IRCloggy #git 2012-05-15

Logs Search ←Prev date Next date→ Channels Documentation

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

2012-05-15

→ MacGyverNL joined00:01
← MacGyverNL left00:01
→ MacGyverNL joined00:01
→ Pupeno_W_ joined00:01
← isao1 left00:01
→ isao joined00:03
sitaram dmlloyd: if you did it at the right spot there should be only conflicts that would have occurred anyway if the right commit was applied earlier. Are you seeing more than that?00:03
→ kerrick joined00:03
← Pupeno_W left00:03
Pupeno_W_ → Pupeno_W00:03
→ dvide joined00:04
→ davidbanham joined00:04
← dreiss left00:05
→ machty joined00:06
← robotmay left00:08
→ robotmay joined00:09
← glennpratt left00:09
← robotmay left00:09
→ h0h0h0 joined00:10
gitinfo set mode: +v00:10
h0h0h0 Sup: What's a good windows tool to draw basic git commit tree pictures with? All the pretty ones seem to be keynote/mac. :(00:10
h0h0h0 wins benign question award?00:11
← kermit left00:11
→ williamcotton_ joined00:11
→ TheShagg joined00:12
dmlloyd sitaram: it seems to have worked. There were no conflicts when I reverted, reapplied, and squashed (of course). Cherry picking the result into master did raise a couple conflicts but they were expected and easy to deal with00:12
gitinfo set mode: +v00:12
← neurodrone left00:12
← williamcotton left00:12
williamcotton_ → williamcotton00:12
→ thursday joined00:12
dmlloyd now I'm just re-checking the resultant diff to be 100% sure that everything went as I expected.00:12
TheShagg so when I do a git rebase -i, i am getting a conflict, and when I try and do a git mergetool, kdiff3 automerges but it does it wrong and I cannot edit it00:13
← Chillance_ left00:13
TheShagg is there a way to make it so kdiff3 does not just automerge and quit to the command line00:13
→ kermit joined00:14
← isao left00:16
← hyperair left00:16
→ flazz joined00:18
← kerrick left00:19
← gmarseliss left00:22
→ project2501b joined00:22
→ JEntrep joined00:24
← JEntrep left00:24
← amoxibos left00:24
flazz when i do rebase -i i get noop in the editor. non-interactive rebase does a bunch of conflicts. a rebase should not do anything because the working branch is up to date with the parent, according to the tree view.00:25
← maletor left00:26
SethRobertson "i get noop in the editor"?00:27
FauxFaux Sounds like a new dance.00:27
← codeM0nK3Y left00:27
flazz sorry, that was fired off too quickly00:27
SethRobertson I suggest looking at the commits after the one specified in the rebase to see if any of them were merge commits00:28
Recreating merges is errorprone00:29
flazz some of them were conflicts00:29
at least a few00:29
→ luke-jr joined00:29
flazz is there a sane way to progress?00:29
SethRobertson So you are saying that there were some merge commits after the commit you were attempting to rebase?00:30
→ mikepack joined00:30
flazz yes, but they were resolved00:30
so all the conflicts are being recreated as the rebase progresses?00:31
→ twisted` joined00:31
SethRobertson Yes. Attempting to rebase-i with intermediary merges is problematic. Most likely, you should just let history be and if there is anything that needs fixing, just create a new commit.00:32
flazz does that mean rebase --skip?00:32
← palyboy left00:32
→ palyboy joined00:32
← ehsan left00:33
← williamcotton left00:33
→ nesneros joined00:34
flazz and at the end fix all the conflicts?00:34
← nesneros left00:34
← hwrd|work left00:34
infogulch → infogulch|afk00:35
→ luxurymode joined00:35
← tchan left00:35
→ rurufufuss joined00:36
← berserkr left00:37
← lonewulf` left00:37
← _W_ left00:39
→ _W_ joined00:39
flazz SethRobertson: could you elaborate on 'let history be'?00:40
→ w09x joined00:40
→ lonewulf` joined00:40
← w09x left00:41
SethRobertson rebase-i performs !rewriting of potentially public history. It is rarely REQUIRED to rewrite history, normally someone is just trying to better perfect what happened in the past. From the perspective of what you are shipping/deploying, you can make a new commit to fix whatever the problem that occurred in the past was00:41
gitinfo [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)00:41
← joaoh82 left00:41
→ Goplat joined00:42
→ kenperkins joined00:42
→ aidan joined00:42
← marc_kdab left00:43
aidan I have a git repo with another git repo in it ... but I don't want my parent repo to know / care about the sub-repo. Is there an easy solution?00:43
→ enmand joined00:44
← wizbit left00:44
→ w09x joined00:45
flazz given that i've resolved the conflicts, can i tell rebase to use the resolutions that i already made without me redoing it by hand?00:45
← QaDeS left00:45
← mikepack left00:45
← alberto56 left00:46
→ mikepack joined00:46
→ wizbit joined00:46
→ jackh joined00:46
→ hwrd|work joined00:46
← mtytel left00:47
← iwoj left00:49
← mikepack left00:49
← catkiller left00:50
→ oogatta joined00:52
→ mtytel joined00:53
offby1 flazz: man git-rerere00:54
gitinfo flazz: the git-rerere manpage is available at http://jk.gs/git-rerere.html00:54
offby1 always thinks of Aretha Franklin when he says that00:55
← mtytel left00:55
→ amigojapan joined00:58
→ mtytel joined00:59
← enmand left00:59
→ gchristensen joined01:00
← anth0ny_ left01:04
→ weiyang joined01:05
← beatak left01:05
→ igotnolegs joined01:05
flazz thanks guys for all the great info, i think i solved my problem with rebase-p01:06
for future reference, why does rebase not preserve merges by default?01:07
→ steve1_ joined01:08
← steve1_ left01:08
→ intripoon joined01:09
→ steve3 joined01:10
→ D4M0NB joined01:11
EugeneKay aidan - First entry in !subprojects01:11
gitinfo aidan: So, you want to add git repositories inside of other git repositories? Well, you have four main options. First is to just do it, add the repo to the outer project's .gitignore, and treat them entirely separately. Best if they are entirely seperate. Otherwise your best options are "!submodule" "!gitslave" and "!subtree" Try typing those commands into this IRC channel.01:11
← intripoon_ left01:12
← EyesIsMine left01:12
gitinfo set mode: +v01:12
← harish left01:13
← SegFaultAX|work left01:13
→ xiangfu joined01:14
→ Vampire0 joined01:14
→ iwoj joined01:16
← Vampire0_ left01:16
← thelazydeveloper left01:17
← Lgb left01:20
← hwrd|work left01:20
← roger_raymond left01:20
← luxurymode left01:21
← Tinti left01:21
← Destos left01:22
← erkules left01:22
→ Lgb joined01:22
→ gearaholic joined01:24
→ dreiss joined01:27
→ EyesIsMine joined01:27
→ mdhopkins joined01:27
← jstout24 left01:28
→ lindenle_ joined01:28
← gearaholic left01:28
← PrinceAMD left01:29
→ roger_raymond joined01:30
→ _huoxito joined01:32
← TheShagg left01:33
← D4M0NB left01:33
→ gtar joined01:33
→ ehsan joined01:34
← gtar left01:34
→ gtar joined01:34
← gtar left01:34
infogulch|afk → infogulch01:35
→ erkules joined01:35
→ tonesfrommars joined01:35
← teweWORK left01:37
→ xxthink joined01:37
← pdtpatrick left01:38
← ColKurtz left01:39
→ kaushal joined01:40
kaushal Hi01:40
gitinfo kaushal: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.01:40
kaushal can git be used for php based codebase ?01:40
matthiaskrgr why not?01:41
I don't know any reasons why it shouldn't work01:41
offby1 well, it might turn green.01:41
matthiaskrgr you can also use git to track changes in configuration files or images :p01:42
images won't work that well, because you won't see any useful diffs01:42
← marhaban left01:42
→ jcontra joined01:44
→ marhaban joined01:44
← tewecske left01:44
SJS we need a good ASCII diff tool for images.01:47
→ zdk joined01:47
← mischat left01:47
→ duckxx joined01:48
→ duckx joined01:49
→ tchan joined01:49
steve3 !backup01:50
gitinfo Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups01:50
→ iamjarvo1 joined01:50
← roger_raymond left01:50
→ roger_raymond joined01:50
← weiyang left01:50
matthiaskrgr !fun01:50
gitinfo [!refund] If you are not satisfied with git for whatever reason, you are entitled to a full refund of the purchase price, and are invited to use another VCS. Elsewhere.01:50
→ glennpratt joined01:51
← iamjarvo left01:51
← roger_raymond left01:52
→ roger_raymond joined01:52
→ chaowang joined01:52
→ tmchaves joined01:53
← duckxx left01:53
← Radium left01:53
← meisth0th left01:53
tmchaves Hi. Is there any way to transform a local copy of a repository into a repository that can be cloned?01:53
← Kunda left01:55
← Cam left01:55
matthiaskrgr you can clone local repos01:55
git clone /path/to/repo01:55
kaushal matthiaskrgr: thanks01:55
matthiaskrgr: is git written in "C" ?01:55
→ Nox_ joined01:56
matthiaskrgr uuhmmm01:56
dunno :P01:56
im just a user01:56
← Targen left01:56
kaushal matthiaskrgr: np01:56
matthiaskrgr: git is source code management software ?01:57
← iwoj left01:57
← Bombe left01:57
matthiaskrgr its a version control system01:57
← Nox_ left01:57
kaushal what is SCM ?01:57
matthiaskrgr !scm01:57
→ Cam joined01:57
matthiaskrgr :(01:57
← cenuij left01:57
matthiaskrgr actually I wondered about this, too :)01:58
http://en.wikipedia.org/wiki/Supply_chain_management01:58
could fit01:58
kaushal nope01:58
matthiaskrgr: source code management01:58
matthiaskrgr mh01:58
ah01:58
← dreiss left01:58
kaushal so whats the deal ?01:58
→ dreiss joined01:59
← ehsan left01:59
← abetaha_ left02:00
→ PrinceAMD joined02:00
← PrinceAMD left02:00
→ PrinceAMD joined02:00
← amigojapan left02:00
→ ehsan joined02:02
→ efreese joined02:02
→ amoxibos joined02:02
← amoxibos left02:03
→ anth0ny joined02:04
→ Bombe joined02:04
iadicicco kaushal: yeah, i think it's written in c02:05
kaushal: http://github.com/git/git02:05
→ Destos joined02:07
→ neurodrone joined02:08
← neurodrone left02:08
→ neurodrone joined02:08
← dreiss left02:08
← oriba left02:09
← Destos left02:09
→ harish joined02:10
→ weiyang joined02:11
← AaronMT left02:11
← d4rklit3 left02:11
← fisted left02:13
→ stliu joined02:13
← Araxia_ left02:13
← jcontra left02:13
→ marhaban1 joined02:14
→ jcontra joined02:14
← marhaban left02:16
→ abetaha joined02:16
→ fisted joined02:17
→ nesneros joined02:19
→ ipoval joined02:20
← roger_raymond left02:20
← pandeiro left02:23
← marhaban1 left02:23
aidan !subtree02:24
gitinfo The git subtree merge method is ideal to incorporate a subsidiary git repositories directly in to single git repository with "unified" git history, where you only need to pull changes in from external sources not contribute your own changes back (which if technically possible is at least difficult). See http://progit.org/book/ch6-7.html Type "!subtree_alt" for more options02:24
← anth0ny left02:24
← xiangfu left02:25
← davidbanham left02:26
← WaltzingAlong left02:29
← orafu left02:29
→ orafu joined02:29
→ anth0ny joined02:30
← nesneros left02:30
← kaushal left02:30
→ friskd joined02:32
← kukks left02:36
← Constant_ left02:37
← flazz left02:40
→ xiangfu joined02:44
← Lulzon left02:47
← glennpratt left02:47
← wizbit left02:49
→ wizbit joined02:50
tmchaves where are my project files kept in a git bare repository ?02:53
milki a sekrit location02:53
offby1 shush02:54
milki they are stored as git objects and some magic pixie dust02:54
offby1 don't even kid about it02:54
milki tmchaves: what do you need them for?02:54
tmchaves oh I see02:54
offby1 tmchaves: they're in .git/objects. Only some of those files are your project files, though; others are history and other things. Oh and they're compressed.02:55
tmchaves I have a machine that is used both as a git shared repository and web server, I'd like to use the files to the web server directly from git repository02:55
milki ya....02:56
tmchaves: !website02:56
gitinfo tmchaves: [!deploy] Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh02:56
tmchaves so that when someone updates it (pushes something into it) the webserver is updated02:56
→ wolftankk joined02:58
RandalSchwartz where's the one thatmentions me?02:59
!deploy02:59
gitinfo Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh02:59
RandalSchwartz no - that's not it02:59
crap02:59
EugeneKay !triggers02:59
gitinfo See http://jk.gs/git/bot/trigger.php02:59
← wizbit left02:59
→ Yuuhi` joined02:59
→ h0bbit_ joined02:59
EugeneKay RandalSchwartz - this one, you eman? https://gist.github.com/171423502:59
I turned it into a post-receive hook and expanded it quite a bit.03:00
RandalSchwartz yeah which one mentions that03:00
Oh!03:00
milki uh huh03:00
RandalSchwartz yeah, it should have been a post-receive hook03:00
milki thats why its different03:00
also includes comments03:00
:o03:00
RandalSchwartz thank you thank you03:00
EugeneKay It's the same basic script, but I decided I wanted to use it.... and I like having comments, and features!03:00
tmchaves thanks..03:01
→ wizbit joined03:01
RandalSchwartz yeah. the rsync is essential to avoid updating files that haven't changed03:01
most people forget that. :)03:01
milki just uses a makefile...03:01
EugeneKay I need to write a bit about how to use the --delete and --exclude family in opts03:01
→ Owner joined03:01
← Yuuhi left03:01
RandalSchwartz rsync is senior magic03:02
Owner how do you ignore a directory from being updated during a pull ?03:02
or permenatly03:02
RandalSchwartz pull = fetch + merge03:02
you are merging upstream03:02
← h0bbit left03:02
RandalSchwartz why are you ignoring upstream?03:02
← xiangfu left03:02
← thearthur left03:02
RandalSchwartz rsync is magic magic when you get to --filter='. -'03:03
Owner well i wanted upstream to ignore a directory03:03
RandalSchwartz why?03:03
every directory is important in a source management system03:03
Owner true03:03
milki until you rm -rf subdir/ *03:04
→ Targen joined03:04
RandalSchwartz in a *deploy* system, you might not want a subdir03:05
but git is *not* a deploy system03:05
→ arvind_khadri joined03:05
← arvind_khadri left03:05
→ arvind_khadri joined03:05
← urbanmonk left03:05
Owner well ideally03:06
i want to pull the directory on initial clone03:06
then ignore it after that03:06
so i can make local changes03:06
RandalSchwartz then you don't have source03:06
you have prototypes, which should be under a different name03:06
foo.proto/*03:07
which *you* copy to foo/*03:07
so that foo.proto/* can be updated upstreadm03:07
Owner i see03:07
RandalSchwartz and also pushed if you have a different idea for the prototypes03:07
→ yizumi joined03:07
RandalSchwartz and locally, you can put .git/exclude foo/*03:08
or it could even be done with a .gitignore upstream03:08
if everyone is doing the same thing03:08
milki this isnt a proper tracking branch -.-03:08
← tmchaves left03:09
RandalSchwartz whatever03:09
Owner i think git ignore might be easier03:09
can i not just type "git ignore conf/*"03:11
← wizbit left03:12
← Eridius left03:15
→ xiangfu joined03:16
← eletuchy left03:17
← mdhopkins left03:18
← Maccer left03:20
→ wizbit joined03:21
→ edude03 joined03:21
→ irqq joined03:22
bremner no, that command does not exist.03:22
← Dreamer3 left03:23
→ Dreamer3 joined03:25
→ Radium_ joined03:27
EugeneKay Yay! GitHub brought back the buttons to see Watchers / Forkers!03:27
I like to think that my Support question about that lead to the fix.03:27
← arvind_khadri left03:28
← iamjarvo1 left03:29
→ amoxibos joined03:29
← mcluskydodallas left03:30
← Radium_ left03:31
→ johnanderson joined03:31
← amoxibos left03:31
← subbyyy left03:31
→ amoxibos joined03:31
← amoxibos left03:32
→ amoxibos joined03:32
→ Davey joined03:34
← Kabaka left03:35
→ Kabaka joined03:37
→ Per4u3e joined03:37
→ subbyyy joined03:39
→ Kunda joined03:39
→ ChoiZ_ joined03:41
→ hegge joined03:41
← ptolemyxi left03:41
← _iron left03:41
← ChoiZ left03:41
→ _iron joined03:41
← SerajewelKS left03:42
→ unnamedwill joined03:42
→ SerajewelKS joined03:42
→ dr0id joined03:43
← az left03:43
→ az joined03:46
→ Maccer joined03:47
← Error404NotFound left03:51
← Per4u3e left03:52
← wolfcore left03:54
→ wolfcore joined03:56
milki aha03:57
except theyll never give you credit03:57
→ dc5ala joined03:58
← jackh left03:58
→ dreiss joined03:59
← yizumi left04:02
← ping-pong left04:02
→ yizumi joined04:03
← mritz left04:05
← amoxibos left04:06
← Spudster left04:07
← wizbit left04:08
← helgikrs left04:09
← Kunda left04:10
→ wizbit joined04:10
← irqq left04:11
← Vert left04:12
← edude03 left04:13
→ edude03 joined04:13
→ helgikrs joined04:16
→ Chib joined04:17
← wizbit left04:18
→ wizbit joined04:18
→ dermoth_ joined04:23
← wizbit left04:23
→ jceb joined04:24
→ wizbit joined04:25
← macmartine left04:25
← OOPMan left04:26
← dermoth|home left04:26
← LekeFly left04:27
→ fuzzybunny69y joined04:28
← cyphase left04:28
→ LekeFly joined04:28
← dr0id left04:29
← harish left04:30
→ macmartine joined04:31
EugeneKay I can dream.04:32
→ acl_ joined04:34
→ BarryCrayon joined04:37
→ pdtpatrick joined04:39
→ ilius joined04:40
→ sangi joined04:41
← BarryCrayon left04:42
← cakehero left04:43
← joepie91 left04:43
← LekeFly left04:44
← Tohuw left04:44
→ joepie91 joined04:45
→ yoklov joined04:45
→ Eridius joined04:47
→ Tohuw joined04:47
← wolfcore left04:49
← wizbit left04:50
← johnanderson left04:50
→ iwoj joined04:50
→ mritz joined04:51
→ wizbit joined04:52
← vitor-br left04:52
→ maleknet joined04:53
→ harshpb joined04:54
← krz left04:55
← wizbit left04:59
← klj613 left05:00
→ syphar joined05:02
← n3m left05:03
→ n3m joined05:04
→ _sundar_ joined05:05
→ dr0id joined05:07
← jceb left05:08
← lonewulf` left05:08
← rchavik left05:08
← tonesfrommars left05:10
← smw left05:10
→ ShooterM_ joined05:10
→ mmc1 joined05:12
→ pantsman- joined05:14
← SeySayux left05:16
← MissionCritical left05:16
→ wizbit joined05:16
← click left05:17
← teotwaki left05:17
→ click_ joined05:17
← duckx left05:17
← preflex left05:17
← Dynom left05:17
← rawtaz left05:17
← geeKISSexy left05:17
→ RickXy joined05:17
→ preflex_ joined05:17
→ Yuuhi`` joined05:17
→ duckx joined05:17
→ Dynom joined05:17
→ rawtaz joined05:17
→ teotwaki joined05:17
← n3m left05:18
← kermit left05:18
← pantsman left05:18
← gxk left05:18
← UukGoblin left05:18
→ smw_ joined05:18
→ UukGoblin joined05:18
← orafu left05:18
← djh_ left05:18
→ n3m joined05:18
→ gxk joined05:18
→ arvind_khadri joined05:18
← Yuuhi` left05:18
← arvind_khadri left05:18
→ arvind_khadri joined05:18
preflex_ → preflex05:18
→ orafu joined05:18
← gavin_huang left05:18
← chris2 left05:18
← mikegerwitz left05:18
→ LekeFly joined05:18
→ chris2 joined05:18
→ mikegerwitz joined05:18
← mikegerwitz left05:18
→ mikegerwitz joined05:18
← Sil4nc4_ left05:19
→ djh joined05:19
→ Sil4nc4 joined05:19
← mmc1 left05:21
→ harish joined05:22
← wizbit left05:22
← smw_ left05:22
→ Dave^| joined05:22
→ lonewulf` joined05:23
→ SeySayux joined05:23
← anth0ny left05:23
→ cyphase joined05:23
→ anth0ny joined05:24
→ wizbit joined05:24
← steve3 left05:24
→ gavin_huang joined05:25
→ barrbrain joined05:26
← macmartine left05:26
← crankharder left05:27
→ toabctl joined05:27
← anth0ny left05:28
→ rchavik joined05:29
← rchavik left05:29
→ rchavik joined05:29
→ marc_kdab joined05:32
→ MissionCritical joined05:32
← xiangfu left05:33
← enr left05:34
→ fisted_ joined05:35
← fisted left05:35
→ crankharder joined05:36
← PrinceAMD left05:36
← Davey left05:36
→ khmarbaise joined05:37
→ PrinceAMD joined05:37
← PrinceAMD left05:37
→ PrinceAMD joined05:37
→ _md joined05:38
→ davidbanham joined05:39
← yoklov left05:39
← yuriks_ left05:40
→ yuriks_ joined05:40
yuriks_ → yuriks05:40
← madewokherd left05:40
← wizbit left05:42
← mritz left05:43
→ wizbit joined05:43
← cyphase left05:44
← khmarbaise left05:46
infogulch → infogulch|afk05:46
← mmattice left05:48
← lonewulftp left05:49
→ dan_p joined05:50
→ lonewulftp joined05:50
dan_p Hi, i'd like to get the patch of the conflcit resolution done in a merge commit, but can't seem to find the way to do it?05:51
→ Sigma joined05:51
_ikke_ dan_p: I don't understand your question05:51
← Bass10 left05:52
sitaram that's almost always useless, because the "before" side would contain conflict markers etc. But if you re-do the exact same merge on a fresh repo and then diff that with the resolved tree you should get it05:52
← duckx left05:52
sitaram dan_p: "fresh repo" because if you do it on the same repo, rerere may kick in and not give you a pristine merge05:53
← idoru left05:54
→ mishok13 joined05:54
← slakmagik left05:56
dan_p _ikke_: the best way I can explain is show a gitweb output - I would like to get this output from the tools: http://git.moodle.org/gw?p=integration.git;a=commitdiff;h=0dda790c05bcab5b57a624ac3f6b9c0229d1e47a05:56
→ twb joined05:56
dan_p sitaram: thanks, makes sense. I'm probrably doing something that there is a better way to do05:56
twb Does "git foo --help" call man directly, or is there a variable like $MANREADER ?05:57
← wizbit left05:57
→ wizbit joined05:57
dan_p I am trying to convert someones merge into a cleanly rebased branch, but I wanted to get the conflict resolution diff so I can do it earlier on in the rebase05:57
→ airborn joined05:57
→ tvw joined05:58
← toabctl left05:58
← weiyang left05:59
→ cyphase joined06:00
→ angelsl joined06:00
← WillMarshall left06:01
← marc_kdab left06:01
→ beautifulmind joined06:02
← wizbit left06:03
→ wizbit joined06:05
← cooloney left06:06
→ kadoban_ joined06:10
EugeneKay sitaram - I'm cleaning out the various hooks which have infested my repositories/ directory over the package upgrades.... with g3, all I should have(from gitolite) is the 'update' hook, and 'post-update' in the gitolite-admin.git repo, correct?06:10
