| 2007-05-11 |
| gitster changed the topic to: 1.5.2-rc3/1.5.1.4 | Everyone asleep or clueless? Try [email@hidden.address] | Channel log http://colabti.de/irclogger/irclogger_log/git". | 00:05 |
|
motion
| damn ive never seen conversations carried on for so long on a mailing list :/ | 00:16 |
|
gitster
| which topic? | 00:20 |
|
motion
| gitster: index handling | 00:25 |
| → yashi joined | 00:25 |
|
DrNick
| you must not read interesting mailing lists | 00:26 |
|
motion
| heh | 00:26 |
|
| gitster agrees. | 00:26 |
|
DrNick
| lkml has a ~400 message thread on fallocate() | 00:26 |
|
| it will not die | 00:26 |
|
motion
| right now im only on git and hg | 00:26 |
|
| lkml ? :x | 00:26 |
|
DrNick
| ok, only 177 messages | 00:26 |
|
| but, still | 00:26 |
|
motion
| heh | 00:26 |
|
DrNick
| linux-kernel mailing list | 00:26 |
|
motion
| mm | 00:28 |
|
| i think plumbing subprojects is exactly what im looking for | 00:28 |
|
| grrrr 500 revisions to go :( | 00:31 |
| → sgrimm joined | 00:36 |
| → Tali_ joined | 00:37 |
| → tokkee_ joined | 00:37 |
|
motion
| can git svn be told to look for branches in 2 locations | 00:40 |
|
| that are like apart of eachother kid onf | 00:40 |
|
| this chick at work insisted on having branches/ AND branches/releasebranches/ | 00:40 |
|
| dont ask me . | 00:40 |
|
| and its screwing up my git svn fetch :( calling the branch tags/releasebranches | 00:42 |
|
| so i really have 1 branch with 2 branches in it | 00:42 |
|
| unless i could like go through git's history and change where it branches | 00:42 |
|
| mm nm | 00:43 |
|
clee
| motion: uh. your repo is broken. | 00:43 |
|
motion
| clee: ... | 00:43 |
|
DrNick
| no, that's perfectly acceptable SVN behavior | 00:44 |
|
motion
| thnx for telling me the obvious . | 00:44 |
|
clee
| DrNick: well, right, but it's stupid. | 00:44 |
|
motion
| drnick: so can i tell git svn to look in branches/releasebranches/* and branches/* (ignoring releasebranches of course) for branhces ? | 00:44 |
|
clee
| sorry, motion. your repo is *stupid* | 00:44 |
|
motion
| its organized ODDLY | 00:44 |
|
clee
| :) | 00:45 |
|
| I don't know how to handle that case, sadly | 00:45 |
|
motion
| i mean even if i decide to branch it now it just breaks it when i try to git-svn fetch later | 00:45 |
|
clee
| it's one of the things that keeps me from implementing proper branch support with KDE | 00:45 |
|
motion
| so its almost useless to convert it to git | 00:45 |
|
| clee: i personally dont like it | 00:46 |
|
| err hmm | 00:46 |
|
| how "regexy" can i get in [svn-remote] | 00:46 |
|
| like: branches = base/branches/{releasebranches/*,*} or soemthing | 00:47 |
|
| but thats still gonna see releasebranches as a branch | 00:47 |
|
| motion cries | 00:47 |
|
motion
| i shoulda just ignored the branch in branches/ | 00:48 |
|
| yeah i am gonan do that, its a blank branch anyway, nothing was done to it :D | 00:49 |
|
| guess i get to re svn init/fetch tomorrow | 00:49 |
|
DrNick
| motion: you can't get regexy at all, sadly | 00:49 |
|
| or globby, for that matter | 00:49 |
|
motion
| heh | 00:50 |
|
DrNick
| * is a special case, not a pattern-matching operator | 00:50 |
|
motion
| drnick: so no way to do that | 00:50 |
|
| yeah | 00:50 |
|
| drnick: i was gonna guess that | 00:50 |
| → sgrimm_ joined | 01:08 |
| ← dduncan left | 01:26 |
| → z3ro joined | 01:46 |
|
motion
| yey | 02:01 |
|
