 left23:07
gnat42 it likely did a fast forward merge if that makes a difference23:07
ayust http://stackoverflow.com/questions/964876/head-and-orig-head-in-git23:07
→ deryl joined23:07
uau revert_or_cherry_pick at builtin/revert.c:605; do_pick_commit at builtin/revert.c:510; do_recursive_merge at builtin/revert.c:350; merge_trees at merge-recursive.c:1545; process_entry at merge-recursive.c:1371; merge_content at merge-recursive.c:1298; update_file_flags at merge-recursive.c:68723:07
cirwin gnat42: ah, yes, that does; wereHamster had it right23:08
uau that's the call path that ends up overwriting files23:08
helo fr0sty: looks like it gets about 200/s... so this will take a while :/23:08
← rcu left23:08
← lazzer left23:08
gnat42 so I should reset ORIG_HEAD ?23:08
wereHamster gnat42: that or HEAD@{1}23:09
gnat42 ok so git reset ORIG_HEAD does nothing23:09
→ rcu joined23:09
cirwin HEAD@{1} will get you back to just having done the merge (before you did reset HEAD^), you might need to do HEAD@{2} to go back to before the merge too23:09
← shyam_iyer left23:09
← xiangfu left23:09
cirwin but it's now at the point where you need to look in the reflog23:09
gnat42 all23:09
cirwin and get the hash out23:09
gnat42 ah23:09
cirwin sorry, that's my fault23:09
← selckin left23:09
wereHamster if in doubt, look at git reflog23:10
gnat42 *saweet*23:10
thanks23:10
← rcu left23:10
adamholt → adamholt_away23:11
adamholt_away → adamholt23:11
gnat42 hmm ok - that's not quite doing what I expect23:12
cirwin in what way?23:12
gnat42 as part of the merge a directory in the master branch was deleted, but it never comes back23:12
← Zhonghua left23:12
→ davetoo joined23:12
cirwin you will need to do a git checkout after the reset23:12
(or a reset --hard)23:12
gnat42 ok so suppose I've run reset a couple times23:13
is it safe to do the reset --hard?23:13
can I get back to my 'just merged' state if I do this wrong?23:13
cirwin yes and yes23:13
← frooh left23:13
cirwin the only thing you can't get back is any uncommitted changes23:14
wereHamster you can get any commited state back23:14
→ rcu joined23:14
gnat42 ok so I haven't committed anything... so would like to get back to the point before I ran all these reset commands23:14
reset --hard ?23:14
davetoo can somebody show me where to read about how "'branch.<foo>.remote=. " works? i.e. remote =' .'23:14
gnat42 or will that get me whatever I've reset to?23:14
→ lazzer joined23:15
davetoo Is this (remote=.) purely a git-svn thing?23:15
cirwin gnat42: I can't visualize what you're saying — either: go to the right commit in the reflog using reset, and then checkout -f; or go to the right commit in the reflog using reset --hard23:15
← xerora left23:16
gnat42 cirwin: thanks, I think I figured out the issue23:16
uau according to git bisect, commit 882fd11aff6f0e8add77e75924678cce875a0eaf "merge-recursive: Delay content merging for renames" introduced the file overwrite bug23:16
gnat42 obviously with your guys help23:16
→ hagabaka joined23:18
← hagabaka left23:18
→ hagabaka joined23:18
← d0k left23:18
← v0n left23:19
← dyer left23:19
← TeckniX left23:20
← bhosie left23:20
← alin-gym left23:20
gnat42 ok so reading man git didn't answer this for me, if I want to merge a couple commits in the branch into master... is there a syntax for that? or do I simply create a diff of those commits and then apply them to master?23:20
jast the 'git' manpage can be found at http://www.kernel.org/pub/software/scm/git/docs/git.html [automatic message]23:20
cirwin gnat42: git cherry-pick23:21
← Quilck left23:21
gnat42 thansk cirwin23:21
← mabrand left23:22
gnat42 just a clarification quesiton... do I just pass the commit hash I don't have to specify that the commits I'm looking at are on a different branch?23:22
cirwin just the hash is fine23:23
ayust commits aren't inherently "on a branch"23:23
branches just happen to point to certain chains of commits23:23
gnat42 ayust: yeah asking that question seems to be solidifying the idea in my head - still mainly used to svn23:23
ayust multiple branches can point to the same commit or commit-parent23:23
← Exposure left23:23
← rcu left23:24
→ rcu joined23:24
khaije|minder so i had some files staged in my index, then did a git co -f , I expected this would restore my workdir to the index state, is that not right?23:24
← WebDragon left23:24