sitaram correct06:10
EugeneKay nukes the rest06:11
→ jceb joined06:11
→ cbreak-work joined06:11
sitaram EugeneKay: you can nuke *all* hooks if you like, then run 'gitolite setup' and it will put back what it needs. Just be sure no one attempts to push in the small window between nuke and setup.06:12
→ jackh joined06:12
EugeneKay I have a few with a post-receive hook(my ! deploy script) that I don't wanna tread on06:13
sitaram (just FYI, in case you were preparing a script or something)06:13
aah ok06:13
EugeneKay I just abused find a bit, did the trick06:13
sitaram there is no abuse; there is only power and those too weak to use it06:15
EugeneKay Am I also correct that upgrading is 'git pull; ./install; gitolite setup' ?06:16
→ ThomasLocke joined06:17
← thierryp left06:17
EugeneKay Actually `./gitolite/install -ln`, looking at my bash history.06:18
→ glcrazy joined06:18
sitaram upgrading from g2 to g3? no....! within g3 from one version to another, yes06:18
EugeneKay The latter06:18
← dcorbin left06:19
EugeneKay I've managed to git all of my gitolite instances up to g3, finally.06:19
← Slush-_ left06:19
→ papegaaij joined06:19
sitaram yes that will work. In fact the documentation doesnt go into details but in the "-ln" case you don't even need that because it's only doing a symlink (which is a no-op if you have not moved your clone)06:19
cool06:20
← Slush-__ left06:20
← wizbit left06:20
EugeneKay I haven't caught cancer or had !weasels yet!06:20
gitinfo The consequences of this proposal are not well-defined. A band of furious weasels may infest your undergarments, or it might work just fine. You should !backup then !tryit and let us know what happens.06:20
→ wizbit joined06:21
sitaram EugeneKay: cancer does not become apparent in less than "months" timeframe06:21
EugeneKay "yet"06:21
← kadoban_ left06:22
→ benjo2 joined06:22
← harish left06:22
← rchavik left06:22
→ idoru joined06:22
EugeneKay Do you know if the people who did the previous (g2) audit have gone through g3, by chance?06:24
← edude03 left06:24
→ fsvend joined06:24
→ flijten joined06:24
→ mmattice joined06:24
← h0h0h0 left06:24
→ DuzyZiemniak joined06:24
← benjo2 left06:24
→ benjo2 joined06:24
sitaram not as far as I know... they'll probably do it if kernel.org decides to upgrade because that is what prompted it last time06:25
I could ask I guess...06:25
← Targen left06:25
EugeneKay You write good code, and it looks good to me, but expert opinions are nice06:25
sitaram absolutely06:26
and thanks!06:26
← flijten left06:26
→ flijten joined06:27
← Tabmow left06:27
← _huoxito left06:27
← pantsman- left06:28
→ senny joined06:28
→ pantsman joined06:28
← wizbit left06:28
← redrocket left06:28
← pantsman left06:28
→ pantsman joined06:28
← nutron left06:28
← Araxia left06:29
← flijten left06:29
→ marc_kdab joined06:29
→ Tabmow joined06:29
→ flijten joined06:29
→ wizbit joined06:29
← Synthead left06:30
← DrNick left06:31
→ Araxia joined06:31
→ Synthead joined06:31
→ DrNick joined06:32
← marc_kdab left06:34
fsvend hi everyone, reg. git bisect. i'd like to apply a custom patch in each step.. how can i ignore these changes after each step without confusing bisect?06:34
← mishok13 left06:34
fsvend .. i could branch in each step but unsure how to get back to the bisect selected commit..06:34
is the bisect "state" kept so I can manually jump using checkout?06:35
→ jaseemabid joined06:35
← Orphis left06:35
← banghouse2 left06:36
→ robotmay joined06:36
→ Orphis joined06:36
_ikke_ Can't you just apply the patch without committing?06:36
← acl_ left06:37
← h0bbit_ left06:38
← igotnolegs left06:38
← benjo2 left06:38
← lapistano left06:38
fsvend _ikke_: yes, but i guess git won't let me continue (checking out a new commit) since i've changed the content.06:38
→ lapistano joined06:38
et why wouldn't it?06:38
fsvend _ikke_: if i use git reset.. i guess the bisect session is over?06:38
_ikke_ git checkout -f HEAD06:39
fsvend _ikke_: wouldn't that just get me have to HEAD.. what happened to bisect than?06:40
→ sq-one joined06:40
fsvend _ikke_: does it still track my good/bad commits.. even if i "jump back" to HEAD?06:40
→ tewecske joined06:40
→ ph^ joined06:40
← Orphis left06:41
fsvend maybe i should try before i bother you :)06:41
_ikke_ fsvend: HEAD is the currently checked out commit06:41
→ redrocket joined06:42
← gchristensen left06:42
fsvend _ikke_: yes of course.. man. thanks. so git checkout -f HEAD after running patch would work then.06:42
→ foofoobar joined06:43
_ikke_ It should indeed work06:43
→ Orphis joined06:44
→ Sitethief joined06:45
→ TheHaven joined06:45
fsvend _ikke_: again thanks. I knew HEAD was relative to co commit, but somehow my brain didn't work properly today :)06:46
wizbit /wc06:48
← wizbit left06:48
← Orphis left06:49
→ shigekimorimoto joined06:51
_ikke_ yw06:51
→ Orphis joined06:52
→ tqmz joined06:53
← airborn left06:53
→ dcorbin joined06:55
→ mishok13 joined06:55
← ipalaus_ left06:55
← Orphis left06:56
→ Orphis joined06:59
← batmanian left06:59
← friskd left07:01
← beautifulmind left07:04
← Orphis left07:09
← glcrazy left07:09
→ glcrazy joined07:09
← pppaul left07:09
→ pppaul joined07:10
← Goplat left07:11
→ psoo joined07:12
← dan_p left07:14
→ weiyang joined07:14
← pppaul left07:15
→ pppaul joined07:16
→ cooloney joined07:16
→ Orphis joined07:17
← DuzyZiemniak left07:17
→ Milossh joined07:19
→ thierryp joined07:19
← Lgb left07:20
→ benjo2 joined07:20
← optixx left07:20
→ Lgb joined07:22
→ optixx joined07:23
← Fandekasp left07:23
← mastro left07:24
← foofoobar left07:24
← mastroWork left07:24
→ malesca joined07:26
→ ali1234 joined07:26
→ quuxman joined07:28
← davidbanham left07:28
← Silowyi left07:28
→ Element9 joined07:31
ali1234 i've got a slightly odd workflow and i'm wondering how to handle it "properly"07:31
← diegoviola left07:32
ali1234 i'm managing some php code in git. the server end has a hook that automatically checks out the latest pushed code to the www directory. basically as described here: http://toroid.org/ams/git-website-howto07:32
→ TomyLobo joined07:33
EugeneKay ali1234 - !deploy is(in my biased opinion) a much better solution than that !blog post.07:33
gitinfo ali1234: Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh07:33
ali1234: Blog posts, while helpful and informative, are quite often outdated, give bad advice, or are just plain wrong. Please don't rely solely upon them, or treat them as authoritative.07:33
ali1234 ok, but that still requires two repositories right?07:34
this is my question: if i make a typo in my code and i don't see it until after i have pushed, then, if i rebase and fixup, i have to git push -f07:35
since i'm the only person using the system this isn't a huge problem. but is there a better way?07:36
jast EugeneKay's script deploys into a plain directory, not into a repository07:36
afaics07:36
ali1234 it's a receive based hook07:36
EugeneKay It rsyncs to a $URI07:36
You can abuse --delete and --exclude to great profit here07:37
→ irqq joined07:37
jast that would take care of your concerns, I guess07:37
→ mastro joined07:37
← thursday left07:37
EugeneKay It really doesn't care about force pushes, the only thing that matters to it is that the SHA1 isn't all-zeros07:38
→ BarryCrayon joined07:38
→ mastroWork joined07:38
← optixx left07:38
→ smw_ joined07:38
← robotmay left07:38
EugeneKay There are also some basic sanity checks(like that the URI is a directory). These may not hold true if oyu're rsyncing to a remote URI, in which case you'll also want to do something with -e 'ssh -i id_deploy'07:39
← Textmode left07:39
EugeneKay I'm working on a more complete version which supports this sort of thing which is suitable to be a post-receive hook in a gitolite repo.07:39
jast doesn't look like any remote rsyncing is going to be happening in this case, though :)07:39
← mastroWork left07:40
ali1234 eh, so what're you telling me? there doesn't seem to be any difference between your script and the blog post i linked, except your script uses rsync and the blog post uses git checkout -f07:40
← subbyyy left07:40
EugeneKay checkout -f is error-prone and causes insanity. Quickly.07:40
jast ali1234: "git checkout -f" works in a repository. rsync doesn't need to.07:40
EugeneKay rsync is more flexible, powerful, and has less edge cases.07:40
quuxman for a build script I'm retrieving the commit time of every file in a particular directory of my source tree with two shell commands: rev=$(git rev-list -n 1 HEAD); git log -n 1 --pretty=format:%ct $rev07:40
EugeneKay "You take the code here, and you put it there."07:40
→ mastroWork joined07:40
jast ohh, it does the GIT_WORK_TREE switcharoo07:41
← Synthead left07:41
quuxman however, it's taking several minutes for just 800 files, which is not acceptable. Is there a faster way to do this in bulk?07:41
jast without an index that sounds like heaps of fun and breakage07:41
→ optixx joined07:41
← redrocket left07:41
EugeneKay God forbid you should add a file to tracking that already exists in the destination07:42
Aborted checkout, anybody?07:42
← JackWinter left07:42
ali1234 yeah... none of this actually answers my question though. which is, if i have two repositories and i'm the only user, will push -f ever cause problems?07:42
← ben225 left07:42
jast quuxman: where does that look at every file07:42
ali1234: what do you mean by two repositories... your local one or an extra one on the server?07:42
quuxman jast: that's inside a loop in a python script I have07:42
← mastroWork left07:42
ali1234 basically i just want to remote repository to always be an exact copy of my local repository no matter what i do to it07:42
EugeneKay "No, not really."07:43
Bigcheese As long as nobody has a branch based off the branch you are modifying history of, you are good.07:43
ali1234 jast: i mean quite simply two repositories... one on my local computer and another one that i push into over ssh07:43
jast and, anyway, the answer you've already received is that it's very hard to predict with "git checkout -f"07:43
it may work, or it may break horribly07:43
→ mastroWork joined07:43
ali1234 git checkout -f has nothing to do with it, please stop going on about it07:43
quuxman I'm sure this is a common problem. The build script distributes static files to a CDN, and I only want to upload if there's a new version. I also want to use the committed time as a cache busting query parameter07:43
jast oh, you're not talking about the checked out files at all, then?07:43
ali1234 no07:44
EugeneKay ali1234 - you should be OK with push -f'ing, as far as the remote bare repo is concerned.07:44
jast then why were you talking about the deployment stuff in the first place?07:44
it has nothing to do with the commit history07:44
ali1234 because if i hadn't, i'd get lots of messages like "that's stupid why you want to do that" and then i'd have to explain, then have the whole convo about why checkout -f sucks... and still not get an answer07:44
jast ali1234: wrong07:44
we talk to lots of people who have legitimate reasons to force push07:45
→ JackWinter joined07:45
jast we do it ourselves, too07:45
_ikke_ checkout -f != push -f07:45
jast yes :)07:45
we established that already07:45
← Orphis left07:45
← Araxia left07:45
← Slix` left07:46
→ Araxia joined07:46
← CellKill_ left07:46
EugeneKay It still makes you a bad, terrible person. Think of all the trauma you're putting that poor repo through! Ripping away everything it knows and shoving strange things into it.07:47
_ikke_ antropomorpisms ftw07:47
→ redrocket joined07:47
_ikke_ antropomorphisms*07:47
ali1234 well it beats having several none-working commits every time i have to find typos07:47
jast quuxman: I don't think there's a convenient command to do that... but it's very likely to be much faster if you parse the output of something like: git log --name-status --pretty=format:%ct07:48
EugeneKay It's OK to do a "typo fix" commit. Consider also !sausage07:48
gitinfo [!sausage_making] Some developers like to hide the sausage making (pretend to the outside world that their commits sprung full-formed in utter perfection into their git repository). `git rebase -i`, `git add -p`, and `git reset -p` can fix commits up in post-production by splitting different concepts, merging fixes to older commits, etc. See also http://sethrobertson.github.com/GitBestPractices/#sausage07:48
← pantsman left07:48
→ dpino joined07:49
ali1234 it's not OK to do a typo fix commit imo07:49
→ Orphis joined07:49
jast it depends07:49
ali1234 it's certainly not ok to do 10 in a row07:49
quuxman jast: is there a way to get the filename and ctime on the same line at least?07:49
jast sometimes you have no other option07:49
EugeneKay Well, that's jsut like, your opinion, man. :v07:49
ali1234 as i initially said, i can't test the code without pushing it07:49
frogonwheels ali1234: as long as it's not pushed, I think it's quite fine07:49
jast maybe you should consider a different deploy option, then07:50
_ikke_ ali1234: That's awefully bad07:50
et if you just want an exact copy, why not use rsync?07:50
jast one that doesn't require you to commit stuff before deploying it07:50
EugeneKay Or a "live" dev environment07:50
_ikke_ Git is not a deployment tool07:50
jast quuxman: can't think of one07:50
frogonwheels ali1234: What I mean is, I like sausage_making - don't push until it's ready, and squish typeo/bugfix commits into the appropriate commits as I go07:50
jast quuxman: but I think that output should be pretty easy to parse with something like python07:51
frogonwheels ali1234: I prefer not to push typo commits.07:51
→ arti_t joined07:51
EugeneKay quuxman - my !deploy script does some magic with git-log and 'touch'. Look at that, maybe?07:51
gitinfo quuxman: Git is not a deployment tool, but you can build one around it for simple environments. Here is an example hook to get you started: https://github.com/EugeneKay/scripts/blob/master/bash/git-deploy-hook.sh07:51
EugeneKay It's inefficient, but filesystem cache keeps it pretty fast.07:51
jast unless you run it on windows, right :}07:51
EugeneKay quuxman - you might also want to use libgit2's python bindings instead07:51
Using windows as a deploy server is.... well.... I hope you believe in a forgiving God.07:52
quuxman I've heard that from a lot of sources, that "git is not a deployment / build tool" and that you should separate your git repository and your live code. However, I don't see the rational for that. In the project I run, we have git remotes for each running live server, and a simple post-commit hook that does some basic stuff and restarts the web server. It's very convenient07:53
→ ben225 joined07:53
jast that kind of thing works in the simplest cases07:53
but usually you have to cram a lot of logic into hooks like that, and that makes the hook the deploy tool and not git :)07:53
EugeneKay 'git' is an object model, at it's core. Your deployment tool should use git to get it's data, but all the scripting is up to you.07:54
→ shruggar joined07:54
→ ipalaus_ joined07:54
quuxman EugeneKay: ah ok. That's exactly what we're up to07:54
→ pandora17 joined07:54
→ Synthead joined07:54
jast for example, if the live environment is structured differently than the source tree (and often it should be)07:54
_ikke_ Deployment is often more than just update files to the latest version07:54
→ thiago joined07:54
jast or if files need to be pre-processed07:54
→ dark_element joined07:55
EugeneKay Those are jobs for Continuous Integration packages07:55
_ikke_ Hmm, well, the CI itself doesn't do that, it only initiates it07:56
→ d3c joined07:57
gitinfo set mode: +v07:57
benjo2 → bwald07:57
← borwar left07:57
→ harish joined07:58
quuxman EugeneKay: is that just resetting the mtimes based on git's author time?07:58
EugeneKay quuxman - exactly so.07:58
quuxman EugeneKay: so that's similar enough to my code that I suspect it also takes a good chunk of time to run07:58
EugeneKay With a hot filesystem cache(which you have, with packfiles and right after doing a git-archive), I've never seen it take >1s07:59
← d3c left07:59
→ diegoviola joined08:00
→ _Qman joined08:00
→ d3c joined08:00
jast quuxman: hard to tell the difference, since the piece of code you shared with us doesn't ever reference any filenames in any way08:00
← teotwaki left08:01
jast unless, of course, the difference is that you're really short on RAM :)08:01
quuxman EugeneKay: how would I come by $branch ?08:01
← yizumi left08:01
→ whitman joined08:01
EugeneKay quuxman - in my script? It's read from stdin, see man githooks for how to call a post-receive08:01
gitinfo quuxman: the githooks manpage is available at http://jk.gs/githooks.html08:01
jast read the script :)08:01
quuxman EugeneKay: in this case I'm not running it from a githook08:02
jast right, in that case just whatever branch you wanted to look at08:02
EugeneKay $branch just needs to rev-parse to the commit which you git-archive'd08:03
→ yizumi joined08:03
EugeneKay I'm doing some fancy logic to parse the config options, but you don't care about any of that junk08:03
→ francisfish joined08:04
← ipalaus_ left08:05
quuxman I'm just going to hard code 'master'08:05
although that's probably a bad idea now that I think about it08:05
jast I suppose HEAD might work for you08:05
EugeneKay I'd go with HEAD08:05
→ Lord_DeathMatch joined08:05
quuxman Yeah, definitely better than master ;)08:05
jast at least as a default08:05
→ zizzfizzix joined08:05
→ marcogmonteiro joined08:06
← Rafusek left08:06
→ Rafusek joined08:06
→ giallu joined08:07
← cyphase left08:08
→ esonyun joined08:08
← pppaul left08:08
→ pppaul joined08:09
← Mowsh left08:09
→ LimeBlast joined08:09
→ adhawkins joined08:10
← kadoban left08:10
← harshpb left08:10
→ OOPMan joined08:10
→ robotmay joined08:11
→ harshpb joined08:11
quuxman EugeneKay: how many files does that loop through, when it only takes around a second?08:12
EugeneKay `find . -type f | wc -l`: 170808:13
quuxman blarg. What's wrong with this? I copied your line exactly, so now I'm just using one shell command, and it's still taking over a minute for 840 files08:13
← jepler left08:13
EugeneKay That box has a SSD and a bitching-fast CPU.08:14
→ teotwaki joined08:14
→ achadwick joined08:14
EugeneKay There's also <100 commits in that repo, which could be a factor.08:15
quuxman that would explain that08:16
EugeneKay Remind me of the exact problem again08:17
There's gotta be ab etter way to approach it08:17
← sq-one left08:17
← bwidmer left08:18
→ blacktulip joined08:18
→ bwidmer joined08:19
quuxman just trying to get commit times out of all the files in a semi-large directory08:19
EugeneKay But you only need the times for files that have changed since the last time you did this, right?08:19
quuxman $(git log --name-status --pretty=format:%ct) is basically it, I just have to parse it08:20
EugeneKay: true... but that doesn't really help08:20
← Milossh left08:20
← esonyun left08:20
EugeneKay It might. You can lessen the number of files you have to check by parsing --stat08:20
← bwald left08:20
← gasbakid left08:21
← preflex left08:21
← sporous left08:21
→ kermit joined08:21
→ HyperNerdV2_ joined08:21
quuxman I'm also looking into using libgit208:21
EugeneKay Are you doing all this with a non-bare repo?08:22
→ irenicus09 joined08:22
quuxman yeah, there's a working directory, but it's never modified08:22
→ ipalaus_ joined08:22
→ gasbakid__ joined08:22
← dreiss left08:23
← irenicus09 left08:23
EugeneKay git pull --stat | some sed magic | (while read filename; do $(git log HEAD -1 --- ${filename}); done)08:23
(and all the rest of the datestamp magic)08:23
← HyperNerdV2 left08:24
→ sporous joined08:24
→ preflex joined08:24
← sporous left08:25
← HyperNerdV2_ left08:25
→ sporous joined08:25
← Vile left08:26
← mastroWork left08:27
← sporous left08:27
→ sporous joined08:27
← mastro left08:28
→ tcurdt joined08:28
→ alexukf joined08:29
← ben225 left08:33
← pppaul left08:34
→ Fandekasp joined08:34
→ pppaul joined08:35
→ jacobrask joined08:36
jacobrask Hi, is subtree included in any public git release yet?08:37
EugeneKay 1.7.10, IIRC08:37
← tcurdt left08:37
jacobrask cool. the latest version in ubuntu backports seem to be 1.7.9.5.. are there any deb/ubuntu repos with newer versions or do I need to compile it?08:39
← pppaul left08:39
→ mastroWork joined08:40
→ mastro joined08:40
← d3c left08:40
← gavin_huang left08:40
→ solofight joined08:40
→ pppaul joined08:40
← arti_t left08:40
solofight people, what do i do if i want to search for a specific word or sentence across several hundreds of log messages ?08:41
EugeneKay man grep08:42
← pppaul left08:42
→ pppaul joined08:42
solofight EugeneKay: no, grep can only give the line which contains the word i am searching for. But i want the associated details like commit id, files changed ...etc08:42
→ flype joined08:42
→ tatsuya_o joined08:43
EugeneKay solofight - yes it can. Look at -A/-B/-C08:43
FauxFaux jacobrask: The ubuntu-unstable version will probably install.08:44
solofight EugeneKay: oh, will look at it now08:44