| git svn fetch just segfaulted | 02:01 |
|
aeruder_
| yea, i'm hitting some issues of my own, git-format-patch really doesn't like one of my commits (and is dumping out an empty .patch file) | 02:05 |
|
motion
| sweet | 02:05 |
|
| i need to go get some beer soon | 02:06 |
| → sgrimm joined | 02:17 |
|
motion
| can i tell svn fetch to treat base as a specific revision vs downloading * | 02:40 |
|
| this repo has looked pretty damn ugly in the past thatd id just avoid fetching completely | 02:40 |
|
| is there an easy way to see where the branch branched form | 02:44 |
|
| from* | 02:45 |
|
Randal
| the easiest way is to visualize it with gitk --all | 02:48 |
|
gitster
| Hi, Merlyn. | 02:49 |
|
Randal
| hey g | 02:50 |
|
gitster
| gitk is a good suggestion. | 02:50 |
|
| If motion does only straight line development (i.e. has master, branches off mine with "git branch mine master", and builds on mine without merging), then a quick answer would be "git merge-base master mine". | 02:51 |
|
| That would give where 'mine' forks off of 'master'. | 02:51 |
|
| If motion did any merge into 'mine', the question stops having a clear meaning, other than "look for the reflog entry that records the creation of the branch". | 02:52 |
|
| Suppose I fork off of 'master' at v1.5.2-rc3 today to create a topic branch 'A', accumulate a few commits. | 02:54 |
|
Randal
| first, make sure motion is still listening. :) | 02:54 |
|
| unless you like talking to dead air. | 02:54 |
|
| I've noticed people like him tend to do drive-by questions | 02:54 |
|
gitster
| I never type anything here to help only one person. That is a waste of my time. I primarily type here for irclogger so that people can find it later. | 02:55 |
|
Randal
| cool! | 02:55 |
|
| me too, but I like to make sure I'm still solving at least *one* person's problem | 02:55 |
|
gitster
| But somebody else forked off of my 'master' a few weeks ago, perhaps at v1.5.2-rc1, and had a good development on it, which my topic A could use. I merge that to A. | 02:55 |
|
| Randal sleeps | 02:56 |
|
gitster
| Then keep building on my topic branch A. Visualizing it in gitk would make it clear why "where does branch A forked" question does not have a clear meaning. | 02:57 |
|
| Especially after that somebody else merged my topic A back to his branch. | 02:57 |
| → rAndMer joined | 03:16 |
| ← rAndMer left | 03:29 |
| → duncanm joined | 03:46 |
|
duncanm
| is there a fast way to import a cvs repo to git? | 03:51 |
|
spearce
| use fast import and fromcvs ? | 03:52 |
|
duncanm
| is there a guide for how to use it? | 03:54 |
|
spearce
| i think i've only seen notes on it on here, maybe you can search the irc logs. | 03:55 |
|
cehteh
| first of all work with a local CVS repo .. copy it .. remote sux | 04:00 |
| → albertito joined | 04:02 |
| → dduncan joined | 04:04 |
|
duncanm
| cehteh: okay, got that | 04:08 |
|
| cehteh has some git-cvsimport running .. initial import wasnt that fast .. but works relyable | 04:08 |
|
duncanm
| how do i tell git-cvsimport about my local mirror? | 04:09 |
|
spearce
| cehteh: if you want fast, use something that is backed by fast-import. ;-) | 04:10 |
|
cehteh
| spearce: i dont need it fast .. it now runs regulary by crontab | 04:11 |
|
spearce
| heh. | 04:11 |
|
cehteh
| git-cvsimport -o master -k -i -m -v -s . -A ../AUTHORS -d$(readlink -f ../) yap | 04:11 |
|
| mhm ugly? :) | 04:11 |
|