ayust no, it restores it to HEAD state23:25
iirc23:25
though i could be wrong.23:25
← unreal left23:26
khaije|minder well it looks like you're right, though i was hoping it wasn't true23:26
→ mrchrisadams_ joined23:27
cirwin the docs do say the index is used if the commit is not specified — but I agree, it seems to use HEAD23:27
ayust i think it only checks out from the index if you specify a path23:27
→ Exposure joined23:27
cirwin ah23:27
who wrote this stuff.. :p23:27
khaije|minder is there anyway for me to get my index back?23:27
→ unreal joined23:28
cirwin I think you can try finding objects that aren't referenced from commits using some jiggery-pokery, and some of those might contain your changes23:29
but I don't know of an easy way23:29
khaije|minder it could be more clear, it would also make sense to just have an --index option23:29
ok, thats what i was thinking23:29
cirwin rarely uses the index, git commit --amend is easier and safer, and I can use --fixup to amend previous commits23:29
khaije|minder not familiar with --fixup , wassat?23:30
← Metathink left23:30
← envi left23:30
cirwin if you have a modern git, rebase -i will squash commits with a "fixup" command, so that they go from being two commits to being one23:30
→ madewokherd joined23:30
← MattDiPasquale left23:31
cirwin git commit --fixup allows you to easily create a commit that is susceptible to being squashed23:31
in that manner23:31
so I do lots of git commit --fixups throughout a session23:31
and then a big rebase -i at the end23:31
→ jmil joined23:31
cirwin to sort out the hitsory23:31
khaije|minder ah cool, i've been doing rolling-amends ,but that clearly better23:32
← pantsman left23:32
→ tvw joined23:32
khaije|minder puts on scuba gear and holds breath, bbl23:32
← bcardarella left23:35
→ pleonasm joined23:35
→ leo2007 joined23:36
← coderdad left23:37
← sternocera left23:38
← jmil left23:39
← chitragupt left23:40
→ chitragupt joined23:41
← t_b left23:44
← priidu left23:45
→ nb joined23:46
ivan is it a good idea to use merge --no-ff to indicate that the commits in the branch are broken versions of the software?23:46
← triptec left23:46
khaije|minder does git preserve any proir index info after a co ? i though it might, but now i'm thinking probably not..?23:46
mrchrisadams_ hi guys, I'm trying to work out in my head what's happening when you add a -f to a git push command, to force a certain push. If you're storing all content in the git objects file, anyway when working with what might you lose if you force a push like so?23:47
← necromancer left23:48
← biodegabriel left23:48
→ linusoleander_ joined23:48
← Determinist left23:49
← rendar left23:49
← linusoleander left23:50
linusoleander_ → linusoleander23:50
← tydeas left23:50
← bronson left23:50
← darkredandyellow left23:51
← cannonball left23:51
kpreid mrchrisadams_: stuff is lost from ordinary view if there are no refs to it (including indirectly through commit parents)23:51
→ lightcap joined23:52
kpreid mrchrisadams_: if you push -f you cause the destination repository to forget about the current history of some ref and use yours instead, so there are some commits with no refs pointing to them23:52
→ darkredandyellow joined23:52
asoltys hi, is there a way i can check-in my svn-remote config stuff to the repo so that when people checkout the branch that's tracking svn, the svn upstream will be configured for them,? i tried copying the options from .git/config into a file called ".gitconfig" in the working tree and pushign that.. no luck23:52
← murr5y left23:52
kpreid and that stuff will go away on a future GC.23:52
← kpreid left23:53
← gnat42 left23:53
← kipras_ left23:54
→ jamiew joined23:54
mrchrisadams_ kpreid: I see, so when I push, I'm still sending my snapshot of the filesystem on my machine, so someone else could check it out on another machine if need be, right?23:55
← cwebber left23:56
mrchrisadams_ and in if I used the git reflog command, I could dig around and find them still if I had the time and inclination?23:56
→ bronson joined23:56
→ murr5y joined23:56
← sraue left23:56
← RobertLaptop left23:57
← nadavoid left23:57
→ sraue joined23:57
→ nadavoid joined23:57
← project2501a left23:57
← skyl left23:58
→ phedhex joined23:58
← ketas left23:58
→ eletuchy joined23:58

Logs Search ←Prev date Next date→ Channels Documentation