EugeneKay solofight - if you want it programmatically(for a script) it gets tougher, but doable.08:44
→ jepler joined08:46
← pppaul left08:47
solofight EugeneKay: thank you. Grep -A and -B and -C definitely fit in. I am worried about dynamically detecting the multi line comment and lines after, before the end,start of the comment08:47
any automatic ways to detect them ?08:47
EugeneKay Write a script08:47
cbreak-work solofight: man git-log08:48
gitinfo solofight: the git-log manpage is available at http://jk.gs/git-log.html08:48
→ Orbitrix joined08:48
→ pppaul joined08:48
← huevos left08:48
cbreak-work it has grep options08:48
(in particular --grep)08:48
solofight wow cbreak-work you are awesome08:49
thank you - exactly what i was looking for08:49
cbreak-work: ++08:49
EugeneKay: ++08:49
← _Qman left08:50
EugeneKay ZZ08:50
→ jmd joined08:53
jmd I have a couple of directories, in which if I do "git log <fielname>" I see absolutely nothing. What gives?08:54
EugeneKay !dashes08:54
gitinfo Use a double-dash(--) to separate refs from paths, especially when dealing with ambiguous file names. Ex: git checkout master -- origin (check out the file "origin" from branch "master")08:54
jmd Also "git log -- <filename>" shows nothing.08:55
FauxFaux jmd: "git ls-files ."08:56
jmd Also nothing.08:56
← harshpb left08:57
cbreak-work jmd: maybe no file in those dirs is tracked08:58
← ipalaus_ left08:58
jmd cbreak-work: But I just committed a change and "git log HEAD" shows they were succesfully changed.08:58
→ Vile joined08:59
_ikke_ jmd: Are you only giving the filename, or the whole path?08:59
jmd just the filename08:59
_ikke_ try the relative path08:59
cbreak-work jmd: pastebin: git log --stat -1 and git log -- filename09:00
← ilius left09:00
cbreak-work all filenames are relative to your current location09:00
jmd -109:00
→ andrewh joined09:00
← kcm1700 left09:01
→ kcm1700 joined09:01
→ srinivas07 joined09:01
← glcrazy left09:03
→ glcrazy joined09:03
← s0ber left09:04
→ berserkr joined09:06
← jaseemabid left09:08
→ h0bbit_ joined09:09
→ Sigma[Mtp] joined09:09
← pppaul left09:10
→ pppaul joined09:11
← Sigma left09:11
→ whgkl joined09:12
Sigma[Mtp] → Sigma09:12
← TomyLobo left09:13
gitinfo set mode: +v09:13
whgkl Hi, is there a way to enable verbose when using git difftool? The mingw32 environment doesn't start my kdiff3 executable...09:13
FauxFaux GIT_TRACE=1 git difftool # maybe.09:13
solofight People, all log related functions we have in Git can be achieved with git log itself ? Ex: git shortlog is nothing but git log --pretty right ?09:14
← srinivas07 left09:16
FauxFaux solofight: Looks very different to me. If your question is "can we build everything on top of log", then why not build everything on top of rev-list or something even lower?09:16
→ f0i joined09:16
→ xiangfu joined09:16
solofight FauxFaux: oh09:17
→ mxweas_ joined09:17
← jacobrask left09:17
_ikke_ FauxFaux: Well, git shortlog is imo just a shortcut09:18
FauxFaux My guess would be a post-processor. To the source!09:19
→ mmc1 joined09:19
→ huevos joined09:19
→ stodan joined09:19
whgkl After "Hit return to launch 'kdiff3':" But nothing happens after hitting return... I only get back to prompt.09:20
FauxFaux Bah, it's not.09:20
et is kdiff3 installed?09:20
whgkl et yap09:20
→ Vert joined09:20
whgkl difftool.kdiff3.cmd=D:/dev/bin/kdiff3/kdiff3.exe $LOCAL $REMOTE09:21
_ikke_ FauxFaux: What is it then?09:21
← Orbitrix left09:21
solofight feels KVIrc is a good replacement for mirc, Xchat09:21
← cyb3r3li0g left09:21
← warthog9 left09:22
← thiago left09:22
→ cyb3r3li0g joined09:23
shruggar what's the fastest / easier way to create a tree-object from an arbitrary directory?09:23
whgkl when I execute D:/... from propmt kdiff starts...09:24
→ cortana2 joined09:24
← stodan left09:25
→ lusers joined09:25
_ikke_ shruggar: Why do you need that?09:25
FauxFaux shruggar: Commit it.09:25
→ stodan joined09:25
jast or put stuff in the index and use write-tree09:25
shruggar _ikke_: snapshotting a directory so that I can easily spot differences between what someone has actually deployed and what they're supposed to have deployed09:26
(also basing some temporary branches off of the "actually deployed", for now)09:26
FauxFaux It's a shame hash-object -w doesn't take a directory.09:27
_ikke_ can't you just copy that dir onto the working tree?09:28
← jackh left09:28
→ ipalaus_ joined09:28
→ d3c joined09:29
shruggar I have solved this sort of thing in the past by setting GIT_INDEX_FILE and GIT_WORK_TREE09:29
→ marthinal joined09:30
← pandora17 left09:31
← pppaul left09:31
→ arti_t joined09:32
→ pppaul joined09:32
→ davidbanham joined09:32
→ srinivas07 joined09:32
→ warthog9 joined09:35
← lusers left09:35
→ jaseemabid joined09:37
← ronny left09:38
→ peda_uni joined09:38
gitinfo set mode: +v09:41
peda_uni hi. I have two svn repos that contain the same project but both are visible to a different set of developers and I have to keep them in sync. However, the directory structures in the repos are quite different09:41
is there a way to handle that with git in an easy way? I do not want to diff/patch/copy everything manually09:41
FauxFaux You can have two git-svn "remotes" in one repo, and rebase/dcommit changes from one onto the other, but it's going to cause tears.09:42
→ mischat joined09:42
← ipalaus_ left09:42
← pppaul left09:42
peda_uni FauxFaux, yeah, that's not the problem. My problem is, that the files I want to sync are in different directories09:43
e.g. in one repo it's called "src/foo.c" in the other it's "sources/foo.c"09:43
FauxFaux So just git-svn those seperate directories?09:43
→ pppaul joined09:43
→ noisebleed joined09:44
← noisebleed left09:44
→ noisebleed joined09:44
peda_uni hm09:44
← juvenal left09:44
peda_uni difficult too, because I have build-files (like cmake, make,...) in the top-level dir, and I have to maintain them, too09:45
← sangi left09:46
← _sundar_ left09:46
→ sangi joined09:46
← neurodrone left09:46
← pppaul left09:48
← diegoviola left09:49
→ pppaul joined09:49
shruggar why are these two trees so different? If they are so different, what about them actually needs to be "sync"ed?09:50
syncing directories usually means ensuring they have the same structure, too09:50
← weiyang left09:51
← glcrazy left09:51
→ benjo2 joined09:52
peda_uni one repository is a win project, the other's a linux project and they have much functionality in common (i.e. many files are/should be the same)09:52
but build-files and some startup/gui-code is different09:53
← d3vic3 left09:53
peda_uni and I'd like to get commits done to common files in one of the repos also in the other repo09:54
→ d3vic3 joined09:54
← twb left09:54
→ ToBeFree joined09:55
FauxFaux Just fix the project to have some kind of non-terrible layout. /o\09:55
peda_uni :)09:57
→ icq joined09:57
icq hey guys09:57
I am trying to migrate an svn repository to git but I am encountering some problems09:58
→ _sundar_ joined09:58
icq first I have trunk/tags/branches/other-branch-folder so I do -T trunk -b branches -b other-branch-folder -t tags09:58
← localhost left09:58
→ thiago joined09:58
icq the problem is that in other-branch-folder there is a README file09:58
how is git going to deal with that file?09:59
will it skip it?09:59
← PrinceAMD left09:59
← n3m left09:59
→ localhost joined10:01
cbreak-work icq: a single file branch?10:03
← machty left10:04
← chaowang left10:06
→ arietis joined10:07
→ ipalaus_ joined10:08
← pppaul left10:10
→ pppaul joined10:11
→ PrinceAMD joined10:11
→ DoniRS joined10:12
← xiangfu left10:12
← andrewh left10:12
← ipalaus_ left10:14
→ andrewh joined10:15
← gasbakid__ left10:15
→ QaDeS joined10:15
← mxweas_ left10:18
→ gasbakid__ joined10:18
← fuzzybunny69y left10:19
→ ipalaus_ joined10:20
← berserkr left10:20
→ berserkr joined10:22
← jaseemabid left10:24
← mischat left10:24
→ mischat joined10:24
→ pirx joined10:25
← jceb left10:25
pirx hello! i have a gitosis-admin repository on a server. my key is not added there, and all people with admin rights have quit the company. i have full access to the server, and need to clean up among the keys and such. any pointers on how to do this?10:26
→ jkal joined10:26
← stliu left10:26
_ikke_ !gitosis10:26
gitinfo gitosis is no longer maintained and supported by the author; we usually recommend gitolite instead which has much better documentation and more features: http://github.com/sitaramc/gitolite10:26
pirx yes, that does not help me much with gitosis:)10:27
new repos use gitolite, but this one still remains10:27
FauxFaux pirx: Probably need to manually fiddle ~gitosis/.ssh/authorised_keys to get it to allow you to push gitosis-admin once, as if you were someone else.10:27
sitaram pirx: gitosis has no "write access control" so you can just clone the repo directly on the server, add your key, and push; it should work10:28
(I'm guessing here)10:29
← angelsl left10:29
pirx i grepped the home dir of gitosis for the admin keys, and they are added in a few places, like ~/repositories/gitosis-admin.git/gitosis.conf and binary match on ~/repositories/gitosis-admin.git/index10:29
but i am not sure i want to edit those files directly in a production server10:29
sitaram: aha, will try that10:29
sitaram pirx: log on to server, git clone repositories/gitosis-admin.git ga, cd ga, <change keys, git add, commit, push>10:30
→ GreekFreak joined10:30
sitaram again, I'm guessing here; it's been ages since I looked at it anyway10:31
MacGyverNL As an aside, if you're fiddling anyway, why not migrate the old repo to gitolite?10:31
← f0i left10:31
sitaram pirx: the reason you should do what MacGyverNL said is that there really isn't any support for gitosis. No one really knows, no one cares10:31
→ wean joined10:32
sitaram but I'm the author of gitolite so I'm clearly biased (and that is why I waited till someone else said it ;-)10:32
jkal hi, i'm trying to do something like this: http://dpaste.de/YOxwu/10:33
any insight on what is the best strategy to do it?10:34
shruggar I just thought it was funny you were giving gitosis advice, since clearly the last time you needed to think about it, it must have ended with "screw this! I'll write my own!"10:34
GreekFreak Hi. I am trying to create a repo on my flashdrive so I can push to that instead of github. I have followed these instructions by someone in this channel, but it fails at the push: (1) $ cd /media/Platform_Code && mkdir admin_base.git && cd admin_base.git && git init --bare (2) I went to my repo (my project folder not the .git folder in that) and: $ git remote add flashdrive /media/Platform_Code && git push flashdrive --all. This all10:34
fails on the push with the error 'fatal: '/media/Platform_Code/' does not appear to be a git repository'10:34
→ mabrand joined10:34
→ cthuluh_ joined10:35
← scientes left10:35
→ linusoleander joined10:35
shruggar GreekFreak: use "git bundle" for sneakernet-based transfers10:35
GreekFreak shruggar, sorry, I'm new to all this. what is a sneakernet-based transfer?10:36
shruggar as in "copy files to an external drive, walk them over (ie, in your sneakers) to the other computer, plug in drive"10:37
→ marhaban joined10:38
gitinfo set mode: +v10:38
benjo2 GreekFreak: git remote add flash drive /media/Platform_Code/admin_base.git10:38
you're missing the repo when you do the remote add10:38
GreekFreak shruggar, lol10:39
benjo2, trying it now10:39
shruggar it's a very common term. I suppose it's gotten less common the more ubiquitous networking has become10:39
_ikke_ https://en.wikipedia.org/wiki/Sneakernet10:40
MacGyverNL shruggar: People are also usually baffled if I use the term "air gap firewall".10:40
GreekFreak benjo2, thanks, that did it10:40
→ wbednarski joined10:41
shruggar heh, as if an air gap could stop anything these days :)10:41
GreekFreak shruggar, you forget how many things are taken for granted these days10:41
← pppaul left10:41
→ xiangfu joined10:41
FauxFaux I bet someone, somewhere, is using wireless because they heard that an air-gap was important.10:41
_ikke_ :D10:41
← robbyoconnor left10:42
MacGyverNL shruggar: Well, for one thing, SCADA-systems should be air-gapped considering the horrendously inadequate security measures on them.10:42
→ pppaul joined10:42
MacGyverNL It doesn't make it impossible to get to them...10:42
Just... Harder.10:42
But this is not about git, so let's leave it there.10:42
← flype left10:42
cthuluh_ → cthuluh10:43
← linusoleander left10:43
shruggar WiFi into a trusted network, trusted because it has an air-gap. Plug in my phone to charge it, laptop detects the network is marked as trusted and shares the connection. Security! :D10:43
← wean left10:45
→ gavin_huang joined10:45
← benjo2 left10:46
→ vals_ joined10:46
← tango_ left10:47
← pppaul left10:47
→ pppaul joined10:47
vals_ → tango_10:48
→ Ryback_ joined10:48
→ n2diy joined10:50
← GreekFreak left10:50
← thiago left10:51
→ amcsi_work joined10:52
→ gchristensen joined10:52
← ali1234 left10:55
→ Athlon^1 joined10:56
→ babilen joined10:58
b00b00 hello10:58
← pppaul left10:58
babilen → Guest9645810:58
→ pppaul joined10:59
→ linusoleander joined10:59
← linusoleander left10:59
shruggar s/\d*$//; giggle();10:59
→ foofoobar joined10:59
shruggar !hi10:59
gitinfo [!welcome] Welcome to #git, a place full of helpful gits. If you have a question, feel free to just go ahead and ask—somebody should answer shortly. For more info on this channel, see http://jk.gs/git/ Take backups (type !backup to learn how) before taking advice.10:59
b00b00 while trying clonning svn to git, i have this error: "fatal: Not a valid object name refs/remotes/tags/17072008-Private Beta Improvements" and "cat-file commit refs/remotes/tags/17072008-Private Beta Improvements: command returned error: 128"10:59
← w09x left10:59
→ s0ber joined11:00
← torarne left11:00
b00b00 any idea what to do here for continuing with clonning?11:02
→ oehman joined11:03
→ hendry joined11:03
ivan maybe you can delete all mentions of the tag from .git/11:03
← xiangfu left11:03
b00b00 ivan: can you assist me how?11:03
shruggar I don't know if clone is resumable11:03
ivan git svn fetch/rebase is resumable11:04
b00b00 shruggar: loosk like it is11:04
ivan b00b00: grep and an editor11:04
← ipalaus_ left11:04
← zdk left11:05
→ killerstorm joined11:06
b00b00 ivan, bit new with this, can be more specific?11:07
ivan grep 17072008-Private -r .git11:08
← Guest96458 left11:08
b00b00 ok, thanks11:09
→ babilen_ joined11:09
killerstorm hi. I've tried process of handling of remote dependencies described here: http://git-scm.com/book/ch6-7.html11:09
babilen_ → Guest1586311:09
killerstorm But I don't get how to share such remotes, i.e. enable developers who clone repo to be able to update libraries when they want to.11:09
Is there some way to achieve identical setup automatically?11:10
b00b00 ivan: what i supposed to do with the list of a few files found? delete them?11:10
← Guest15863 left11:11
→ Guest15863 joined11:11
ivan nope, just the lines with 17072008-Private11:11
Guest15863 → babilen11:11
ivan (it may still not work after you do this, I don't know which phase of git/git-svn is breaking)11:11
b00b00 ivan: yes, meant delete lines, thanks11:11
→ Constant_ joined11:11
← berserkr left11:12
→ CannedCorn joined11:12
← Constant_ left11:13
← shruggar left11:13
→ shruggar1 joined11:13
← _iron left11:13
→ thiago joined11:13
→ glennpratt joined11:13
← dr_lepper left11:13
→ juvenal joined11:14
→ berserkr joined11:14
→ rtuin joined11:15
→ dr_lepper joined11:16
← Orphis left11:17
← stodan left11:18
→ enmand joined11:18
← whgkl left11:19
→ Davey joined11:20
← OOPMan left11:22
→ stodan joined11:23
← Eryn_1983_FL left11:23
sangi running git repack throws an error like this "fatal: Out of memory, malloc failed"11:23
thiago how much RAM Do you have available?11:23
sangi its a 1GB RAM11:24
← jkal left11:24
← CannedCorn left11:24
sangi thiago, sorry I have 512MB of RAM11:24
MacGyverNL That's not much.11:26
→ Peste_Bubonica joined11:27
thiago sangi: how big is your repository?11:27
→ w09x joined11:27
sangi thiago, its more than 2GB and have a single image of size 900MB11:28
MacGyverNL You could try fiddling with deltaCacheLimit, deltaCacheSize and windowMemory.11:28
← arti_t left11:28
MacGyverNL Though if there's a file larger than the available memory, I'm unsure if repack can ever succeed.11:29
→ arti_t joined11:29
← Tohuw left11:29
→ denysonique joined11:29
← papegaaij left11:30
→ beautifulmind joined11:30
sangi MacGyverNL, I have tried increasing window memory and max_pack_size11:31
→ Tinti joined11:31
→ skylamer` joined11:31
sangi but the same error only11:32
← yizumi left11:32
sitaram MacGyverNL: I think the limit is RAM+swap not just RAM but of course the amount of thrashing might crash the disk on a machine old enough to have only 512 MB :)11:34
thiago sangi: you need at least as much RAM as your biggest file.11:34
← pppaul left11:34
→ fuzzybunny69y joined11:34
sitaram thiago: not (technically) RAM+swap?11:34
sitaram may be wrong; wants to know...11:34
thiago I don't think so11:34
sitaram hmm11:34
thiago I don't know the code for sure either11:34
RAM+swap would indicate that it's dirty memory11:35
sangi thiago, ok11:35
thiago i.e., the output of the packer11:35
since I know that files are mmap()ed to memory, they're clean11:35
→ pppaul joined11:35
sitaram and it'll thrash the disk but *eventually* it should complete, no?11:35
thiago then again, when repacking an already-packed file, it needs to expand the contents, so it needs dirty memory11:35
→ ripper joined11:36
MacGyverNL You should decrease windowMemory, not increase it.11:36
thiago and all files in .git/objects are already zlib-compressed at least11:36
← andrewh left11:36
MacGyverNL Also, it could be that he's running a 32-bit build and hitting a memory space limit.11:36
→ Tohuw joined11:36
thiago MacGyverNL: that's 3 GB on Linux11:36
MacGyverNL (My main annoyance with msysgit, as a matter of fact.)11:36
sitaram thiago: I'm not even thinking at that level. When does malloc() fail? When RAM+swap is gone low, not just RAM, right? (please God don't let me be wrong about something so basic!)11:36
thiago But, in any case, running a program that requires working set size larger than RAM is a bad idea11:36
it'll thrashs11:36
← ripper left11:37
thiago sitaram: malloc() fails when the allocation request fails. Depending on some options, it might be brk() or mmap() (or both).11:37
that means the kernel decided it could not provide more virtual memory11:38
→ jceb joined11:38
thiago which it does only when it has exhausted its page supply. That's RAM+swap.11:38
→ stuntmachine joined11:38
→ benjo2 joined11:38
← teotwaki left11:38
thiago but note that we assumed that the error message comes from malloc() returning NULL. It's possible that the error message comes from repack trying mmap() and that returned NULL too.11:39
MacGyverNL And then there's still the possibility of overcommit, further complicating matters a bit.11:39
thiago overcommitting doesn't cause that error11:39
sitaram I guess... anyway even if it works the thrashing would be horrendous11:39
thiago it causes the program to exit with SIGSEGV or to get OOM killed11:39
→ letharion_ joined11:39
thiago or SIGBUS, depending on the platform11:39
indeed11:40
gitinfo set mode: +v11:40
← oehman left11:40
thiago so my recommendation is: for large repositories, use a 64-bit machine with a few GB of RAM.11:41
→ solussd_ joined11:41
letharion_ → letharion11:41
← arti_t left11:41
→ arti_t joined11:42
→ oehman joined11:43
FauxFaux sitaram: On x32, you can get heap fragmentation such that large (implicitly contiguous) allocations will fail, even though the actual amount of free everything is quite high. Happens in practice in other areas, i.e. very unlikely to get over 1gb contiguous (for the JVM ¬_¬) on x32 Windows as it poops processes everywhere.11:44
thiago FauxFaux: by x32, do you really mean x32? Or did you mean x86?11:44
sitaram jvm+windows = poop anyway :)11:44
FauxFaux x86_32, yeah.11:44
thiago FauxFaux: x32 = 32-bit x86-6411:44
← dark_element left11:44
thiago FauxFaux: i.e., 32-bit code running in long mode11:44
FauxFaux Really? Probably ambigious enough that it shouldn't be used anyway.11:45
MacGyverNL Doesn't that also happen on x86-64?11:45
thiago though it's quite possible that x32 suffers from the same problem11:45
MacGyverNL (The heap fragmenting)11:45
FauxFaux MacGyverNL: Sure, but the chance of it being a problem is vanishingly small for any relatively sensible allocation.11:45
→ mtphys joined11:45
thiago MacGyverNL: heap fragmentation happens. But you can surely find a 1 GB contiguous virtual space in the 47-bit region available for the program.11:45
← juvenal left11:46
→ Wooga joined11:46
MacGyverNL Ah, of course.11:46
thiago on x86-64 with 64-bit pointers, 1 GB is 0.000762939453125% of the possible addressing space11:47
which isn't 64-bit wide: it's actually 47-bit wide on current machines.11:47
← n2diy left11:47
MacGyverNL But what I meant was, it's not like heap fragmenting is truly an architecture artefact of 32-bit x86.11:48
→ ilius joined11:48
thiago on x86, it would be 1/3rd; on x32, it would be 1/4.11:48
← ilius left11:48