| duncanm: you see the -d option .. just dir to local CVS | 04:15 |
|
duncanm
| yeah | 04:16 |
|
| i thought i had to do something like :local: or something | 04:16 |
|
cehteh
| just dir suffice | 04:18 |
|
duncanm
| is it useful to have -m? | 04:18 |
|
cehteh
| dunno .. i just added is, maybe itd useful and it runs with that since some time now ... | 04:19 |
|
| whats -m for? :) | 04:19 |
|
duncanm
| Attempt to detect merges based on the commit message. This option will enable default regexes that try to capture the name source branch name from the commit message. | 04:19 |
|
cehteh
| well i think on the yap tree merges never happend | 04:20 |
|
| its rsynced CVS from sourceforge | 04:20 |
|
| mhm :local: is optional i think | 04:21 |
|
| well .. works for me as is :) | 04:21 |
| → rkaway1 joined | 04:25 |
|
duncanm
| if i made some changes to my repo, and i haven't made a commit and just wanna revert all of it, what do i do? | 04:44 |
|
spearce
| if you want to chuck all of your uncommitted changes: git reset --hard. but that chucks, with no warning! ;-) | 04:45 |
| → sgrimm joined | 05:00 |
| → masterdriverz joined | 05:42 |
| → robtaylo1 joined | 06:06 |
|
motion
| shit ! | 06:16 |
|
| can i umm | 06:16 |
|
| remove the last two commits :x | 06:16 |
|
dilinger
| git-reset --hard <sha1> | 06:27 |
|
| where the <sha1> is the commit that you want to be the new HEAD | 06:27 |
|
motion
| kk | 06:28 |
|
| do i need to gc --prune then ? | 06:29 |
| → mithro joined | 06:30 |
|
mithro
| hi! would adding git-cvsserver to git-shell be evil? | 06:31 |
|
| i specifically want users to be able to push/pull via both git and cvs | 06:31 |
|
| preferably without being able to break into a shell | 06:31 |
|
motion
| gitster: yeah , i understand what you mean, your suggestion is exactly what i wanted though, the merge-base thnx :D | 06:35 |
|
| randal: i backlog when asking a question and dont get the answer right away. i appreciate gitster's response very much :D | 06:36 |
| → aroben joined | 06:54 |
| → dedekind joined | 06:59 |
|
motion
| stupid question but it goes through my head: | 06:59 |
|
| is there any reason one should not use git for small projects | 07:00 |
| → jeremyb joined | 07:08 |
| → strangy joined | 07:08 |
|
strangy
| mornin' | 07:08 |
| → mithro joined | 07:20 |
| → etnt joined | 07:25 |
|
etnt
| I read somewhere that it is possible to invoke an external diff pgm (e.g kdiff3) when merging, how do i do that ? | 07:31 |
|
pigeon
| i have git cloned a repo, then i branched it, and i have made changes to this local branch, but i don't want to commit any of these changes yet. now there are new changes from the original repo, how can i fetch those and have those changes into my new local branch? | 07:33 |
|
gitster
| etnt: I would suggest to study t/t6026-merge-attr.sh and do something similar. | 07:33 |
|
etnt
| gitster: thanx, but where do I find that script ? | 07:35 |
|
gitster
| In the t/ directory. | 07:35 |
|
| That's where all the tests live. | 07:35 |
|
etnt
| in the git repos ? | 07:36 |
|
gitster
| Yup. You are talking about per-path custom merge, so you should be using git built from the source ;-). | 07:36 |
|
| etnt will start tracking the git repos | 07:37 |
|
gitster
| ... because no packaged version has that feature yet. It will be in 1.5.2 and we are in the stabilization period for it. The more testers we have, the better ;-). | 07:38 |