→ ilius joined11:48
FauxFaux MacGyverNL: True. I was focussing on the "malloc failing" case.11:48
→ andrewh joined11:48
→ juvenal joined11:48
thiago heap fragmentation happens all the time, but that's usually not a problem. The problem is when you can't allocate because there's no contiguous block big enough.11:49
heap fragmentation becomes a problem of itself in long-running applications, because they tend to keep a higher RSS than they actually need (they can't return memory to the OS)11:49
← oehman left11:49
FauxFaux Luckily, the JVM fixes all this. /me runs.11:51
→ bowarv joined11:51
← ilius left11:53
← Tohuw left11:53
← gasbakid__ left11:53
→ gasbakid joined11:53
← dv310p3r left11:54
← benjo2 left11:54
← cooloney left11:55
→ sgncfl joined11:55
← pppaul left11:56
← solofight left11:56
→ Tohuw joined11:56
→ pppaul joined11:56
→ angelsl joined11:59
← pppaul left11:59
→ pppaul joined12:00
← juvenal left12:01
→ flype joined12:01
→ dark_element joined12:02
→ juvenal joined12:02
← skylamer` left12:02
← peda_uni left12:04
→ dbolser joined12:04
→ cannonball joined12:04
← pppaul left12:06
→ pppaul joined12:06
→ Milossh joined12:08
← mabrand left12:09
← btree__ left12:10
← pppaul left12:11
→ iamjarvo joined12:11
→ pppaul joined12:11
← Milossh left12:11
→ Milossh joined12:12
→ Kunda joined12:12
→ ipalaus_ joined12:12
← marhaban left12:13
← ehsan left12:14
← sgncfl left12:15
→ freeed joined12:15
→ OOPMan joined12:17
← jmd left12:18
→ _Qman joined12:19
→ johnanderson joined12:20
← johnanderson left12:20
→ papegaaij joined12:20
← stuntmachine left12:20
← QaDeS left12:21
→ aboudreault joined12:22
aboudreault Hi. Can I do a git diff of the files of "Changes to be committed:" section?12:23
FauxFaux git diff --cached12:23
→ ronniy joined12:23
aboudreault thans12:23
*ks.12:23
→ erichynds joined12:24
← shruggar1 left12:28
→ proxinix joined12:30
← proxinix left12:30
→ marhaban joined12:30
→ AaronMT joined12:31
→ yoklov joined12:32
→ ISF joined12:33
← andrewh left12:33
← sangi left12:33
→ shruggar joined12:36
← mtphys left12:36
← wolftankk left12:38
→ CannedCorn joined12:39
← marhaban left12:40
← kaen left12:42
→ d0k joined12:42
← w09x left12:42
→ w09x joined12:42
→ duckxx joined12:42
← arti_t left12:43
← w09x left12:43
→ w09x_ joined12:43
→ dailylinux joined12:44
dailylinux hi12:44
gitinfo dailylinux: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.12:44
dailylinux can i get some help to pull a remote branch i've added to another project?12:45
← w09x_ left12:45
→ w09x joined12:45
_ikke_ dailylinux: What is your problem?12:45
dailylinux i did a git remote add files_odfviewer git://gitorious.org/my-owncloud-apps/files_odfviewer.git12:45
and then git fetch files_odfviewer12:46
this doesn't seem to be a correct way of doing it12:46
← w09x left12:46
dailylinux my config file http://fpaste.org/ewkz/12:46
→ w09x joined12:47
_ikke_ dailylinux: The only step you are still missing is a merge or a rebase (or creating a local branch from the remote branch)12:49
→ raz joined12:49
← solussd_ left12:49
raz gnah12:49
_ikke_ depending on what you want12:49
← duckxx left12:49
→ solussd_ joined12:49
raz so once again i 'git add'ed a directory in the past that itself contained a .git12:49
it's been committed a couple times, and checks out as an empty directory12:50
dailylinux _ikke_, hmm. i want to merge12:50
raz i removed the .git in there now, but how do i re-add it as regular directory? git status shows nothing, even after git adding it again12:50
dailylinux _ikke_, not gonna edit and change, just test the codes12:50
→ w09x_ joined12:51
← w09x left12:51
← fuzzybunny69y left12:51
_ikke_ dailylinux: What branch are you talking about in particular12:51
→ Bass10 joined12:51
← Kunda left12:51
dailylinux [remote "files_odfviewer"] _ikke_12:51
→ shruggar1 joined12:52
→ w09x joined12:52
← w09x_ left12:52
_ikke_ dailylinux: That's a remote, not a branch12:52
dailylinux: git branch -r12:52
raz ok, apparently the offending directory is checked in as "Subproject commit <hex>" (according to git log -p)12:52
← w09x left12:53
raz curses the submodule trainwreck yet again12:53
← shruggar left12:53
→ tommyvyo joined12:53
dailylinux _ikke_, i need to clone it or merge it locally12:54
→ w09x joined12:54
_ikke_ raz: subproject != submodule12:54
← solussd_ left12:54
_ikke_ Why do you think you want to merge it?12:54
raz _ikke_: git behaves absolutely brain damaged when you run "git add; git commit" on a directory that contains a .git12:54
← killerstorm left12:54
→ chffelix joined12:54
_ikke_ raz: Can be, don't have experience with it. But you don't add a submodule that way12:54
raz unless there is a way to do anything useful out of the resulting commit12:54
← w09x left12:54
→ w09x joined12:55
raz _ikke_: ofcourse you don't. but it happens very easily by accident and is easy to go unnoticed until you clone the repo to some other place and end up with empty directories.12:55
→ sq-one joined12:55
raz i don't understand why git can't do the natural thing and simply add such a directory as a submodule when you 'git add' it. because that's very likely what the user actually wants when doing that.12:56
→ tyatpi joined12:56
_ikke_ raz: that's not how submodules work12:56
chffelix hi everyone..12:56
gitinfo chffelix: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.12:56
dailylinux _ikke_, hmm, not sure. I need it locally cloned at least :)12:57
raz _ikke_: perhaps i'm missing the forest for the trees yet again. can you explain to me in what scenario the current behaviour makes any sense?12:57
_ikke_ raz: I'm not saying this behaviour makes sense. I'm only saying your proposed alternative is not going to work12:57
raz _ikke_: then throw an error "you really don't want to do that"12:58
_ikke_ raz: Propose that on the mailing list12:58
raz don't silently pretend you added something which you didn't :/12:58
→ stuntmachine joined12:58
raz yes, i think i'm actually gonna type something up12:59
← pppaul left12:59
→ hellopat joined12:59
→ pppaul joined13:00
← w09x left13:00
→ ptolemyxi joined13:00
dailylinux _ikke_, you know how to fix it?13:00
→ w09x joined13:00
gitinfo set mode: +v13:00
ptolemyxi what does tracking info mean for a branch? is it needed?13:00
← chffelix left13:01
_ikke_ ptolemyxi: It's not crucial to git, but it's a convenience for you13:02
dailylinux: Depends on what needs to be fixed13:02
← h0bbit_ left13:02
dailylinux _ikke_, i need to clone the git git://gitorious.org/my-owncloud-apps/files_odfviewer.git13:03
and get it locally13:03
← w09x left13:03
dailylinux to run the code13:03
→ marhaban joined13:03
→ w09x_ joined13:03
→ h0bbit_ joined13:03
_ikke_ dailylinux: !xy13:03
gitinfo dailylinux: [!doinitrong] It sounds like you're approaching this problem in the wrong manner. Let's step back for a minute - What are you actually trying to achieve? Why are you trying to do it this way?13:03
← dvide left13:04
dailylinux _ikke_, trying to add a branch ( git://gitorious.org/my-owncloud-apps/files_odfviewer.git)to main project i've cloned git://gitorious.org/owncloud/owncloud.git13:05
← pppaul left13:05
→ pppaul joined13:05
_ikke_ dailylinux: You seem to mix up some terms13:06
← maleknet left13:06
_ikke_ git://gitorious.org/my-owncloud-apps/files_odfviewer.git is a repository, not a branch13:06
→ ehsan joined13:06
→ SkAzZ joined13:06
dailylinux _ikke_, ok, sorry... i need that repo added to my existing repos13:07
so i can do git pull and pull all along..13:07
_ikke_ well, you do that with git remote13:07
← mansour left13:07
dailylinux git remote add files_odfviewer git-addr?13:08
_ikke_ git remote add my-own-cloud git://gitorious.org/my-owncloud-apps/files_odfviewer.git13:08
→ Jackneill joined13:08
→ vervic joined13:08
_ikke_ yeah13:08
dailylinux that's done13:08
← Jackneill left13:08
→ Jackneill joined13:08
dailylinux so now git pull ?13:08
→ SkAzZ1 joined13:08
dailylinux i did a git fetch, maybe that screwed up things?13:08
_ikke_ dailylinux: !pull13:09
gitinfo dailylinux: pull=fetch+merge (or with flags/config also fetch+rebase). It is thus *not* the opposite of push in any sense. A good article that explains the difference between fetch and pull: http://longair.net/blog/2009/04/16/git-fetch-and-merge/13:09
← arietis left13:09
_ikke_ So you did only one part of the pull13:09
the second part is merge13:09
dailylinux ah, ok13:09
→ revagomes joined13:10
dailylinux git merge git-addr?13:10
this is "merging" locally?13:11
_ikke_ dailylinux: No, not git-addr13:11
you merge branches, not repositories13:11
→ machty joined13:11
← SkAzZ left13:11
→ alexisabril joined13:12
← foofoobar left13:12
← kcj left13:12
← w09x_ left13:12
→ vitor-br joined13:12
dailylinux git merge files_odfviewer ?13:12
→ juanfer joined13:13
_ikke_ git merge files_odfview/master13:13
dailylinux fatal: 'files_odfview/master' does not point to a commit13:13
_ikke_ dailylinux: Try reading a !book about git13:13
gitinfo dailylinux: There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable13:13
← milli left13:13
_ikke_ dailylinux: What does git branch -r return?13:13
dailylinux i can't see a [branch ......] for this branch13:13
http://fpaste.org/OnK6/13:14
_ikke_ dailylinux: Because it's not recorded in the config by default13:14
→ Gentlecat joined13:14
dailylinux ah, ok13:14
now i got error13:14
→ w09x joined13:14
_ikke_ git merge files_odfviewer/master13:14
dailylinux http://fpaste.org/uaqO/13:14
yep, i did that13:15
_ikke_ They don't apply cleanly13:15
dailylinux ah, ok13:15
_ikke_ git status13:15
→ kor-ilquelit joined13:16
→ teotwaki joined13:16
← w09x left13:16
→ w09x joined13:17
dailylinux http://fpaste.org/71hi/13:17
← stuntmachine left13:17
_ikke_ dailylinux: .gitignore merged13:17
← w09x left13:17
_ikke_ conflcited*13:17
→ w09x_ joined13:17
_ikke_ if you open it, you see conflict markers13:17
← sq-one left13:17
_ikke_ >>>>> and <<<<<13:18
→ hyperair joined13:18
dailylinux so things are ok now?13:18
i've opened .gitignore13:19
← Lgb left13:20
→ bartek joined13:20
dailylinux http://fpaste.org/L8aY/13:20
some conflicts it seems13:20
gitinfo set mode: +v13:20
kor-ilquelit hi~! Koreans here??13:20
→ Desproges joined13:20
kor-ilquelit how make *.git file?13:21
← Athlon^1 left13:21
→ ph^_ joined13:21
→ w09x joined13:21
← w09x left13:22
→ w09x joined13:22
_ikke_ ctrl-Uctrl-U13:22
← w09x_ left13:22
→ Lgb joined13:22
_ikke_ dailylinux: Some things seem duplicate13:23
→ Balreign joined13:23
← davidbanham left13:23
← ph^ left13:23
_ikke_ kor-ilquelit: What do you mean?13:23
→ w09x_ joined13:23
← bartek left13:23
dailylinux _ikke_, ok, never mind, i'll just clone and use it :)13:23
easier13:23
kor-ilquelit _ikke_ ex... test@localhost\project\test.git13:24
→ milli joined13:24
kor-ilquelit I'll make it in window..13:24
→ dv310p3r joined13:24
_ikke_ kor-ilquelit: That's a directory, not a file13:24
→ stuntmachine joined13:24
kor-ilquelit aha ..13:25
_ikke_ kor-ilquelit: Are more precise, that's a repository13:25
s/are/or/i13:25
→ bartek joined13:25
→ Ecesis joined13:25
← Ecesis left13:25
← w09x_ left13:25
→ w09x_ joined13:25
→ banghouse2 joined13:26
← w09x left13:26
→ nadavoid joined13:27
kor-ilquelit I installed freeSSHd in window.13:27
how connect .git??13:27
.git directory is D:\PROJECT13:28
→ zloy1 joined13:28
→ sq-one joined13:29
dailylinux _ikke_, thx for your help :)13:29
→ kukks joined13:29
← ConstantineXVI left13:30
← w09x_ left13:30
← berserkr left13:31
← OOPMan left13:33
→ berserkr joined13:33
→ TeckniX joined13:33
→ w09x joined13:34
← marhaban left13:35
→ asu_123_asu joined13:36
← ehsan left13:38
→ dcarper joined13:39
→ Per4u3e joined13:40
← LimitedAtonement left13:41
← tatsuya_o left13:42
← pppaul left13:42
← efreese left13:42
← asu_123_asu left13:43
← vervic left13:43
→ pppaul joined13:43
← yoklov left13:43
← axl_ left13:44
→ cooloney joined13:47
→ TomyLobo joined13:47
→ pandeiro joined13:48
← beautifulmind left13:48
← teotwaki left13:49
benbergman the company i work for wants to transition from cvs to git and wants to continue using some scripts (appropriately modified) to checkout the repository by date13:49
unfortunately, the conversion process seems to have not preserved the dates from cvs13:50
does anyone have any insight?13:50
i believe the tool used for the conversion was git-cvsimport, but i was not the one that did the conversion13:50
cbreak-work if git doesn't have the dates in git log (and you can't see them in the commits with git show -p commitid13:50
) then you probably lost the date during the conversion13:51
maybe redo the conversion properly?13:51
benbergman can you recommend anay resources for doing the conversion "properly"?13:51
cbreak-work I have no idea.13:51
I've not used CVS in decades13:51
for svn I'd use git-svn13:51
benbergman ok13:51
→ neurodrone joined13:52
← w09x left13:52
et a lazy person could import the cvs repo into svn and then use git-svn ;)13:52
→ ConstantineXVI joined13:53
cbreak-work two impedance mismatches for the price of one!13:53
→ sxie joined13:54
← pppaul left13:54
→ pppaul joined13:54
← tyatpi left13:55
← d0k left13:56
→ d0k joined13:56
offby1 where by "lazy" you mean "masochistic"13:57
→ ehsan joined13:57
→ stliu joined13:57
icq I have the next situation now:13:57
I converted a svn repo to a git repository13:57
it works correct13:58
then I cloned this git repository13:58
but git branch -r does not show the branches from the original git repo13:58
← cybersphinx left13:58
icq also if I setup git svn on this one I get a hang and doing git svn rebase13:58
with the next error:13:58
Unable to determine upstream SVN information from working tree history13:59
et don't do that13:59
icq I've read in a lot of places and I saw everywhere they were suggesting this solution13:59
I want a git svn repo so I can start the migration of the services to git while still pushing to svn in the meantime14:00
et cloning a repo that was created by git-svn and then using git svn again on either side is a recipe for pain in my experience14:00
BinGOs pure clusterfuck in mine.14:00
← pppaul left14:01
← flype left14:01
→ pppaul joined14:02
→ mjt joined14:04
charon it's "easy" if you know exactly how git-svn works. you "merely" need to copy over all of git-svn's remote-tracking branches into the exact same place, use the same branch config, and 'git svn fetch'14:04
← cagedwisdom left14:04
← alanp left14:05
charon if there are no other (git) remotes involved, something like git fetch origin refs/remotes/*:refs/remotes/* may take care of the remote branches (but note !backup)14:05
gitinfo Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups14:05
mjt how to "force-merge" an upstream-tracking branch to be exactly the same as current upstream, removing all local changes and additions?14:05
→ agileadam joined14:06
← pppaul left14:06
RandalSchwartz git checkout thatbranch; git reset --hard origin/thatbranch14:06
→ cakehero joined14:06
→ jackh joined14:06
→ pppaul joined14:07
→ alanp joined14:07
→ gtrak joined14:08
mjt um. It does something different. It removes local changes, instead of "reverting" them.14:08
RandalSchwartz yes14:08
isn't that what you want?14:08
you want to discard all local work, right?14:08
→ mrwalker joined14:09
mjt not exactly14:09
RandalSchwartz then ask for what you want. :)14:09
→ beatak joined14:09
→ bglusman joined14:09
RandalSchwartz or better yet, explain what you want without using words like "merge"14:10
goal - not mechanism14:10
→ ology joined14:10
← berserkr left14:11
→ beatak_m joined14:11
mjt upstream does fun games with stable branches, so that stable_x_y changes conflicts with master. There's "upstream" local branch which is now at some x_y stable version, and on top of that, there's some distribution-specific stuff in another branch. I want to preserve current local upstream _state_ but "forward" it to current upstream master branch, so that the "merge commit" will revert all x_y changes and return "back" to upstream/master.14:12
RandalSchwartz so you don't need the local modifications. why do you care about the history of those modifications?14:12
mjt i want to preserve current state of it since it is related distribution-specific development14:12
RandalSchwartz why not just resync with upstream/master?14:12
→ berserkr joined14:12
RandalSchwartz then just shove that stuff into a side branch14:12
if you force-merge, you'll forever be out of sync SHA1 wise14:13
AlexMax Is there a way to see branches with no branchname or tag that haven't been garbage-collected yet?14:13
RandalSchwartz you'll never be able to send new local mods upstream14:13
a branch without a name is just called a "commit" :)14:13
AlexMax I just used hg-fast-export and I'm missing some anonymous heads and bookmarked heads14:13
mjt um.14:13
SethRobertson AlexMax: Look here for the lost&found !fixup14:14
gitinfo AlexMax: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!14:14
← beatak left14:14
→ codeM0nK3Y joined14:15
mjt there was some trick, like checkout -b tmp upstream/master; merge -m ours upstream; branch -M tmp upstream iirc. lemme just try that...14:15
SethRobertson mjt: If you rewrote history, that will probably do bad things14:15
RandalSchwartz that'll still just get you a commit that *looks* like upstream, but is totally *useless* for shared development14:15
why do you want that?14:15
mjt that branch is not supposed to be used in shared development14:16
AlexMax SethRobertson: To make clear what I am trying to do, I am trying to transition from mercurial to git with an SVN upstream, and I'd like to get all of my old history as well.14:16
mjt actually i've no idea why, it is how debian used git for packages, and I still can't understand why.14:16
SethRobertson AlexMax: I'm not sure why you need to transition from hg to git if svn is upstream. Just reclone from svn14:17
AlexMax SethRobertson: I am14:17
However I'd like to transition all of my old commits from my old repository14:17
I do not have commit access to the SVN upstream14:17
SethRobertson Ah14:17
RandalSchwartz I'm still confused as to why you need history of changes that you're then throwing away14:17
if you need access to those commits, tag it or put it in another branch14:17
AlexMax Because I'd like to 'show off' my work on github14:17
RandalSchwartz then just reset to match upstream14:17
→ Fullmoon joined14:18
← Fullmoon left14:18
AlexMax instead of there simply being a single SVN revision that says "Merged AlexMax's patch"14:18
RandalSchwartz if someone at debian is proposing a wonky dev strategy like that, they need to be edicated. :)14:18
AlexMax I'd like to keep all that history14:18
RandalSchwartz educated14:18
→ DoBeL joined14:18
RandalSchwartz AlexMax - in case it's not clear, I'm talking to mjt. :)14:18
mjt um. there are two discussions going on, i'll just shut up and wait till the interruptee AlexMax will finish.14:19
→ fly9 joined14:19
→ joshteam joined14:19
AlexMax RandalSchwartz: Ah my paologies14:19
gitinfo set mode: +v14:19
AlexMax I haven't had enough coffee14:20
DoBeL hi, i've some question about git, i have a repo on github, and i want to take 4 commit from another repo (in github too) and add that 4 commits in my repo, it's possible ? how ?14:20
AlexMax I'll wait until I'm highlighted again14:20
with a response14:20
so sorry14:20
SethRobertson RandalSchwartz: They are trying to avoid compiler/typo commits by doing squash instead of rewriting history and perfecting is too hard to explain for them and do for most of their contributers14:20
Or so I assume14:21
RandalSchwartz yeah - still seems confusing though14:21
← kenperkins left14:21
SethRobertson AlexMax: Well, it isn't clear what your question is. I already told you how to find your anonymous commits14:21
mjt RandalSchwartz: but i think yes, just tagging should be enough. Now it's interesting how to push that new branch state (with reset --hard) to the shared repository which has disable non-ff set to true :)14:21
RandalSchwartz why have commits that don't actually influence the current state?14:21
RandalSchwartz shakes head14:22
RandalSchwartz wanders off14:22
mjt heh14:22
AlexMax SethRobertson: Well, I stated my intentions in full, in the hopes that perhaps there would be an easier way14:22
than using hg-fast-export, having to find lost commits by hand, then uh....somehow getting thsoe commits into the git-svn repo14:23
→ luxurymode joined14:23
SethRobertson mjt: You need to get agreement from everyone else and the admins to disable it because you will get into rewriting wars otherwise14:23
RandalSchwartz well - wait, why does the *shared* repository also have those commits that aren't influencing history either?14:23
maybe you guys need to learn to use something other than "master" for a branch name. :)14:23
SethRobertson AlexMax: Why not name the anonymous commits in hg then?14:24
→ SkAzZ joined14:24
RandalSchwartz git push origin master:work-in-progress14:24
mjt i rarely use "master" (or "origin" for that matter) myself.14:24
→ axl_ joined14:24
→ chrisgeorge joined14:24
→ legnaleurc joined14:25
RandalSchwartz I rarely create local tracking branches now14:25
I have my "master", and a bunch of origin/ branches, and that's it14:25
mjt ditto14:25
RandalSchwartz the only thing the local tracking branch does is get out of sync. that seems silly :)14:26
mjt usually not origin/* but upstream/* and debian/* (and a few others when needed)14:26
RandalSchwartz I can easily type git diff master...origin/maser14:26
→ Destos joined14:27
← pppaul left14:27
RandalSchwartz with gitolite, we're enabling user/merlyn/foo branches14:27
→ pppaul joined14:27
RandalSchwartz and user/merlyn/bar repos14:27
for communication outside the normal jenkins builds14:27
mjt but that resetting will be.. um.. since there's also debian branch which is based on this "local" upstream branch, and that one too will have to be synced somehow... um.14:27
RandalSchwartz mjt -= then why isn't the same sha1 coming back at you?14:28
mjt and i can't "just" reset --hard it, since that one does actually have something in there14:28
← SkAzZ1 left14:28
→ kevwil joined14:28
RandalSchwartz somewhere, someone else merged something14:28
adn if so, your local sha1 is now meaningless14:28
mjt yes it is meaningless14:28
RandalSchwartz anyway - must get in car to go to work. :)14:28
mjt thank you for the helo!14:29
help too! ;)14:29
→ tharkun joined14:29
← tharkun left14:29
← gavin_huang left14:30
← gchristensen left14:31
← kukks left14:31
← Weiss left14:31
→ airborn joined14:32
← pppaul left14:32
→ w09x joined14:32
← srinivas07 left14:32
→ pppaul joined14:34
AlexMax SethRobertson: I don't think hg-fast-export picks up on bookmarks14:34
and these were already put in branches14:34
erm14:34
and branches are forever14:34
cbreak-work until someone renames, rewrites or deletes them14:34
AlexMax Mercurial branches14:35
SethRobertson If hg-export picks up on *anything*, then simply do that same thing to the commits that are not being picked up on14:35
AlexMax cool14:35
← DoBeL left14:37
→ intransit joined14:38
→ macmartine joined14:38
← j0ran left14:39
→ minijupe joined14:40
minijupe how can I see the history of one line or group of lines in a file. I basically want to know when it was changed to its current state, and what it was before.14:41
offby1 minijupe: "git blame" will give you that, sort of14:41
← frogonwheels left14:41
← pppaul left14:41
offby1 probably better to find some sort of GUI tool that will show you the various layers of "git blame". "git gui" probably does this but I've never figured out how to use it.14:42
→ pppaul joined14:42
minijupe offby1: tnx, i tried that first but couldn't make it work. But just found this, which seems to be close enough. Pretty cool: git log -p14:42
git log -p filename14:42
→ stkrzysiak joined14:42
AlexMax SethRobertson: I found the commits I was looking for, they were dangling14:43
SethRobertson Yes, that was expected14:43
offby1 minijupe: hm, I don't see how "git log -p" does what you wanted, but ... as long as you're happy14:43
← Lord_DeathMatch left14:43
AlexMax Now, how do I transplant them to my git-svn repository, keeping the branching data intact?14:44
I expect the sha1's will be different14:44
but I don't care14:44
SethRobertson AlexMax: Add the hg exported repo as a remote14:44
→ edude03 joined14:44
SethRobertson You can then create local branches out of the remote tracking branches if you want14:44
Or just leave them as a remote.14:45
← sq-one left14:45
minijupe offby1: it doesn't do it exactly. I'll try your gui.14:45
AlexMax SethRobertson: Right, but how do i reattach the commits at the right place, since the sha's will be different14:45
SethRobertson You would need to use git replace, or cherry-pick, or the three argument version of rebase to attach the commits in the right place14:46
AlexMax Ah14:46
← jargon- left14:46
AlexMax Thank you so much, by the way14:46
← w09x left14:46
← Destos left14:47
← stliu left14:47
→ catphish joined14:47
← tqmz left14:48
→ Lord_DeathMatch joined14:48
→ Destos joined14:48
→ jargon- joined14:49
→ eguzman_ joined14:49
→ teotwaki joined14:50
← Lord_DeathMatch left14:52
→ w09x joined14:53
→ hwrd|work joined14:55
→ dcope joined14:55
← cooloney left14:55
← axl_ left14:55
→ axl_ joined14:56
← dcope left14:56
→ dcope joined14:56
← airborn left14:56
dcope Are submodules generally pretty bad? Every time I ask someone about them they say that they're more trouble then they're worth.14:56
→ gchristensen joined14:56
→ madewokherd joined14:57
← codeM0nK3Y left14:57
→ codeM0nK3Y joined14:58
→ Kunda joined14:58
→ n3m joined14:58
← Araxia left14:59
← machty left14:59
→ g0bl1n joined14:59
← g0bl1n left14:59
→ g0bl1n joined14:59
← ology left15:00
← rurufufuss left15:00
→ h0h0h0 joined15:00
← _Qman left15:00
← erichynds left15:01
→ ns5 joined15:01
→ tatsuya_o joined15:01
← flijten left15:02
gitinfo set mode: +v15:02
solussd with korma I'm getting cyclic load errors because my model namespaces are co-dependent- seems like this would be a problem in just about any database driven app using korma with relationships between entities housed in different namespaces. Are we suppose to keep our entire model in one namespace?15:04
← zloy1 left15:04
→ camdenx_ joined15:04
ns5 I found that after I push something to orig master, the git-daemon-export-ok file in the original repo will be removed. Any idea?15:04
← camdenx left15:04
solussd ^ sorry, wrong channel.15:05
→ mischat_ joined15:06
← w09x left15:07
← ISF left15:08
→ alex__c2022 joined15:09
← mischat left15:10
mischat_ → mischat15:10
→ popcorn joined15:10
← psoo left15:10
← ThomasLocke left15:11
← localhost left15:12
→ ology joined15:12
← n3m left15:13
← cbreak-work left15:13
→ tyatpi joined15:13
→ n3m joined15:13
← pppaul left15:14
AlexMax SethRobertson: Cherry pick seems to _almost_ work for me15:14
← dpino left15:14
← hyperair left15:14
SethRobertson How dopes it not work?15:14
variable → trout15:15
AlexMax SethRobertson: Well, in the original bookmark, I not only branch, but then I merge from upstream as well15:15
And when i try to cherrypick that range of commits, it brings the upstream commits as well15:15
(the old hgsubversion upstream, not the git-svn upstream)15:15
← n3m left15:16
SethRobertson ns5: Seems...unlikely. Can you !repro? ls -l .git/git-daemon-export-ok; git push; ls -l .git/git-daemon-export-ok15:16
gitinfo ns5: Please paste (using https://gist.github.com/ or similar) a transcript (https://gist.github.com/2415442) of your terminal session, or at least explain exactly what you did that led up to the problem. This will help immensely with troubleshooting.15:16
→ pppaul joined15:16
SethRobertson AlexMax: When you use `git log <range>` do you see the list of commits you expect to see?15:17
→ urbanmonk joined15:17
AlexMax SethRobertson: Well, not what I _wanted_15:17
← hwrd|work left15:18
AlexMax but i suppose i should have expected it15:18
SethRobertson What range did you use?15:18
AlexMax Just fe3cc5d..4f322c215:18
where fe3 is the first commit in the hg bookmark from the original hgsubversion15:18
← Sitethief left15:18
AlexMax and 4f is the latest commit in the hg bookmark15:19
SethRobertson Try using three dots and see if `git log` returns what you want then15:19
AlexMax There is a commit in between those two f23 that 'merges from upstream'15:19
→ cybersphinx joined15:19
AlexMax no it does not15:19
SethRobertson Oh, then cherry-pick the range before and after the merge, but not the merge15:19
AlexMax SethRobertson: Okay, but how do I get that merge then?15:20
→ hyperair joined15:20
SethRobertson Recreate it15:20
AlexMax Ah i see15:20
→ Silowyi joined15:20
AlexMax cherry pick the original range15:20
do the merge in git15:20
then cherry pick the rest15:20
gotcha15:20
makes sense15:20
← CannedCorn left15:21
← twisted` left15:21
→ okiwan joined15:22
AlexMax And I don't think this will change any of the shas in my existing repository15:22
← wolog left15:22
→ nrichards joined15:23
→ mikepack joined15:23
AlexMax so once this is done (in my copy of the git-svn repo with a remote of the hg-fast-export repo) i can then add the copy of the git-svn repo as a remote, fetch just that branch, and voila)15:23
→ kenperkins joined15:26
→ wolog joined15:26
popcorn → popcorn|away15:27
← kermit left15:27
popcorn|away → popcorn15:27
← jefferai left15:28
← ipoval left15:28
→ psoo joined15:28
← d3c left15:29
→ alberto56 joined15:30
← jargon- left15:31
→ kadoban joined15:31
→ kermit joined15:31
← amcsi_work left15:31
← cakehero left15:32
→ mtphys joined15:33
infogulch|afk → infogulch15:33
→ cakehero joined15:34
← jackh left15:35
← rtuin left15:35
→ jargon- joined15:37
→ thelazydeveloper joined15:38
→ _iron joined15:38
→ gtrak` joined15:39
→ wbednarski_ joined15:40
→ doki_pen joined15:40
→ beatak joined15:40
→ xref_ joined15:40
→ SeySayux_ joined15:41
→ ppppaul joined15:41
← gtrak``` left15:41
← pppaul left15:41
← beatak_m left15:41
← xref left15:41
← styx_ left15:41
← doki_pen_ left15:41
← SeySayux left15:41
← hays left15:41
xref_ → xref15:41
← defaultro left15:41
→ styx__ joined15:41
← papegaaij left15:41
→ yatiohi joined15:41
→ maletor joined15:42
→ hays joined15:42
→ defaultro joined15:43
← wbednarski left15:43
→ CannedCorn joined15:44
← maletor left15:44
← thierryp left15:45
→ jefferai joined15:46
← mishok13 left15:46
→ rath joined15:49
← blacktulip left15:49
← jargon- left15:50
→ HyperNerdV2 joined15:50
← jceb left15:51
← urbanmonk left15:52
→ SkAzZ1 joined15:52
neersighted|AFK → neersighted15:53
← SkAzZ left15:53
→ dp_ joined15:55
dp_ I accidentally just pulled changes from master in to another branch. How can I revert those?15:55
Vinnie_win dp_: git reset --hard {ref}15:55
→ airborn joined15:55
RandalSchwartz !fixup15:56
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!15:56
RandalSchwartz read that!15:56
→ raatiniemi joined15:56
RandalSchwartz a good url to bookmark for any kind of fixup15:56
→ dfr|mac joined15:56
dp_ Vinnie_win: thanks15:56
tango_ I accidentally another branch15:57
dp_ RandalSchwartz: will do. I read the git from the bottom up document. it was a great help to the basics15:57
tango_ this needs a meme.jpg15:57
← pdtpatrick left15:57
→ Anubisss_ joined15:57
← Anubisss left15:58
→ anth0ny joined15:58
← ns5 left16:00
← barrbrain left16:00
← dfr|mac left16:01
tango_ something like this http://memegenerator.net/instance/2044940916:01
← alexukf left16:01
← dark_element left16:02
← robotmay left16:02
← alexisabril left16:03
dp_ heh16:03
→ via joined16:04
via is there a way to clone a remote repo straight into a tarball?16:04
RandalSchwartz not clone, but look at git-archive16:04
git archive --remote ...16:04
via i was trying that, but it seems to request the tarball from the remote server instead of cloning it to produce it16:05
RandalSchwartz right16:05
do you want a *clone*, or a *tarball*16:05
← iamjarvo left16:05
via does that require the remote repo to be configured in some way?16:05
RandalSchwartz it has to be fairly recent16:05
→ flype joined16:05
← giallu left16:06
→ marhaban joined16:06
RandalSchwartz the git-archive remote protocol is relatively new16:06
I'm not sure of the revision required16:06
you *are* talking about ssh-git protocol right?16:06
→ alexukf joined16:06
RandalSchwartz not http:16:06
via and not git:// either?16:06
→ oddraise1t joined16:06
RandalSchwartz a git-server?16:06
not sure if that knows how16:06
→ Balreign_ joined16:07
→ Anubisss__ joined16:07
← hays left16:07
← mikepack left16:07
← nrichards left16:07
← g0bl1n left16:07
← Balreign left16:07
← ptolemyxi left16:07
← juvenal left16:07
← chris2 left16:07
← UukGoblin left16:07
← quuxman left16:07
← xxthink left16:07
← jrdavid left16:07
← Slush- left16:07
← ska-fan left16:07
← anildigital left16:07
← lindenle_ left16:07
← sjefen6 left16:07
← bitslip left16:07
← [E]sc left16:07
← talexb left16:07
← luka12345 left16:07
← OPK left16:07
← avar left16:07
← esc left16:07
← egeste left16:07
← solussd left16:07
← lpsmith left16:07
← narthollis left16:07
← oddraisent left16:07
← mnemoc left16:07
RandalSchwartz maybe someone else here knows16:07
dcope is there a decent tutorial for subtree?16:07
→ egeste joined16:07
RandalSchwartz if you google for "git subtree", there's a few example pages.16:08
I'm using subtree heavily for $work if you have a question16:08
← dr0id left16:08
→ esc joined16:08
← esc left16:08
→ esc joined16:08
→ mritz joined16:08
→ Espreon joined16:10
→ maletor joined16:10
← alexukf left16:10
Espreon Hey guys. How do I merge a local commit with a commit that's already in the remote repository? And yes, I do actually need to do this.16:11
← Anubisss_ left16:11
← mmc1 left16:11
→ hays joined16:12
→ mikepack joined16:12
→ nrichards joined16:12
→ g0bl1n joined16:12
→ quuxman joined16:12
→ chris2 joined16:12
→ UukGoblin joined16:12
→ xxthink joined16:12
→ lindenle_ joined16:12
→ solussd joined16:12
→ jrdavid joined16:12
→ Slush- joined16:12
→ lpsmith joined16:12
→ mnemoc joined16:12
→ ska-fan joined16:12
→ anildigital joined16:12
→ sjefen6 joined16:12
→ bitslip joined16:12
→ OPK joined16:12
→ [E]sc joined16:12
→ luka12345 joined16:12
→ avar joined16:12
→ narthollis joined16:12
→ fangism1 joined16:12
→ Fudge joined16:12
→ aaronky joined16:12
→ pigeon joined16:12
← UukGoblin left16:13
← [E]sc left16:13
← hays left16:13
FauxFaux Espreon: Using rebase -i and squash (or any other squash, i.e. merge --squash)? !rewriting16:13
gitinfo Espreon: [!rewriting_public_history] Rewriting public history is a very bad idea. Anyone else who may have pulled the old history will have to `git pull --rebase` and even worse things if they have tagged or branched, so you must publish your humiliation so they know what to do. You will need to `git push -f` to force the push. The server may not allow this. See receive.denyNonFastForwards (git-config)16:13
→ UukGoblin joined16:13
→ stkrzysiak_ joined16:13
→ [E]sc joined16:13
→ Anubisss joined16:14
← stkrzysiak_ left16:14
→ stliu joined16:14
→ hays joined16:14
→ Proctoglas joined16:14
← Anubisss__ left16:14
→ stkrzysiak_ joined16:15
Espreon Ugh...16:15
← stkrzysiak left16:15
stkrzysiak_ → stkrzysiak16:15
Espreon ... This is *not* going to end well.16:15
← whitman left16:15
Espreon I probably should just nuke my fork, refork, and make it perfect for the... gatekeeper.16:15
→ yoklov joined16:17
→ wolfcore joined16:17
← dunpeal left16:17
→ DofLeFou joined16:19
→ john_anderson joined16:20
← Proctoglas left16:21
← okiwan left16:21
← SkAzZ1 left16:21
→ talexb joined16:22
← popcorn left16:22
→ btree joined16:22
← dailylinux left16:22
← tyatpi left16:22
→ umc joined16:22
← senny left16:23
→ SkAzZ joined16:23
Espreon FauxFaux: ... Yeah... I think I'll just get one of these perfect commit-obssessed people to hold my hand through the demonic process. Thank you, though.16:24
← dcarper left16:24
← Kunda left16:24
← LimeBlast left16:25
← stliu left16:25
← Espreon left16:25
← revagomes left16:26
→ jceb joined16:26
→ dermoth|work joined16:27
→ sacarlson joined16:28
← marhaban left16:29
sacarlson seems when I checkout an older version that some of the files that weren't in the old are still in the git directory how can I remove them?16:29
→ ymasory joined16:30
← BarryCrayon left16:30
→ BarryCrayon joined16:30
← adhawkins left16:31
← Milossh left16:32
→ pdtpatrick joined16:32
← anth0ny left16:33
→ SkAzZ1 joined16:33
→ killerstorm joined16:33
→ drake01 joined16:33
← SkAzZ left16:33
→ alexisabril joined16:34
← BarryCrayon left16:34
sacarlson I tried git clean -f but it returns Not removing raddb/mods-enabled/ not sure why16:35
drake01 Do I lose something when I use http/https instead of git protocol while cloning a repository? What exactly is the difference16:36
← mikepack left16:36
→ juvenal joined16:36
RandalSchwartz sacarlson - perhaps some .gitignore'd files in there16:36
→ mikepack joined16:36
RandalSchwartz use -x if you *really* want things like a brand new checkout16:36
sacarlson RandalSchwartz: good Idea I'll check that16:36
← mikepack left16:37
← IcePic left16:37
→ mikepack joined16:37
← francisfish left16:37
← Vert left16:38
→ revagomes joined16:38
← sureaint left16:38
→ iamjarvo joined16:38
RandalSchwartz or the equally dangerous git checkout HEAD .16:39
← achadwick left16:40
→ eletuchy joined16:40
← jceb left16:40
killerstorm hi16:41
gitinfo killerstorm: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.16:41
→ Araxia joined16:41
→ avinash joined16:41
killerstorm Can somebody explain me how to use subtree-style foreign repo handling when there is more than one developer?16:41
I.e. both developers want to be able to update libraries from upstream and share stuff with each other.16:42
← drake01 left16:42
killerstorm This article: http://git-scm.com/book/ch6-7.html covers only one developer case16:42
→ sureaint joined16:42
→ jas4711 joined16:43
← gtrak` left16:43
→ nesneros joined16:44
→ gtrak` joined16:44
→ hellopat_ joined16:45
killerstorm Also I wonder which steps are really necessary. Do I need to create branches or remotes will do?16:45
← dr_lepper left16:46
← h0bbit_ left16:46
killerstorm Can I updates those branches without checking them out.16:46
← hellopat left16:46
→ dcarper joined16:47
← yoklov left16:47
← angelsl left16:48
→ kukks joined16:48
← kukks left16:48
→ kukks joined16:48
→ na3r joined16:48
→ gdoteof joined16:49
gdoteof idk what happened. i keep being told i am not on a branch16:49
RandalSchwartz killerstorm - git fetch -a (fetchs remote branches)16:49
gdoteof so i checkout a branch, and i am still not on a branch16:49
RandalSchwartz gdoteof - what kind of command for checkout?16:49
gdoteof git branch -a always lists my branches and * (no branch)16:49
git checkout remotes/origin/deploy16:49
RandalSchwartz yeah - that's a detached head16:49
you can look, but not touch16:50
gdoteof i want my head to come back on16:50
RandalSchwartz if you want to make commits, make a *local* branch16:50
gdoteof i don't.16:50
i guess, i did make some commits16:50
RandalSchwartz not on the remote branch you didn't16:50
gdoteof but i don't necessarily need to be doing that16:50
right16:50
RandalSchwartz does "git branch -a" show your local branch too?16:50
gdoteof RandalSchwartz: well, i am not sure. i might have made a localbranch called "origin/deploy"16:51
← nesneros left16:51
gdoteof i have 5: master, origin/deploy, remotes/origin/HEAD -> origin master, remotes/origin/deploy,remotes/origin/master16:51
← alberto56 left16:51
gdoteof what i am trying to do is pull a recompiled executable from remotes/origin/deploy16:52
RandalSchwartz so maybe you want "git checkout master"16:52
and then do what with it?>16:52
gdoteof the executable isn't in the master branch16:52
RandalSchwartz: and then ./ it16:52
→ jstout24 joined16:52
← abetaha left16:52
RandalSchwartz uh ./ ?16:52
gdoteof start it16:52
sorry16:52
RandalSchwartz git checkout origin/deploy16:52
cd (right dir)16:52
and run it16:52
killerstorm RandalSchwartz: git fetch will update remote branches, but article recommends creating local ones which track them16:53
RandalSchwartz which article?16:53
because I don't agree. :)16:53
killerstorm http://git-scm.com/book/ch6-7.html16:53
RandalSchwartz don't create local branches unless you plan to commit16:53
it just makes it more confusing16:53
killerstorm Yeah, definitely. I suspected that they are unnecessary :)16:53
→ RyuGuns joined16:53
← solussd left16:53
RyuGuns What is unstaging on git?16:54
RandalSchwartz RyuGuns - resetting the index for those items to the prior commit16:54
so that they won't be updated on a new commit16:54
← tvw left16:54
gdoteof now git says that i am not telling it which branch to merge with16:55
RandalSchwartz gdoteof - why are you trying to merge?16:55
gdoteof RandalSchwartz: i am not16:55
i did git pull16:55
RandalSchwartz NO16:55
pull = fetch + merge16:55
gdoteof mk16:55
good to know16:55
RandalSchwartz you didn't say you wanted to merge it somehwere16:55
you said you just wanted to run it16:55
my instructions did that16:55
now what *else* are you trying to do?16:56
RyuGuns Everytime I try to push something on github I get "Permission denied (publickey)."16:56
gdoteof hrm. i guess i didn't realize the checkout was giving me waht i needed16:56
sorry16:56
thanks16:56
RyuGuns I did everything according to the docs.16:56
→ machty joined16:56
RandalSchwartz RyuGuns - did you check a simple ssh ?16:56
→ scientes joined16:57
RandalSchwartz ssh github.com16:57
should return back a verification that all is well16:57
oh weird. didn't for me either.16:57
maybe github is broken :)16:57
cmn RyuGuns: what url are you using?16:58
→ dfr|mac joined16:58
RyuGuns Just sec16:59
→ IcePic joined16:59
← maletor left16:59
→ SkAzZ joined16:59
RyuGuns $ git push -u origin master17:00
Permission denied (publickey).17:00
Everything else works fine...17:00
← tatsuya_o left17:00
RandalSchwartz RyuGuns - try "ssh [email@hidden.address]17:00
oops17:00
RyuGuns - try "ssh -T [email@hidden.address]17:00
RyuGuns okay...17:00
RandalSchwartz if that doesn't work, go back to setting things up17:00
→ tyatpi joined17:01
← SkAzZ1 left17:01
→ SkAzZ1 joined17:01
RyuGuns Same thing.17:02
→ grayhoof joined17:02
gitinfo set mode: +v17:03
grayhoof is having no tracking information for a branch ok? what do I lose if I don't have one?17:03
cbreak RyuGuns: did you set up your public key?17:03
grayhoof: tracking infos are not required17:03