|
etnt
| ok, btw saw that there is a 'git-mergetool', is it alive ? | 07:38 |
|
motion
| mm log --first-parent doesnt work how i expected | 07:39 |
|
gitster
| 've never used it in the real-life project, but haven't heard complaints about it either, so people must be happy with it. | 07:39 |
|
motion
| is there an easy way to git-log to stop where to branch was last merged | 07:39 |
|
| like svn log --stop-on-cop | 07:40 |
|
| y | 07:40 |
|
| mm come to think of it i really do want git on solaris :x | 07:40 |
|
| im gonna play with that more tomorrow night | 07:40 |
| ← dedekind left | 07:42 |
| → segher_ joined | 07:48 |
|
motion
| how do i tell git i fixed a conflict | 07:49 |
|
| git update-index --remove | 07:51 |
|
| :D | 07:51 |
| → qiyong joined | 07:52 |
|
qiyong
| how to show a commit given the commit ID | 07:52 |
|
mithro
| is there a script which will fix permissions after a push on a shared repository | 07:54 |
|
pigeon
| qiyong: git show <commit id> is one way | 07:55 |
|
| g'day mithro | 07:55 |
|
qiyong
| > git-show 9185cfa92507d07ac787bc73d06c42222eec7239 | 07:55 |
|
mithro
| hey pigeon | 07:55 |
|
qiyong
| pigeon, this commit makes me trouble | 07:56 |
|
| is it by purpose? my box no longer power off after hibernate (off topic) | 07:57 |
|
motion
| git can be confusing at times :x | 07:58 |
|
| wtf | 08:00 |
|
| ok that didnt work | 08:01 |
|
qiyong
| pigeon, is it you that kernel hacker nick pigeon? | 08:01 |
|
| nikk | 08:02 |
|
pigeon
| qiyong: probably not, is there a kernel hacker named pigeon? | 08:04 |
|
qiyong
| pigeon, not the exactly spelling, but similar | 08:05 |
|
mithro
| ahh, it appears I should have made the repositories with --shared=group | 08:08 |
|
motion
| mithro: thnx. that just helped me a bit too :D | 08:15 |
| → ofri joined | 08:19 |
|
motion
| mmm | 08:21 |
|
| git reset is cool | 08:21 |
| → johnw joined | 08:26 |
| → Oejet joined | 08:29 |
| → johnw_ joined | 08:41 |
|
motion
| is there a way to have gitk show multiple branhces, so the very top would show 2 paths instead of just one | 08:43 |
| → nud joined | 08:53 |
|
tronical
| motion: try gitk --all | 08:55 |
| → nud_ joined | 08:58 |
| → crich1999 joined | 08:59 |
|
johnw_
| I have a git repository that I push to via ssh; I'm the only one who pushes there, and that's all it's there for. For some reason, after I a "git-svn dcommit" on my repository, I can no longer push to my git-over-ssh repository without using --force. Anyone know why git-svn is changing the HEAD tag enough to make my ssh repository think that it's no longer a direct descendent? | 09:12 |
|
kha
| johnw_: When you dcommit, git-svn reimports the commits from Subversion and points HEAD at the new commits. | 09:13 |
|
johnw_
| ahhh | 09:14 |
|
kha
| johnw_: I think the reason is that git-svn needs to add metadata to the commit comment. | 09:14 |
|
johnw_
| so is doing git push -f the right thing after that? | 09:14 |
|
kha
| johnw_: And obviously it needs to mangle the author/committer names too. :-) | 09:14 |
|
| johnw_: Yes, it is. | 09:14 |
|
johnw_
| thanks | 09:15 |
|
kha
| johnw_: Try tagging just before dcommitting, and view the result with gitk. | 09:15 |
|
| johnw_: You'll see both the old and the new commits that way. | 09:15 |
|