→ SkAzZ2 joined17:03
cbreak you won't get status infos and you won't be able to pull/push without specifying where to push/where to pull from17:03
unless you rely on less reliable modes17:03
← mischat left17:04
← SkAzZ left17:04
← shruggar1 left17:04
← killerstorm left17:04
→ SkAzZ joined17:04
RyuGuns cbreak: I set it up..17:05
Exactly as the docs on github said too..17:05
cbreak what did you do to set it up?17:05
← defaultro left17:05
→ defaultro joined17:05
← osX64 left17:05
← SkAzZ1 left17:06
→ Shusshu joined17:06
→ abetaha joined17:06
→ Kunda joined17:07
→ isao joined17:07
cmn clearly you didn't, what did you put in .ssh/config?17:07
→ samtuke joined17:07
← SkAzZ2 left17:08
RyuGuns They didn't mention anything about .ssh/config ...17:08
samtuke best way to bring specific changes from master back to a branch that is behind master on those files, without merging all changes?17:09
← arvind_khadri left17:09
→ osX64 joined17:09
→ alvaro_o joined17:11
cbreak you don't need to change .ssh/config17:11
← Vile left17:11
RandalSchwartz cbreak - you might if you're using a separate ssh key just for github (recommended)17:11
cbreak samtuke: there's only merging history, rebasing on history, or patching stuff in between with cherry picking17:11
RandalSchwartz: why's that recommended?17:12
← irqq left17:12
RyuGuns brb17:12
cbreak even if it is, ssh-add ing an other key shouldn't be that tragic17:12
← alvaro_o left17:12
cmn if you have many ssh keys, you do need to change .ssh/config17:13
→ erichynds joined17:13
← Peste_Bubonica left17:13
samtuke cbreak, cherry picking is what I need I think - just apply changes from a few specific commits from master to the unstable branch, as fixes that I need occurred in master since the branch occurred17:13
cmn otherwise the key negotiation will take too long17:13
← cortana2 left17:13
cmn erm... will try too many wrong keys, I mean17:13
cbreak samtuke: be aware that that'll create a historic mess17:13
← Davey left17:14
samtuke cbreak, the history may be confusing, but it wont stop a full blown merge between the master and unstable branch later right? its a cosmetic mess not a functional mess17:15
→ Peste_Bubonica joined17:15
← ShooterM_ left17:15
→ ShooterM_ joined17:15
→ martinjlowm joined17:15
MacGyverNL RandalSchwartz: Why is it recommended to use a separate SSH key for github?17:15
RandalSchwartz well - call me paranoid17:16
→ killerstorm joined17:16
RandalSchwartz I use a different pub key for every one of my clients :)17:16
cmn MacGyverNL: you should use a separate key for each comptuter/remote computer combination17:16
cbreak samtuke: it might cause conflicts17:16
MacGyverNL I am paranoid, but I use one SSH key per client I connect from, not per server I connect to.17:16
cbreak don't know, I never did that17:16
→ drev1 joined17:16
→ mishok13 joined17:17
RandalSchwartz because occasionally, I've had to install my private key for that client into one of the clients machines to do multi-hops17:17
← ShooterM_ left17:17
samtuke cbreak, well I can always merge them if necessary. nothing more serious than that right?17:17
FauxFaux cmn: What does that protect against? Misrepresentation and (the chance of a key being factored) aren't interesting to me.17:17
cbreak RandalSchwartz: one key per client makes sense, but not one key per server17:17
MacGyverNL RandalSchwartz: But that's what agent forwarding is for. Not ideal, but still better than putting the key on the hop.17:17
RandalSchwartz "client" here in the sense of "paying customer"17:17
MacGyverNL Ah.17:18
→ dreiss joined17:18
cmn if the service or machine is compromised, only one link is compromised17:18
RandalSchwartz sorry - overloaded terms17:18
→ d4rklit3 joined17:18
cbreak samtuke: you'll have "duplicate" commits.17:18
that's it.17:18
samtuke cbreak, I think its worth it - I can't see a better option. I need the functionality from master in my unstable branch17:19
cbreak rebase it? :)17:19
← axl_ left17:19
cbreak I rebase my own topic branches all the time17:19
→ gusnan joined17:19
cbreak works fine.17:19
→ axl_ joined17:20
samtuke cbreak, can I rebase just a few files, or better yet just files within one directory?17:20
→ h0bbit joined17:20
cbreak no17:20
that makes no sense17:20
← killerstorm left17:20
cbreak rebasing is history rewriting17:20
it has nothing to do with files17:21
samtuke cbreak, right, so to get just the changes I want across I have to cherrypick17:21
cbreak why not rebase/merge?17:21
you'll have to do it eventually17:21
and it's easier if you do it early :)17:21
→ sphex joined17:23
→ mikaello joined17:23
→ toabctl joined17:24
sphex hey. when using git-subtree to include several sub-projects, IIUC I should fetch them with --no-tags or else tag names could collide?17:26
→ DuzyZiemniak joined17:26
← dreiss left17:27
RandalSchwartz sphex - you aren't fetching them17:27
subtree is17:27
→ FreeArtMan joined17:27
RandalSchwartz either git subtree add (the first time)17:27
or git subtree pull (to refresh)17:27
in neither case will you be getting the tags17:28
← thorbjornDX left17:28
→ osmosis joined17:28
→ arietis joined17:29
sphex oh.. ok17:29
→ Davey joined17:29
← toabctl left17:30
sphex hrm. I'm confused.17:31
← Bass10 left17:32
EugeneKay That means you're learning.17:32
← juvenal left17:32
→ fuzzybunny69y joined17:33
sphex RandalSchwartz: is it that if I give the remote repo URLs directly to git subtree, it'll do the equivalent of fetch --no-tags internally?17:33
→ abhinavmehta joined17:34
→ juvenal joined17:34
RyuGuns Funk yes17:35
→ HyperNerdV2_ joined17:35
← Desproges left17:35
← d3vic3 left17:35
→ oehman joined17:35
abhinavmehta I've 1 git-repo, since form sometime(months) I'm not using that repo…and now I want to free that repo, by taking that code to another repo, which I call as 'misc', now how to do that..? I want my all commits and history of my subject-repo.17:36
RandalSchwartz sphex - I suspect so17:36
it's only fetching the branch you specify17:36
← osX64 left17:36
RandalSchwartz not all the rest of the branches or the tags17:36
RyuGuns I did "ssh-add ~/.ssh/id_rsa" and it worked17:36
→ Bass10 joined17:36
← HyperNerdV2 left17:36
HyperNerdV2_ → HyperNerdV217:36
→ shruggar joined17:37
abhinavmehta so how to move one repo(assume named as 'x') to another(as a sub branch/folder) with preserving all my commits to 'x'-repo??17:37
→ marc_kdab joined17:37
abhinavmehta am i clear..?17:37
RandalSchwartz just a single branch17:37
← Bass10 left17:37
sphex RandalSchwartz: ah! ok. yeah, makes sense. I forgot you could even do that.17:38
→ toabctl joined17:38
→ Bass10 joined17:38
PerlJam abhinavmehta: it sounds like you want to subtree merge.17:38
abhinavmehta PerlJam: give me sometime, I've to read what is subtree merge..I've no idea about this.17:39
← fir31ight left17:39
RandalSchwartz are you trying to change the paths within the commits?17:39
if so, filter-branch or the new shiny git-subtree would be easiest.17:39
PerlJam oh, yes. that too.17:39
RandalSchwartz or are you just moving commits around17:39
← vitor-br left17:40
RandalSchwartz is deploying piles of code based around git-subtree17:40
→ tvw joined17:40
← g0bl1n left17:40
→ EzeQL joined17:40
sphex RandalSchwartz: thanks. say I wanted the tags though. is there a way to have fetch prefix their name with the remote name somehow?17:40
→ d3vic3 joined17:41
EzeQL Hi, im receiving this error while trying to update a submodule : fatal: reference is not a tree: 18a1d3188683263a2a8ac6eef3aedccb0ece9b14 , Unable to checkout '18a1d3188683263a2a8ac6eef3aedccb0ece9b14' in submodule path 'shared'17:41
→ txomon joined17:41
RandalSchwartz sphex not really17:42
the tags are meaningless because the local commits have a different sha1 anyway17:42
← HyperNerdV2 left17:42
Vinnie_win RandalSchwartz: Do you mean you are deploying code to the git project, or your own projects?17:42
→ osX64 joined17:42
RandalSchwartz I'm deploying code for $client17:43
txomon I think I found a git annex bug17:43
→ killerstorm joined17:43
← ronniy left17:43
RandalSchwartz I suppose you could argue that the synthetic commit mapping to a remote tagged commit might want to get the same tag17:43
txomon if you run: git init; touch a; git annex init thisrepo; git annex add *; git annex unlock *; git annex add *;17:43
RandalSchwartz but then you'd have potential collision problems anyway17:43
← dc5ala left17:44
txomon it will add twice the same element, without having in account the modifications17:44
init >_>17:44
txomon If anyone oc17:44
← marcogmonteiro left17:44
→ fir31ight joined17:45
→ dr_lepper joined17:45
← killerstorm left17:45
→ Chillance joined17:45
← shruggar left17:46
sphex RandalSchwartz: ah ok. yeah I would only want them to make exploring the history a bit easier. can be helpful to see when releases happened, even if they only match the "real" history of the sub-project. I just found out about git-ls-remote though, I guess that's what I should be using...17:46
→ thearthur joined17:46
→ shruggar joined17:48
← Chib left17:50
txomon so any idea to know if git annex had already added it or not? or is there anything I can do about this? or anything more reliable than annex?17:50
← samtuke left17:50
→ mischat joined17:51
→ hwrd|work joined17:52
→ maletor joined17:52
← shruggar left17:52
← eletuchy left17:53
→ SkAzZ1 joined17:53
bremner txomon: did you try asking on #vcs-home on oftc? some git annex users (and the author) hang out there.17:54
txomon hum didn't know about it, will go there ty bremner17:54
bremner, was that correct channel?17:54
bremner yes17:55
← SkAzZ left17:55
→ txomin joined17:56
← txomin left17:56
← txomon left17:57
→ txomon joined17:57
txomon sry bremner xchat crashed. Didn't see your reply17:58
→ ShooterM_ joined17:58
← h0bbit left17:58
← dp_ left17:58
→ JeroenDeDauw joined17:59
← raatiniemi left18:00
← sacarlson left18:00
bball- I have created a branch off of master.. checked out the branch and made some chagnes (did not commit the changes)...18:00
→ williamcotton joined18:01
→ bitkiller joined18:01
txomon bball-, and what?18:01
bball- now I checked out master again but I still see the modified files from the branch Iw as working on earlier18:01
txomon bball-, did you add them to git?18:01
EzeQL Hi, im receiving this error while trying to update a submodule : fatal: reference is not a tree: 18a1d3188683263a2a8ac6eef3aedccb0ece9b14 , Unable to checkout '18a1d3188683263a2a8ac6eef3aedccb0ece9b14' in submodule path 'shared'18:01
bball- they are already being tracked18:01
all I did was modify them..18:01
← gusnan left18:02
txomon EzeQL, so check it exists18:02
bball-, git status what tells you?18:02
→ StamEhad joined18:02
← abhinavmehta left18:03
bball- ok.. it shows them as modified and that I need to use git add to update what will be committed18:03
→ krz joined18:03
bball- so if I add them while I have branch 'a' checkedout.. when I checkout branch 'b' I won't see them as being modified anymore ?18:03
→ subbyyy joined18:04
← StamEhad left18:04
← subbyyy left18:04
→ subbyyy joined18:04
← subbyyy left18:04
→ shruggar joined18:05
← TTSDA left18:05
← babilen left18:05
← flype left18:05
→ subbyyy joined18:06
← subbyyy left18:06
→ efreese joined18:06
→ cleardance joined18:06
→ subbyyy joined18:06
← subbyyy left18:06
← efreese left18:07
→ subbyyy joined18:07
bball- txwikinger: ok.. I did git add -u while I was in my branch18:07
and then checked out master and they are still showing up as modified ?18:07
err txomon18:07
cleardance What is the standard way to get a new version of a project? Like if i pushed a version to github, a partner did some changing and pushed it, how should i get the newest? do i just clone and overwrit the old files? can git switch back to older versions easily(i guess that is the point...)?18:08
PerlJam cleardance: git pull18:08
txomon puf, never had that problem before bball- ... can't help you now atm, wait till someone else replys18:08
← rrn left18:08
cmn cleardance: the point of version control is that you don't have to overwrite anything18:09
you fetch from them18:09
and then see how you want to approach updating, maybe a merge, maybe you do want to abandon local changes18:09
→ TTSDA joined18:10
JeroenDeDauw I have 2 git repos created from doing a git to svn thing. One repo has an incomplete history for some reason and a pile of changes. The other has the full history and no changes (since the svn to git thing). How can I apply the changes from the former onto the later?18:10
cleardance so clone is just for the first time? or could you fetch the first time as well? ( i mean if your partner started a project and you dont have any of it yet)?18:10
→ devth joined18:10
← tfittsy left18:10
← Dreamer3 left18:10
cmn clone is setting up the repo and then fetching18:10
← fisted_ left18:11
→ rrn joined18:11
→ Weiss joined18:11
→ adhawkins joined18:12
txomon cleardance, the idea is that fetch will assume at least the first commit to be the sames18:12
same*18:12
cmn huh?18:12
fetch doesn't assume that anything is the same18:13
→ imachuchu joined18:13
bball- txomon: just FYI.. once I commited the changes in my branch they were no longer visible when I checked out master18:14
← agileadam left18:14
→ sacarlson joined18:14
txomon cmn, to be able to merge yes, I don't know if it allows you having data from another repo with no common point in history18:14
cmn fetch doesn't have anything to with merge18:15
if you have commits in common, you'd download less data, nothing else18:15
→ fisted joined18:15
cmn and fwiw, merge doesn't assume or need common commits18:15
txomon ... you are correct... doesn't make sense.. Don't know where I got that idea from18:17
→ abhinavmehta joined18:17
cmn you're mixing up fetch, pull and sensible use or merging18:17
→ robotmay joined18:18
→ rubyruy joined18:18
→ zastaph joined18:19
← bitkiller left18:19
→ Mowsh joined18:19
← jas4711 left18:20
← williamcotton left18:20
abhinavmehta In master-branch I added new branch aka 'develop'.(using git checkout -b) and committed and pushed that to git-servers. Now today I've again created new branch aka 'other' from 'develop'..now I want to push this branch too to git-server….but I want that later my git log should show this 'other' branch as a branch born out from develop rather than master..how to do that...?18:20
zastaph http://git-scm.com/book/en/Git-on-the-Server-The-Protocols .. even Local Protocol requires you to install git on the server? I considered putting my source on a SMBFS/CIFS NAS, which doesn't currently have git installed18:20
abhinavmehta I know I may write "git push other origin/develop"..but not sure..that's why asking you guys..18:20
→ andreagelati joined18:21
→ alberto56 joined18:21
abhinavmehta sorry I mean "git push -u other origin/develop"18:21
→ vbajpai joined18:22
vbajpai should I delete a branch, once I am done with it and have merged it with master?18:22
→ Dreamer3 joined18:22
← alberto56 left18:22
← dv310p3r left18:23
abhinavmehta vbajpai: if you delete, where will you end-up putting your code…either you have to merge it with master or have to take it out.18:23
vbajpai: so its better to merge with master, with appropriate commit-message IMO18:24
→ Vile joined18:24
← dcope left18:24
← rubyruy left18:24
vbajpai abhinavmehta: as I said, I already have merged it with master, should I delete the branch?18:24
→ rubyruy joined18:24
→ Dreamer3__ joined18:25
imachuchu vbajpai: completely up to you. A branch costs ~40 bytes to keep, and git can hold many many (many!) branches at a time18:25
← Dreamer3__ left18:25
← marthinal left18:25
cmn zastaph: that makes no sense, local doesn't talk with the "server"18:25
vbajpai imachuchu: i see it more as a namespace issue18:25
→ eletuchy joined18:25
imachuchu vbajpai: so delete whenever you don't want to see one again, and create/recreate whenever you want to18:25
vbajpai: what do you mean "namespace"18:26
zastaph cmn, i agree.. but if you read the link it says that only "Http" of the 4 doesn't "require git to be installed"18:26
vbajpai imachuchu: well, the branch names are shared as well, because i pushed the branch to remote as well18:26
← jpalmer left18:26
cmn dumb http doesn't require git to be installed, smart http does, and dumb http should die in a fire18:26
→ jpalmer joined18:26
← Dreamer3 left18:27
zastaph cmn, but the link says that Local Protocol requires it.. does that mean that if I use a NAS for remote repository, that NAS needs git installed?18:27
imachuchu vbajpai: do you/others plan on using this branch again?18:27
cmn zastaph: no it doesn't18:27
← Per4u3e left18:27
vbajpai imachuchu: you mean the branch name? or the branch itself?18:27
zastaph cmn, so Local+http doesn't require git, ssh+Git does ?18:28
cmn local doesn't require any other git than the one you're using anyway18:28
any other requires git because you're talking to another computer18:28
except for dumb http which is a pile of crap18:28
zastaph but I'm talking about local protocol, so that you actually talk to a local network resource (a NAS)18:29
cmn that's not git's job in the local case18:29
← machty left18:29
→ MANCHUCK joined18:29
zastaph ok, i'm confused :)18:29
cmn you tell your OS to talk to whatever other system is providing the filesystem18:29
and you tell git to use that filesystem18:29
← hwrd|work left18:29
imachuchu vbajpai: the branch is just a pointer to a commit (and when you add a new commit it just increments the pointer) and a name, so really it is just a name.18:29
cmn it doesn't matter if the filesystem is on the other side of the world, as long as it's accessible to git via the filesystem18:30
vbajpai imachuchu: good point18:30
zastaph right18:30
cmn, someone should update those docs i guess :p18:30
cmn there is nothing wrong on that page18:30
imachuchu vbajpai: does that help?18:30
vbajpai imachuchu: yes, it does18:31
bremner txomon: yes, #vcs-home on network oftc18:31
← vbajpai left18:31
→ Dreamer3 joined18:32
cmn other than being outdated18:32
← andreagelati left18:32
→ SkAzZ joined18:33
txomon bremner, yeah, got there ty18:33
← shruggar left18:33
→ khmarbaise joined18:35
← SkAzZ1 left18:35
← osmosis left18:35
← thiago left18:38
→ dpino joined18:38
← Dreamer3 left18:39
zastaph for Local Protcol, is SMBFS/CIFS the best choice if both windows, mac and linux needs access?18:41
→ doug joined18:42
→ Targen joined18:44
cmn it's better to have git run on your actually local filesystem and push out with ssh or wahtever else you have18:44
→ alberto56 joined18:44
← sdeboey left18:44
→ flakblas joined18:45
→ hwrd|work joined18:45
zastaph im only going to access the git repository when home, so don't really need the authenticaiton etc. and if I need ssh I need a new server besides the NAS18:46
← banghouse2 left18:46
cmn but if you're going to use a shared filesystem for the main repo and not the ones you work on, whatever works for your OS will be fine18:47
git won't care18:47
→ banghouse2 joined18:47
cmn it will still be faster over ssh often18:47
← Macca- left18:47
→ osmosis joined18:48
← oehman left18:48
← alberto56 left18:48
→ macabre joined18:48
→ alberto56 joined18:49
→ CptCaptain joined18:49
CptCaptain Hey guys18:49
How can I change the language of git?18:50
← abhinavmehta left18:50
flakblas CptCaptain: What do you mean?18:50
← alberto56 left18:50
cmn you set your environment settings to whatever language you prefer18:50
and then upgrade to a git version with l10n support18:50
← fuzzybunny69y left18:51
CptCaptain A recent update introduced the german translation but I want to use it in englsh18:51
flakblas Oooh. I get it. Real languages. I'm thinking like C, Perl, etc.18:51
cmn right18:51
CptCaptain I would prefer not to touch18:52
cmn so set the environment variables to en_US or C18:52
← Vinnie_win left18:52
CptCaptain *$LANG so the rest stays the way it is18:52
cmn that's the only way to tell gettext to change the language18:53
if you want the rest to stay the way it is, change it only for git, for example in a wrapper script18:53
flakblas Could you alias the git command and have it run a wrapper that sets LANG temporarily for the duration of the git command?18:53
doug i'm in the kneejerk habit of "git fetch; git merge --ff-only origin/mybranch"18:53
what's the Right Way To Do That?18:53
Aaeriele_ → Aaeriele18:53
cmn if you're never going to inspect the changes, set up tracking and pull --ff-only18:54
CptCaptain So there is no option in gitconfig?18:54
→ mmc1 joined18:54
cmn why would there be?18:54
→ sdeboey joined18:55
→ marhaban joined18:55
CptCaptain cmn: Why not?^^18:55
→ h0bbit joined18:55
cmn language is an environment setting18:55
raek command line applications are usually in english. it feels pretty weird when they are not.18:55
I mean, the computer speaks english...18:55
cmn the computer speaks whatever you tell it to speak18:56
doug cmn: set up tracking? what's that?18:56
raek GUIs are usually translated though18:56
→ HyperNerdV2 joined18:56
cmn and if your environment says that your preferred language is German, it shouldn't speak in English if it can18:56
doug: man git branch --set-upstream18:56
gitinfo doug: the git-branch manpage is available at http://jk.gs/git-branch.html18:56
← rath left18:56
raek CptCaptain: I guess you could set the LANG variable in your .bashrc file (or equivalent)18:57
flakblas I think if you really need it you should just write a wrapper script and alias git to it. Have the wrapper set LANG to whatver you want.18:57
raek then it would only affect command line apps (or apps started through a shell)18:57
flakblas raek: He only wants it in effect for git. Not the whole env.18:57
→ twisted` joined18:58
→ Vinnie_win joined18:58
raek flakblas: I assumed CptCaptain had the experience as me, namely that git is the only (or one of the few) command line app on the system which is not translated18:58
s/not//18:59
cmn git wasn't translated until recently; once it could, it started respecting the language settings18:59
flakblas Well, he stated that he didn't want to change LANG for the whole env so that's why I suggested a wrapper script.19:00
raek is thinking about hacking together a "en_SE" locale again19:00
cmn yes, we've been talking about wrapper scripts for a while now19:01
→ SegFaultAX joined19:01
→ adamcooke joined19:03
→ mdhopkins joined19:04
→ whitman joined19:06
← mmc1 left19:07
← teotwaki left19:09
→ machty joined19:10
← cleardance left19:10
→ jameswarren joined19:10
→ f0i joined19:11
→ g0bl1n joined19:11
→ diegoviola joined19:14
← _md left19:14
→ thiago joined19:15
→ tatsuya_o joined19:16
→ alberto56 joined19:16
← doug left19:16
→ ben225 joined19:17
← Sky[x] left19:17
← ToBeFree left19:19
← wbednarski_ left19:19
→ blacktulip joined19:19
→ d3c joined19:19
← Gentlecat left19:19
← Lgb left19:20
← psoo left19:20
→ mcluskydodallas joined19:20
→ tqmz joined19:21
→ mmc1 joined19:21
← syphar left19:21
→ efreese joined19:22
→ Lgb joined19:22
→ loco joined19:23
→ toofar joined19:23
← stodan left19:23
gitinfo set mode: +v19:23
← Kunda left19:24
loco Hello. If I work on a separate branch, to test new features and not to demolish master branch, what is the best way to come back to master and commit changes I've done on a separate branch to master?19:25
→ dv310p3r joined19:25
← d3c left19:26
→ axl__ joined19:26
← hwrd|work left19:27
gitinfo set mode: +v19:27
← noisebleed left19:28
← ipalaus_ left19:28
imachuchu loco: if a commit logically holds changes, then checkout master and cherry-pick the commits you want19:28
← alberto56 left19:29
loco you refer to the last commit on my separate branch? I'm reading about cherry-pick now, but it seems an advanced option and I'm a noob19:29
→ alberto56 joined19:29
← minijupe left19:30
← axl_ left19:30
axl__ → axl_19:30
← kor-ilquelit left19:30
imachuchu loco: cherry-pick takes the changes introduced in a commit and applies them to the current commit to create a new commit. This is used if you only want some of the changes, if you want everything then merge would be the correct command19:30
→ cantoma joined19:31
imachuchu loco: checkout master, merge feature. This brings over all of the changes in feature, though you may want to look up the difference between an ff-merge and a merge node19:32
→ ISF joined19:33
→ d3c joined19:33
← erichynds left19:33
← HyperNerdV2 left19:33
← efreese left19:34
loco I'd like to make some changes on the separate branch, make a few commits then on this branch, and after the last commit on the separate branch, when I know I want to have these changes on my master, switch back to master and make a commit with changes from the last commit on the separate branch19:34
← isao left19:35
→ isao joined19:35
← dv310p3r left19:35
← aboudreault left19:36
→ josefig_ joined19:36
→ wbednarski joined19:36
→ NiklasFiekas joined19:37
imachuchu loco: so what you want to do is have one commit with all of the changes introduced in feature, or just the changes in the last commit (commits are, conceptually, changes so each commit could be thought of as changes from the previous state of the data)19:37
loco: *last commit on feature19:38
← Targen left19:38
loco changes in the last commit19:38
← iwoj left19:38
→ Targen joined19:39
→ bitkiller joined19:40
← gerard0 left19:40
imachuchu loco: then cherry-pick the last commit, rebase -i afterwards if you need to fixit up at all (like remove some changes, or change the commit message)19:40
→ JohnyBaby joined19:40
loco ok, thx19:41
← josefig_ left19:41
← sphex left19:41
→ nowhereman joined19:41
← _W_ left19:41
← nowhere_man left19:42
← CptCaptain left19:42
→ _W_ joined19:42
loco but after cherry-pick I can modify my commit message, without the need to rebase? I don't understand rebase very well yet19:42
→ sjbrown joined19:43
loco with amend for example19:43
sjbrown hello, i have a problem - i accidentally did a git checkout of a checksum, and i want to get my working dir back to their previous contents19:43
NiklasFiekas sjbrown: do you know the name of the branch you've been working on? master? then "git checkout master"19:44
sjbrown master, yeah.19:44
kevlarman loco: yes you can use amend19:44
imachuchu loco: right, sorry, rebase is more general purpose (it can reword any commit message), amend can only do the previous one19:44
sjbrown ok sweet, i'll try that.19:44
← bitkiller left19:45
← catphish left19:46
← machty left19:46
→ eidolon1 joined19:46
EugeneKay sjbrown - !reflog is worth a look19:46
gitinfo sjbrown: The git reflog (`git log -g`) records the SHAs of your HEADs for 2+ weeks. `git checkout -b myrestore OLDSHA` and `git reset --hard OLDSHA` will relink to that state via a new and current branch respectively, see http://sethrobertson.github.com/GitFixUm/ for full details. WARNING: reset --hard will trash any uncommitted changes! Visualize with: gitk --all --date-order `git log -g --pretty=%H`19:46
→ bitkiller joined19:47
← na3r left19:47
→ machty joined19:48
→ sntxrr joined19:48
NiklasFiekas EugeneKa, sjbrown: however git reset --hard will still leaf HEAD detached19:48
→ snearch joined19:48
← snearch left19:48
→ snearch joined19:48
← Ryback_ left19:48
← CannedCorn left19:49
← eidolon1 left19:49
→ eidolon joined19:50
eidolon hi folks, we're trying to clone a git local repo (no remote server) from one machine to another. they've got it up on apache httpd - i can see it in the browser, btu if i do 'git clone http://blah/path' i'm seeing 'refs not found: did you run git update-server-info on the server' ? - we do that, but nothing happens.19:50
← machty left19:50
sntxrr hello all. it seems I checked in some sensitive code to my repo - I deleted and recreated the repo, removed the offending file via: http://help.github.com/remove-sensitive-data/ - and pushed back to my new master.19:50
i'm still seeing a live link to the commit.. and not a cached one - any suggestions?19:50
FYI, asked in #github w/o any response19:50
← nowhereman left19:50
→ thebignoise joined19:51
← alberto56 left19:51
← jave left19:51
→ nowhereman joined19:51
haarg where are you seeing the link?19:51
sntxrr github19:51
haarg where on github?19:51
← bitkiller left19:51
→ alberto56 joined19:51
sntxrr in the commit history19:52
→ machty joined19:52
sntxrr https://github.com/<MY GITHUB USER>/<MY PROJECT>/commit19:52
haarg are you certain the commit doesn't exist in your local repo?19:52
→ SkAzZ1 joined19:52
→ bitkiller joined19:52
→ jave joined19:53
→ toofar_ joined19:53
← alberto56 left19:53
← toofar left19:53
← DoniRS left19:53
cmn how are you sure it's not cached?19:53
→ DoniRS joined19:53
→ Pistos joined19:53
← isao left19:53
sntxrr cleared the browser cache, visible on multiple computers in multiple locations19:53
← macabre left19:54
→ yoklov joined19:54
← snearch left19:54
imachuchu eidolon: I have no idea how to fix your problem (so others feel free to diagnose), but I've suggested multiple times to use git bundle if there is an issue with git over the network19:55
eidolon: bundle, copy, clone/merge on the target19:55
haarg sntxrr: are you still able to see the actual problem commit?19:55
not just the link in the history19:55
← machty left19:55
← SkAzZ left19:55
cmn fair enough, how did you push? what are you using to see this commit?19:55
are you going directly to the commit's hash?19:56
because /user/repo/commit isnt' a valid URL19:56
or path, rather, so you're probably going to /usr/repo/commit/obj-id ?19:56
→ Dreamer3 joined19:56
sntxrr seems that the removed file does NOT appear in a fresh clone of the repo19:58
→ iwoj joined19:58
← bitkiller left19:58
← g0bl1n left19:59
sntxrr cmn - web browser, from emailed comment from a repo watcher noting the commit of sensitive data19:59
← enmand left19:59
← john_anderson left19:59
NiklasFiekas sntxrr: does it contain the sha?20:00
sntxrr i deleted the repo (via github site). created an identially named one. git push origin master --force20:00
benbergman stuntmachine: from the page you linked: "If the commits were viewed online the pages may also be cached."20:00
woops, sntxrr ^20:00
cmn so what url are you using?20:00
stuntmachine no worries :)20:00
cmn directly to the commit?20:00
then it does look like it's a cache issue20:00
benbergman that sounds like github might cache it20:01
cmn if the repo has the same name, the github backend probably hasn't cleared it up and is serving stale data20:01
if it's that sensitive, you should contact them20:01
benbergman they say to file a ticket if the page is cached on their end20:01
→ john_anderson joined20:01
sntxrr its not *that* crazy sensitive. not like a root password.20:01
← Pistos left20:02
← juvenal left20:02
FauxFaux A password is barely sensitive at all, just change it.20:02
← ymasory left20:02
sntxrr yes, that will be done.20:02
← eletuchy left20:02
sntxrr having a hard time explaining this I guess20:03
https://github.com/rrxtns/Chillout-Tent-bot is my repo20:04
file in question is chillbot-kjc.js, which does appear to be removed20:04
← d3c left20:04
← jefferai left20:04
sntxrr the reason I can still see the sensitive data is via a comment from the fella I forked my repo from20:05
→ williambr joined20:05
sntxrr i'd link it here.. however..20:05
← Targen left20:05
→ Poapfel joined20:05
→ amrx joined20:06
SethRobertson sntxrr: Use git-filter-branch to rewrite the comment, but note that this will not affect the person you forked from of course20:06
← nrichards left20:06
cmn so you have a link directly to the commit or file version that caused it20:06
→ jefferai joined20:06
→ ping-pong joined20:06
cmn that doesn't mean you can grab it with git20:07
amrx How can I merge a git repository obtained via "git-svn clone" with a remote "bare" git repository? do I simply add the new remote and do a pull?20:07
sntxrr SethRobertson: i did this git filter-branch --index-filter 'git rm --cached --ignore-unmatch chillbot-kjc.js'20:07
SethRobertson amrx: Yes, though if you pull (meaning merge) you cannot push back to svn again20:07
sntxrr: That obviously will not change commit messages, there is a different filter for that20:08
→ Ecesis joined20:08
← Ecesis left20:08
← mmc1 left20:09
sntxrr cmn: you can still grab the info via the github website :(20:10
→ nutron joined20:10
→ JohnyBaby_ joined20:10
cmn then you should complain to them20:10
that has nothing to do with git20:10
← blacktulip left20:10
→ ipalaus_ joined20:11
amrx How can I merge changes from a remote, bare git repository with a git-svn clone and still be able to commit to the svn repository? or is this possible?20:11
sntxrr SethRobertson: luckily, I didn't "pull" to the person I forked from, so the info is not in his repo20:11
cmn: fair enough. since I'm a bit of a n00b, thought I'd see if there was something I had missed in git20:12
cmn amrx: you'd have to push all of their changes as well to the svn repo20:12
SethRobertson No luck about it. You cannot pull to someone. You can push to someone and you can ask them to pull, but that is of course different20:12
← JohnyBaby left20:12
sntxrr thats what I meant. like I said, n00berator ;)20:13
SethRobertson amrx: git cherry-pick is probably your best bet20:13
→ CannedCorn joined20:14
amrx can I check the alternate remote out into a different branch and merge things using cherrypick?20:14
SetRobertson: is that what you are suggesting?20:14
→ cjuner joined20:14
← mnemoc left20:14
zastaph is the (Tortoise)Plink mentioned as alternative to OpenSSH during Git Windows instlalation the same as the regular plink.exe from putty homepage?20:14
← williambr left20:14
→ mmc1 joined20:16
← FreeArtMan left20:16
← eidolon left20:16
→ dv310p3r joined20:16
→ teweWORK joined20:17
sntxrr contacting github support. appreciate your help20:18
→ dreiss joined20:19
← yoklov left20:19
canton7 zastaph, it might be modified slightly iirc? something to do with reducing human interaction? vague memory there... basically the same, anyway20:19
zastaph so is it embedded with git or I have to find it elsewhere? and really, which option would you recommend?20:20
for now im not going to use ssh20:20
→ beeriz52 joined20:20
← diegoviola left20:21
← Jackneill left20:21
→ yoklov joined20:21
cmn it might be bundled with tortoise, but it's definitely not embedded in git20:21
canton7 msysgit does ship with plink iirc20:21
← beeriz52 left20:22
→ lacrymology joined20:23
zastaph i got tortoiseHg installed.. probably wouldnt be a good idea to install both20:23
→ solussd joined20:23
zastaph anyway, seems openssh is the bst answer20:23
→ ymasory joined20:23
→ HyperNerdV2 joined20:24
← JohnyBaby_ left20:24
canton7 tortoisegit is a pretty bad piece of software20:24
imachuchu amrx: did you get your question answered? (I've only been tangentally listening, so before I jump in and mess things up is everything already fixed?)20:25
→ rcu joined20:25
← ipalaus_ left20:25
→ JohnyBaby joined20:25
→ berwin joined20:27
← txomon left20:30
→ d3c joined20:32
← JohnyBaby left20:34
← sjbrown left20:34
→ SegFaultAX|work joined20:34
→ JohnyBaby joined20:35
→ gerard0 joined20:35
← osmosis left20:35
→ osmosis joined20:36
→ cleardance joined20:36
cleardance how do i switch back to an older version? how can i view what verions there are?20:36
amrx imachuchu: Based on the feedback, I was going to try checking out the git remote into an alternate brach then using cherry pick to merge20:36
← hellopat_ left20:37
aried3r When I delete remote branches, is the branch really deleted, as in, space gained?20:38
kevlarman aried3r: nope20:38
→ hellopat joined20:38
imachuchu amrx: sounds like the best option. The only thing I'd add would be to make a temporary merge branch, cherry-pick/merge + rebase that branch till it's pretty, then merge/cherry pick it into the master (think of it as a staging branch for your master)20:39
→ alexukf joined20:39
← ConstantineXVI left20:39
cleardance how do i switch back to an older version? how can i view what verions there are?20:39
aried3r Thanks. So what good does deleting remote branches do? Besides not showing in the list of branches?20:40
← cannonball left20:41
imachuchu aried3r: nothing but references are deleted until 1. a commit is not referenced, 2. The commit hits the timeout limit (I think it's ~60 days), and 3. the repo is garbage collected20:41
aried3r: unless you manually git gc and remove the unreferenced commits20:41
→ eletuchy joined20:41
kevlarman imachuchu: reflog entries normally last 30 days20:42
aried3r When I 'git gc' and then push, do the other only have to pull or re-clone?20:42
← RyuGuns left20:43
→ ColKurtz joined20:44
← thebignoise left20:44
← Eridius left20:44
→ mathieui joined20:44
mathieui Hi20:44
gitinfo mathieui: welcome to #git, a place full of helpful gits. If you have a question, just ask it–somebody should answer shortly. For more information about git and this channel, see the links in the topic. It can't hurt to do a backup (type !backup for help) before trying advice you receive here.20:44
olivneh → ori-away20:44
imachuchu kevlarman: so reflog's last 30 days, do commits last 60? Alternatively just point me at the man page which lits these20:44
← Shusshu left20:45
→ JEntrep joined20:45
kevlarman imachuchu: commits don't "last"20:45
any objects that are unreferenced are fair game for git gc20:45
reflog normally causes them to be referenced for 30 days20:46
← JEntrep left20:46
mathieui I have pulled from a repo, then screwed up the merge, but merged anyway, then reverted the merge ; how do I pull again? (git tells me that everything is up-to-date)20:46
imachuchu kevlarman: ahh I see, thank you20:46
cmn unreferenced objects are left alone for a while20:46
kevlarman even after there are no branches from which they are reachable20:46
cmn: are you sure?20:46
cmn yes20:46
→ gv joined20:46
→ irqq joined20:47
cmn man git gc --prune20:47
gitinfo the git-gc manpage is available at http://jk.gs/git-gc.html20:47
imachuchu quote - "Prune loose objects older than date (default is 2 weeks ago, overridable by the config variable gc.pruneExpire). This option is on by default."20:48
→ amya joined20:48
← JohnyBaby left20:49
→ cmyk_ joined20:49
imachuchu mathieui: It sounds like git is correct, the remote data is still in your repo just not merged into your main branch. Look through your repo to find the remote, then merge that manually (with git merge)20:49
cmn mathieui: you /are/ up to date, because you already merged that branch20:50
← stuntmachine left20:50
mathieui cmn: I know that20:50
Nikoli_ → Nikoli20:50
cmn !fixup to see how to undo this20:50
gitinfo So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!20:50
mathieui I am reading http://git-scm.com/2010/03/02/undoing-merges.html right now20:50
cmn the question indicated that you did not20:50
mathieui I am aware that the commits are already in the history, but I still want them to be applied :)20:51
cmn then you shouldn't have reverted the merge20:51
mathieui I want them applied, but not the commited conflict resolution20:52
cmn if you haven't pushed yet, the best way would be to reset back to before the merge and retry20:52
← dpino left20:52
→ hwrd|work joined20:53
cleardance hmm so when i have a version in eclipse open change it after having switched branch in git, what happens a mistake?20:53
mathieui both are pushed (since the first one broke things)20:53
imachuchu mathieui: step 1. Apologize to anyone else who has touched this branch20:54
cleardance u didnt talk to me or?20:54
← john_anderson left20:54
mathieui I think I will revert the revert, and edit the diffs where the conflicts were20:54
← hellopat left20:54
→ john_anderson joined20:56
imachuchu cleardance: Inevitably there will be multiple conversations going on so if someone is writing directly to you they will (mostly/hopefully) put your name in front of their comment20:56
← rcu left20:57
→ Textmode joined20:57
← freeed left20:59
→ thebignoise joined20:59
← Peste_Bubonica left20:59
← mritz left20:59
imachuchu cleardance: I'm afraid your question isn't very clear, where are the mistakes? When you edit something in eclipse it is modifying your checked out copy. When you checkout a new commit then it changes the local checked out copy of the code/data. Do a diff to see how your checked out code differs from the commit it is based off of20:59
← kpreid left21:00
imachuchu cleardance: but a commit, once it is recorded in the repo, cannot be changed so you can't accidentally change an already recorded commit21:00
→ rcu joined21:01
→ kpreid joined21:01
→ Xander joined21:02
→ pantsman joined21:02
← pantsman left21:02
→ pantsman joined21:02
→ macabre joined21:02
← tommyvyo left21:02
← phishy left21:03
→ tommyvyo joined21:03
← tommyvyo left21:03
→ `Assassin joined21:03
← whitman left21:04
→ Sky[x] joined21:04
← Dave^| left21:05
→ mwrosen joined21:06
← Sigma left21:07
← bartek left21:08
← _iron left21:10
cleardance I don't really undertand merge, isn't it just putting two branches together to one consistent repo?21:10
I changed in master and merged master and branchX and it says up to date YEAH! but they are different21:10
canton7 cleardance, you don't merge branxhX and master by typing 'git merge master branchX'21:11
If you want to merge branchX into master, checkout master, then 'git merge branchX'21:11
merging is about taking the changes from some other branch, and incorporating them into your branch21:12
cmn cleardance: merge doesn't mean to overwrite changes21:12
cleardance so what does git merge master branchX do? merge master with master and branchX?21:12
then it should work...?21:12
canton7 it creates a merge with 3 parents: master, branchX, and the current branch21:12
cmn that checks out branchX and merges master into it21:12
hm, am I thinking of rebase then?21:13
canton7 yeah, sound likely21:13
*sounds21:13
← eguzman_ left21:13
cleardance if master and branchX are the same except a method is added in master then it would just be already up to date?21:13
canton7 of course, here, if you've got master checked out, then master and the current branch are the same21:13
← txomon|home left21:13
canton7 cleardance, ignore merges with 3 or more parents for now: they're hardly ever used21:14
cmn cleardance: it has nothing to do with the sameness21:14
→ LucasTizma joined21:14
← Tinti left21:14
cmn the commit you're trying to merge is already part of the history21:14
so your request to incorporate that commit into the current branch's history is answered with "up to date"21:15