| johnw_: Hmm. I guess an even more fundamental reason is that the Subversion repository could have gained new commits in addition to those you dcommitted. And of course it could have a commit hook that rewrites your commits. So there's really no choice but to re-import them. | 09:17 |
|
johnw_
| ah, I get it, it's the re-importing that happens after the dcommit that's messing with me; that makes sense I guess | 09:17 |
| johnw_ → johnw | 09:18 |
|
kha
| johnw: Exactly. It spooked me too at first, until I got it. | 09:19 |
|
| johnw: Basically, dcommitting is like mailing patches to upstream and then pulling from them; most of the time your commits go thorugh unchanged, but even so they'll always have different commit hashes than the ones you generated. | 09:20 |
| janm → nand | 09:20 |
| nand → xor | 09:20 |
| → chris2 joined | 09:20 |
| xor → xnor | 09:21 |
| xnor → janm | 09:21 |
| → janm joined | 09:22 |
| → jeffpc joined | 09:22 |
| → hein joined | 09:49 |
| hein → Sho_ | 10:05 |
|
anders_
| Anyone knows what "Incomplete data: Delta source ended unexpectedly at /usr/bin/git-svn line 2975" means? | 10:19 |
|
kha
| anders_: Hmm. Sounds like git-svn thinks that the svn server stopped in mid-sentence. But I have no idea really. | 10:22 |
| → robfitz joined | 10:23 |
| → Romster joined | 10:34 |
| ← dduncan left | 10:41 |
| dwmw2_gone → dwmw2_LHR | 10:43 |
| → meyering joined | 11:00 |
| → johnw joined | 11:31 |
| → ofri joined | 11:33 |
| → Romster joined | 11:49 |
| → ferdy joined | 12:09 |
| → Sho_ joined | 12:10 |
| → robewald joined | 12:39 |
|
robewald
| hello, is there a recipe to diff to trees with a diff visualizer like kompare? git diff branch1 | kompare -o - is not very satisfactory... because the local files are not taken into account | 12:41 |
| dwmw2_LHR → dwmw2_gone | 12:46 |
| → lcapitulino joined | 12:48 |
|
kha
| robewald: I don't think it's a one-liner. External diff tools typically expect files (or directories) in the filesystem as arguments, so you'd have to check out the args, run the diff tool, then wipe the args. | 12:50 |
| → jasam joined | 12:51 |
|
halfline
| clee: yea | 12:53 |
| → oneman joined | 13:33 |
| → GyrosGeier joined | 13:38 |
| → orospakr joined | 13:41 |
| → Romster joined | 13:41 |
| → sgrimm joined | 13:57 |
| ← duncanm left | 14:10 |
| ← etnt left | 14:14 |
| robtaylo1 → robtaylor | 14:27 |
| → etnt joined | 14:32 |
| → etnt` joined | 14:48 |
| → rphillips joined | 14:53 |
| jwb_gone → jwb | 15:02 |
| → albertito joined | 15:04 |
| → krh joined | 15:10 |
| → mneisen joined | 15:23 |
| → sgrimm joined | 15:32 |
| → alley_cat joined | 15:45 |
| → tcoppi joined | 15:46 |
| → hackbert-hh joined | 15:49 |
| → dirker_ joined | 15:57 |
| → Oejet joined | 16:04 |
| → Sho_ joined | 16:30 |
| → lyakh joined | 16:36 |
| → puzzels joined | 16:38 |
| → sgrimm joined | 16:41 |
| → bronson joined | 16:57 |
| → lyakh joined | 17:23 |
| → Pistahhh joined | 17:42 |
| → lyakh joined | 17:42 |
| → GyrosGeier joined | 17:45 |
| → strangy joined | 18:01 |
|
strangy
| mornin' | 18:01 |
|
fonseca
| Good morning. | 18:09 |
| → cartman joined | 18:09 |
| → sgrimm joined | 18:25 |
| → lyakh joined | 18:33 |
| → orospakr joined | 18:40 |
| → tcoppi joined | 18:40 |
| → ofri joined | 18:41 |
|
lu_zero
| how to make git-svn track externals ? | 18:50 |
|
| I was used to the .4.4 one and now I'm feeling quite lost | 18:51 |
| jwb → jwb_gone | 18:56 |
| → JanHudec joined | 19:12 |
| → Glanzmann joined | 19:16 |
|
Glanzmann
| Has someone tried the windows executables? | 19:16 |
| → fultilt joined | 19:23 |
|
Oejet
| Glanzmann: I tried the one from lilypond.org. | 19:24 |
|
| Glanzmann: Had the same problems as was reported on the mailing list: libiconv missing, path not set, etc. | 19:25 |
| → riddochc joined | 19:27 |
|
riddochc
| Hi, folks. | 19:27 |
|
| I believe I've found another issue in fast-import. | 19:27 |
|
Glanzmann
| Doesn't sound that bad. | 19:29 |
|
riddochc
| Heh. Well, looks like both gitte and speare are off-channel, so I'm composing a description of the issue for the list. | 19:31 |
|
| So... I've got a logfile that's definitely too big for an attachment, and I'm not in the habit of maintaining a website of my own. | 19:41 |
|
| Any suggestions on where to put it? | 19:42 |
|
motion
| pastie.caboo.se | 19:42 |
|
| rafb.net/paste | 19:42 |
|
riddochc
| Thanks. | 19:46 |
| → careym joined | 19:51 |
|
careym
| hi all - am new to git - looking to do a manual bisect (I have to update three libraries as part of the bisect process) and need to be able to determine what the tag of my current location in the repository is, is the recommend command to do this `git-log --pretty=oneline | head -n 1` | 19:54 |
|
gitster
| "git-rev-parse HEAD"? "git show -s --pretty=oneline"? | 19:55 |
|
careym
| gitster: thanks all show the same commit hash, I presume that when I rewind away from current HEAD that "git-rev-parse HEAD" will no longer work? | 19:57 |
|
| next question if git-checkout selects which branch I am working on what is the git command for moving back to a specific date/time on that branch? | 20:21 |
| ← cartman left | 20:26 |
| → dduncan joined | 20:31 |
|
gitster
| http://www.kernel.org/pub/software/scm/git/docs/user-manual.html#detached-head | 20:39 |
|
careym
| gitster: thanks, do I have to search through the history to specify a commit or can I give "git checkout" a date like --since"6 Months Ago" | 20:45 |
|
gitster
| http://www.kernel.org/pub/software/scm/git/docs/user-manual.html#recovering-lost-changes | 20:46 |
|
| The manual is good. At least you should read it once before straying into alley on your own. | 20:47 |
|
careym
| wont learn if I dont stray into the alley, best way to learn is to figure specific things out - thanks for your pointers to date | 20:49 |
|
gitster
| I did not say "read it to prevent you from straying". | 20:49 |
|
| Instead, I meant "read it before you go stray, so that you know what are possible ways out, once you go into alley." | 20:50 |
| → ofri joined | 21:18 |
| → nud joined | 21:19 |
| → cartman joined | 21:34 |
| → nud_ joined | 21:39 |
| → GyrosGeier joined | 21:57 |
|
motion
| gitster: you should arm your self too. you never know whats hiding in the alley | 21:59 |
|
GyrosGeier
| aaargh | 22:00 |
| → tglx joined | 22:00 |
|
GyrosGeier
| I just did a cg-update in a dirty tree | 22:00 |
|
| now it is confused | 22:00 |
|
| and it seems I lost the changes I made | 22:00 |
|
| well | 22:01 |
|
| third time's the charm | 22:01 |
|
tglx
| any gitweb experts around ? | 22:04 |
|
| http://git.kernel.org/?p=linux/kernel/git/tglx/rt-tests.git;a=snapshot;h=afb46d37f279ac1503e1039fefacf306a895b134 | 22:04 |
|
| results in a 29 byte file, which is not even remotely a tarball | 22:04 |
|
motion
| whats file say | 22:06 |
|
tglx
| motion: oops. its a tar.gz, but empty | 22:07 |
| → PurpleHaze joined | 22:09 |
| puzzels → puzzles | 22:10 |
| → tcoppi joined | 22:11 |