→ mritz joined21:16
→ joaoh82 joined21:17
→ BryanRuiz joined21:17
BryanRuiz i have a file in one of my branches21:17
and i know the name of the file21:17
but i dont know the name of the branch21:17
any thoughts on how to find the branch?21:17
imachuchu cleardance: I'm confused as to what the exact situation is, you/we/I may understand things a bit better if you look at the repo and see what it says is the current state, something like: "log --oneline --graph --decorate --all"21:18
← rcu left21:18
→ titom_ joined21:18
→ Eridius joined21:18
← revagomes left21:18
canton7 BryanRuiz, git log --all -- path/to/file? Check --decorate in there as well actually21:19
*chuck21:19
BryanRuiz you amaze me. thanks canton721:19
canton7 when you've got a hash, git branch --contains <hash>21:19
np :)21:19
(that is, if --decorate doesn't show you the branch first)21:19
← mindjiver left21:20
← mwrosen left21:21
← mikaello left21:21
→ solussd_ joined21:21
← xou_ left21:23
→ ipalaus_ joined21:24
← mishok13 left21:24
→ sebastorama joined21:26
← mmc1 left21:27
← airborn left21:28
← d3c left21:29
→ kcj joined21:30
→ mmc1 joined21:30
← TheHaven left21:31
→ GameGamer43 joined21:32
zastaph a dev repo and a remote repo are just 2 project folders right? I need to set a "remote add origin" for all my separate projects, not just once for the whole repo?21:32
← khmarbaise left21:32
← toabctl left21:33
→ toabctl joined21:33
zastaph or maybe origin is not nescessary, to replace it with just the local folder name21:34
← h0h0h0 left21:34
← hwrd|work left21:34
cmn what's the difference between "whole repo" and "seaparate projects" here?21:34
they're all just repos21:34
← amrx left21:34
← pantsman left21:34
→ mneorr joined21:34
← DoniRS left21:35
← mneorr left21:35
→ eidolon joined21:35
→ shruggar joined21:35
zastaph mmm right, and remote add is what links the dev and remote repo21:35
cmn yes21:35
zastaph just i've seen examples of both "remote add local_repo" and "remote add origin", not sure about the difference21:36
presuming you are in the local repo folder when running these21:36
← shruggar left21:36
cmn the difference is the name of the remote21:37
zastaph which one would I use21:37
cmn there is no technical difference, though git will at times default to origin if you omit data21:37
you should use the names that make sense for you and your project21:37
← mmc1 left21:37
← ph^_ left21:38
cmn origin is the default for the remote from which you cloned this repo21:38
zastaph well, my local repo has a folder name, and so has my remote repo .. so the "origin" part is just an alias?21:38
cmn yes21:38
each directory has a path, but that has nothing to do with the name you give it, unless you make it that way21:38
zastaph i made a bare remote repo, and then just git init my dev repo, then i guess remote add origin is most appropriate21:38
← gchristensen left21:39
cmn yes, though a clone will do all this for you21:39
zastaph well the remote repo is bare, no code there yet21:40
cmn the setting up doens't depend on the repo being populated21:41
→ thesearchengine joined21:42
← thesearchengine left21:42
→ aboudreault joined21:42
← alexisabril left21:42
→ alberto56 joined21:44
← irqq left21:46
← alberto56 left21:46
→ ConstantineXVI joined21:46
→ btanaka joined21:46
← luxurymode left21:46
← pandeiro left21:46
← loco left21:47
← solussd_ left21:48
zastaph git remote add origin //server/repo/test.git worked, but git push origin master gave: error: src refspec master does not match any. error: failed to push some refs to '//server/repo/test.gif'21:49
→ nrichards joined21:49
← cyb3r3li0g left21:51
cmn first of all, that path is suspicious21:51
zastaph its a SMBFS/CIFS21:51
cmn and the error happens because you don't have a master branch yet21:51
so it's \\server\repo ?21:51
zastaph well Git Bash says (Master) after I did git init21:51
→ alberto56 joined21:51
cmn that doesn't mean the mater branch exists, it means it's the current/active one21:52
have you committed?21:52
zastaph no cd //server/repo/test.git takes me there, however cd //server doesn't work21:52
no i didn't commit :|21:52
the docs didnt say so :)21:52
cmn what docs didn't say what?21:53
← dcarper left21:53
cmn if you don't commit, you don't have anything to push21:53
zastaph i just followed them blind :)21:53
cmn and the master branch is unborn21:53
if anything says git init should be followed by push, it's complete nonsense21:53
and I'd be weary of the rest of it21:53
SethRobertson zastaph: The docs to read are !docs, especially !book21:54
gitinfo zastaph: There are several good books available about git; 'Pro Git' is probably the best: http://progit.org/book/ but also look at !bottomup !cs !designers !gitt !vcbe and !parable21:54
SethRobertson cmn: I am wary of people who say they are weary21:54
→ mmc1 joined21:54
cmn I am able to be worn, what's your problem with it?21:54
;)21:54
→ cyb3r3li0g joined21:54
zastaph ouch, VIM is default editor, gotta remedy :)21:55
SethRobertson .trigger_edit docs @!doc21:55
gitinfo SethRobertson: Okay.21:55
SethRobertson !docs21:55
gitinfo [!doc] A list of useful documentation of all kinds is here: http://git-scm.com/documentation -- or try the wiki at http://git.wiki.kernel.org/. Also try typing "!book" "!cs" "!bottomup" "!parable" "!best_practices" or "!vcbe" or "!designers" here in IRC. !book is probably the most helpful.21:55
← cjuner left21:55
← alberto56 left21:56
← cyb3r3li0g left21:57
← fsvend left21:58
→ walkerneo joined21:58
← toabctl left21:58
walkerneo Hell, I need some help if anyone is here21:58
*hello21:58
cmn !ask21:58
gitinfo Yes, it's okay to ask questions here.... in fact, you just asked one! ;-) Pretty much any question is fine. We're not terribly picky, but we might be asleep. Please be patient and you should get an answer soon.21:58
← kermit left21:59
walkerneo I'm new to version control and have little experience with linux or shell at all for that matter21:59
anyway, I'm trying to backup my site on bitbucket.org21:59
← Destos left21:59
walkerneo but I'm getting a fatal error about the certificate?21:59
← MANCHUCK left21:59
→ krzkrzkrz joined22:00
→ kermit joined22:00
walkerneo Says:22:00
cmn that will probably be from curl, telling you that the site isn't providing a proper HTTPS certificate22:00
walkerneo error: error setting certificate verify locations:22:00
..22:00
oh, so what do I do?22:01
← d0k left22:01
cmn install some certificates, I'm guessing22:01
→ d0k joined22:01
walkerneo where? how?22:01
← imachuchu left22:01
cmn on your computer, using your package manager22:01
SethRobertson walkerneo: You can also use `GIT_SSL_NO_VERIFY=1 git clone <url>`22:01
cmn on Debian and look-alikes the ca-certificates package usually helps22:01
← krz left22:02
krzkrzkrz → krz22:02
cmn that would also work22:02
walkerneo Well, I'm doing this on my webhost's server through ssh22:02
I'll try that command22:02
cmn then you should complain, because bitbucket's cert seems to be pretty valid from here22:02
SethRobertson Then `cd repo; git config http.sslVerify false` (or perhaps 0)22:02
← iamjarvo left22:02
→ Zenith77 joined22:03
← alexim left22:03
SethRobertson But getting good certs installed would be best22:03
walkerneo It's uploading after that command SethRobertson :D22:03
→ LazyTown joined22:03
SethRobertson Or even downloading22:03
← thebignoise left22:04
LazyTown if I merged a branch from a remote repo, which removed and stopped tracking an important file that is in my .gitignore file, how can I get that file back?22:04
*remote repo* is actually *remote branch*22:04
SethRobertson LazyTown: !fixup22:04
gitinfo LazyTown: So you lost or broke something or need to otherwise find, fix, or delete commits? Look at http://sethrobertson.github.com/GitFixUm/ for full instructions, or !fixup_hints for the tl;dr. Warning: changing old commits will require you to !rewrite published history!22:04
SethRobertson Note the file being in your .gitignore is irrelevant to the problem22:04
→ ispeak_ joined22:04
→ gchristensen joined22:05
← smw_ left22:05
cmn why is an important file in a list of files that's fine to remove?22:05
→ smw joined22:05
LazyTown cmn, its a parameters file that holds database connection info22:05
SethRobertson He said the file was tracked, so it was irrelevant that it was in .gitignore22:05
You should be using !config22:05
gitinfo [!configfiles] the recommended way is to change your application so that you can (perhaps optionally) store site-specific configuration/data in separate files, then keep those files out of version control. Still, see https://gist.github.com/1423106 for ways people have worked around the problem.22:05
walkerneo SethRobertson, I would be able to do this with cron right? And is there a way to do the same with my sql databases?22:06
SethRobertson walkerneo: See !backup link for information about automation.22:07
gitinfo walkerneo: Taking a backup of a git repository is always a good idea, especially when taking advice over IRC. Usually, the best way to TACTICALLY back up a git repo is `git clone --mirror`. However, some unusual maintenance might require `tar cf repo-backup.tar repodir`. Testing in a clone is also an excellent idea. See also http://sethrobertson.github.com/GitBestPractices/#backups22:07
cmn though putting sql dumps into a repo is generally not what you want to do22:07
SethRobertson sql is outside the scope of git, but yes, you can use your sql server's shell (psql mysql sqlplus etc) or table backup command to backup your tables22:08
LazyTown SethRobertson, Which command do I need to pay attention to on this page you sent me?22:08
← eidolon left22:08
← mathieui left22:09
LazyTown I think I found it.22:09
SethRobertson Ah, the sample commands is only if you have ssh access. Hmm. I guess you just want to `git clone --mirror URL` and then in cron `(cd /repo; git fetch)`22:09
The last line does the update. The lines before it are just looking for unmirrored repos.22:10
← DuzyZiemniak left22:11
walkerneo SethRobertson, I can't find anything about automation in the link22:11
SethRobertson There is a sample shell script in the #backup section22:12
But as I said, everything but the last line is just there to probe for unmirrored repos and requires ssh access, and you can run a simplified version of the last line as I mentioned above.22:13
→ bnjmn joined22:14
← juanfer left22:14
walkerneo I don't really understand what any of that is. I never learned any shell script22:14
I'd rather not just copy and paste things I don't understand22:14
bnjmn i have a merge conflict and i want to resolve it by choosing all of the HEAD changes, is there any way to do this automatically?22:14
← amya left22:15
→ Slix` joined22:15
← Element9 left22:15
← thiago left22:17
← notfunk left22:18
→ notfunk_ joined22:18
← iwoj left22:19
Sky[x] i read man and i still dont understand what mean upstream ?22:19
→ yhager joined22:19
← axl_ left22:20
→ vitor-br joined22:20
cmn upstream is a software development term (mostly in OSS) to mean the primary project, to differentiate it with whatever you or your team are doing22:22
in the git manpages, the term "upstream branch" is used to mean the branch on which yours is based22:22
← twisted` left22:22
→ yashshah joined22:23
← scientes left22:24
← fisted left22:24
→ twisted` joined22:25
← titom_ left22:25
← nadavoid left22:25
→ alberto56 joined22:25
← Davey left22:26
→ fisted joined22:27
← beatak left22:27
← macabre left22:28
→ orngchkn joined22:28
← orngchkn left22:30
→ iwoj joined22:30
← harish left22:33
← tvw left22:33
← shoerain left22:34
← cleardance left22:34
← yoklov left22:34
← zastaph left22:35
sureaint → shoerain22:36
walkerneo I added a cron job: cd /home/walkerneo/public_html;git commit -m "Daily backup";GIT_SSL_NO_VERIFY=1 git push -u origin master22:37
will that work?22:37
bob2 why are your certs broken22:37
walkerneo bob2, Is that you from PYthon :D ? And idk22:37
bob2 yes22:38
→ alexim joined22:39
← aboudreault left22:39
← umc left22:39
→ Destos joined22:39
raek walkerneo: you need to add the changes too22:39
walkerneo so git add . ?22:40
← chrisgeorge left22:44
→ Targen joined22:44
← kpreid left22:45
→ kpreid joined22:45
→ thebignoise joined22:46
← twisted` left22:46
← bgerber left22:46
→ babilen joined22:47
babilen → Guest9492622:47
→ yosinof_ joined22:47
→ bgerber joined22:48
bowarv → borwar22:49
← adamcooke left22:49
← stkrzysiak left22:50
← BryanRuiz left22:50
→ BryanRuiz joined22:53
← yashshah left22:53
→ yashshah__ joined22:53
← yashshah__ left22:54
← thelazydeveloper left22:54
→ yashshah__ joined22:55
← heroux left22:55
← dfr|mac left22:56
→ cryptopsy joined22:56
cryptopsy can i "git clone https://" or does it only take the form "git clone git://..."22:56
?22:56
→ heroux joined22:57
← BryanRuiz left22:57
→ Alumin joined22:58
← kevwil left22:58
→ machty joined22:59
← mikepack left22:59
cmn if the other end has a web server configured to call git, you can use https22:59
and if the other end has ssh set up, you can use ssh://22:59
→ mikepack joined22:59
Alumin This question is about "git commit -av". I noticed that once, I typed a commit message and I forgot to remove the comments and diff output...but they were still pruned from the commit message when it was submitted. This is what I want it to do, but I'm wondering what the exact algorithm is as far as how git determines what to prune from a commit message...22:59
← yashshah__ left22:59
cmn commands and diffs23:00
→ BryanRuiz joined23:00
cmn comments23:00
← Guest94926 left23:00
Alumin but, what if I (for some strange reason I can't think of now) wanted to include a diff in my commit message?23:00
cmn you do what any sane person would do and indent it23:00
same as you'd do for any piece of code23:01
← d0k left23:02
Alumin well, I'm not sure about "sane" 'cause when you really come down to it I should theoretically be able to include any data I want in my commit message without having to modify it to suit the VCS. That said...thanks for the info. :) So it actually does some kind of parsing to detect the output of diff and surgically remove it? Obviously comments would be easy but finding diffs seems overkill. :)23:03
→ Lord_DeathMatch joined23:03
cmn commit messages aren't where you dump data, that's what files are for23:03
← mikepack left23:03
← martinjlowm left23:03
Alumin fair enough23:03
← mmc1 left23:04
→ derwin joined23:04
derwin why do "git diff" and "git diff HEAD" have different output?23:05
cmn and it's primarily for humans, so you don't need to be so strict about it, and it's general convention to indent anything that's quoting something, which includes the diff23:05
← Twey left23:05
Alumin anyway, really the important part is just that I know what the rules are so when I do have a complicated commit message or something I don't get a surprise23:05
→ Twey joined23:05
cmn derwin: because they're comparing different things23:05
derwin cmn: what does "git diff" compare against implicitly?23:05
→ benjo2 joined23:05
cmn the idnex23:05
derwin assumed it was HEAD23:05
cmn idnex*23:05
index, I mean23:05
derwin so in the case of a file add, why does that not show anything?23:05
cmn because it's in the index23:06
that's what the add did, it put it in the index23:06
derwin oh, that makes sense.23:06
and git diff HEAD is "show me what the result of the pending commit would be, if diffed against HEAD"23:07
which is both index and content23:07
cmn no23:07
diff --cached shows you that23:07
diff HEAD compares the working directory with HEAD23:07
derwin ah!23:07
I looked for --cached but did not find, awesome, thx.23:07
cmn it's also aliased to --sated23:08
--staged23:08
I can't type tonight23:08
← Chillance left23:09
derwin yeah, just read that in the man page23:09
→ cyphase joined23:09
← bglusman left23:09
derwin a friend suggests that "git add -N newfile.txt" might make "git diff" work as I expect?23:10
(as it puts a null file in the index, I guess?)23:10
cmn -N marks it as intent-to-add23:11
→ kaen joined23:11
cmn it lets you do things like add -p before committig23:11
derwin sweet, I learn new things every day23:11
and now am less likely to be calle a moron by my peers!23:11
cmn: thx!23:11
→ metalfan_ joined23:13
← metalfan_ left23:13
← cmyk_ left23:14
→ linusoleander joined23:15
kenperkins is there a way to easily identify which branches are fully merged into a specific branch23:17
← Balreign_ left23:17
← EToS left23:17
← lonewulftp left23:18
cmn man git branch --merged23:19
gitinfo the git-branch manpage is available at http://jk.gs/git-branch.html23:19
← fly9 left23:19
← yoh left23:20
→ WillMarshall joined23:21
← adhawkins left23:21
← derwin left23:21
← tatsuya_o left23:22
cryptopsy cmn: tnx23:23
← cryptopsy left23:23
walkerneo I'm trying to automatically backup my site with git and bitbucket23:23
What do I do when it asks for the password?23:23
FauxFaux Configure keys properly.23:24
cmn you make it so it doesn't23:24
either set up ssh or store the password in .netrc23:24
walkerneo I don't really know how to do either23:24
→ txomon|home joined23:24
cmn no better time to learn about ssh23:25
→ robk joined23:26
← marhaban left23:26
walkerneo Alright, I'll learn about ssh23:26
also though, what's the command to remove any files that were remove?23:27
← intransit left23:27
robk question, i'm in the process of converting over my svn w/ commit history into git, i've already created the authors.txt & imported the base stuff, but Now i'm confused when i look at http://stackoverflow.com/questions/79165/how-to-migrate-svn-with-history-to-a-new-git-repository to help me complete the process.23:27
walkerneo If I add files I can do git add .23:27
but what if I remove files23:27
would git rm .23:28
just remove all the files?23:28
← john_anderson left23:28
→ SeaSalt joined23:28
← Alumin left23:29
← nrichards left23:29
← tqmz left23:30
← Lord_DeathMatch left23:30
cmn yes, because you asked it to23:31
→ thursday joined23:31
→ marhaban joined23:32
→ frogonwheels joined23:34
← TeckniX left23:34
← vitor-br left23:34
← gv left23:35
← arietis left23:35
→ iamjarvo joined23:38
← dreiss left23:38
walkerneo How can I only remove the files from the project that were removed from the directory?23:39
← iwoj left23:39
walkerneo sorry, from the repository that were removed from the project23:39
← mtytel left23:41
→ EToS joined23:41
cmn clean23:42
and you don't remove them from the repository, you remove them from the working directory23:42
→ txomon joined23:43
← teweWORK left23:43
← dv310p3r left23:44
← iamjarvo left23:45
robk I just followed this tutorial on setting up an svn-rebase to git from this site: http://blokspeed.net/blog/2010/09/converting-from-subversion-to-git/ I now want to push my local svn base copy to my remotely hosted unfuddled git, any idea on how to go about this?23:46
FauxFaux What do you mean by "local svn base copy"? How did you get it? Why is it not like the git repo?23:46
walkerneo cmn, no I mean the other way around. I remove a file from the cwd, do I need to do something so that git doesn't think it's missing?23:47
→ mtytel joined23:47
← JeroenDeDauw left23:47
→ krzkrzkrz joined23:48
cmn bring it back?23:48
git rm removes it from the working dir and the index23:48
so it will be removed from the next commit23:48
← krz left23:50
krzkrzkrz → krz23:50
← DofLeFou left23:51
← alexukf left23:52
walkerneo I mean if I'm making daily backups23:53
→ dcarper joined23:53
walkerneo and I want removed files to be kept out of the daily commit23:53
what do I do?23:53
I know they're not actually removed23:53
→ cagedwisdom joined23:53
walkerneo but I know if I remove them and don't do anything, git thinks they're missing right?23:53
cmn git add -A updates the index with the current state of the working directory23:54
← marc_kdab left23:54
kpreid walkerneo: sounds like you want git add -A23:54
walkerneo sounds like I do23:54
thanks :D23:54
cmn what do you mean if you don't do anything?23:54
← yosinof_ left23:54
cmn status will report them as missing, but git doesn't generally care about the worktree whenc ommitting23:54
→ yosinof joined23:55
walkerneo Yeah, that's what I mean, think they're missing when I acutally just removed htme23:55
So to not need a password with ssh, what do I need to do?23:55
I'm connected to my server via ssh and need to set its ssh...23:55
cmn you tell ssh to use a key that doesn't need a password23:55
← kenperkins left23:56
walkerneo By key, are we talking RSA? And would that mean the password is already in the key?23:57
I'm confused..23:57
→ efreese joined23:57
← TomyLobo left23:57
cmn RSA is the usual method, yes23:57
if you don't set a password, then there is no password in the key23:57
which means ssh can use it directly23:57
← dcarper left23:57
← tewecske left23:57
← rubyruy left23:58
← ben225 left23:58
→ ben225 joined23:59
← ben225 left23:59
haarg the password isn't in the key; the key is used instead of the password.23:59
→ beneggett joined23:59
cmn the password we're talking about is a different one, the one used to unlock the key23:59
though it's generally called a passphrase, isn't it?23:59

Logs Search ←Prev date Next date→ Channels Documentation