|
gitster
| gzip says it is not in gzip format... | 22:11 |
|
cartman
| works fine here | 22:11 |
|
| its empty though | 22:11 |
|
tglx
| why is it empty ? | 22:12 |
|
| cartman is no gitweb expert, sadly | 22:12 |
|
gitster
| k.org gitweb seems to be seriously busted. | 22:14 |
| → ooook joined | 22:15 |
|
gitster
| I picked a random other repository and tried snapshot (I never use it usually, so I do not know when the breakage started nor if it even ever worked) and got garbage. | 22:15 |
|
ooook
| ok I want to get the xgl-0-0-1 head from freedesktop | 22:15 |
|
| what's the right command? | 22:15 |
|
| git clone git://blahblah checkout -f headfile ?? | 22:17 |
|
| anyone around? | 22:22 |
|
motion
| tglx: well what is the commit | 22:24 |
|
tglx
| ? | 22:24 |
| → hackbert-hh joined | 22:25 |
|
motion
| tglx: thats because the commit IS empty ... | 22:25 |
|
tglx
| http://git.kernel.org/?p=linux/kernel/git/tglx/rt-tests.git;a=commitdiff;h=afb46d37f279ac1503e1039fefacf306a895b134;hp=dfbefb8056e363bca648a50274abe9983bee2da3 | 22:25 |
|
motion
| err hmm | 22:25 |
|
tglx
| it's not | 22:25 |
|
motion
| all the snapshots return 29 bytes | 22:26 |
| → z3ro joined | 22:26 |
|
motion
| maybe they're disabled ? | 22:27 |
|
| tglx looks at hpa | 22:27 |
|
tglx
| motion: then the link should be disabled as well | 22:27 |
|
motion
| tglx: yeah i thought so too | 22:28 |
|
| but maybe not | 22:28 |
| → z3ro joined | 22:31 |
|
gitster
| tglx: the right person to ask about this I suspect is warthog9; k.org runs his fork of gitweb. | 22:31 |
|
tglx
| ok, will poke him | 22:33 |
| → cods_ joined | 22:36 |
| cods_ → cods | 22:36 |
|
motion
| umm | 22:42 |
|
| does git-daemon have to run from / | 22:42 |
|
| cant i like run it with a base of /var/git | 22:42 |
|
| --base-path i guess :D nm | 22:43 |
|
hpa
| tglx: Yes? | 22:44 |
|
tglx
| hpa: are snapshots disabled on git.k.org ? | 22:46 |
|
hpa
| tglx: I belive so | 22:46 |
|
tglx
| then we should remove the snapshot link | 22:46 |
|
hpa
| Actuall, no, they aren't | 22:47 |
|
tglx
| hpa: then gitweb on k.org is hosed | 22:48 |
|
hpa
| Just broken, I guess | 22:48 |
|
| Well, snapshots are, anyway | 22:48 |
| ← careym left | 22:48 |
|
tglx
| hpa: ? | 22:48 |
|
hpa
| Snapshots are hosed | 22:49 |
|
tglx
| on kernel org or in general ? | 22:49 |
|
hpa
| On kernel.org | 22:49 |
|
tglx
| ok | 22:49 |
|
| can you poke warthog9 to fix that ? | 22:49 |
|
hpa
| Can I? Sure. | 22:50 |
|
| It's likely to go to the bottom of his very long list. | 22:50 |
|
tglx
| will you do it, please ? | 22:50 |
|
motion
| mmm why cant git:// just be encrypted in its own right :( | 22:50 |
|
hpa
| Even so, I'm not sure if snapshots should get enabled, at least not with the current server config. On the other hand, what warthog9 is spending his time on right now is bringing new servers online, which is much higher priority. | 22:51 |
|
| It's pretty sad that the upstream maintainers aren't interested integrated caching, which is obligatory on kernel.org. | 22:51 |
|
| Snapshots are nasty, because they're custom-generated and big. | 22:52 |
|
tglx
| true | 22:53 |
|
hpa
| Makes them hard to cache. | 22:53 |
|
tglx
| then simply disable them | 22:53 |
|
hackbert-hh
| hi | 22:53 |
|
hpa
| We probably should, yes. | 22:53 |
|
tglx
| and remove the link | 22:53 |
|
hackbert-hh
| how can I push tags into a remote repository? | 22:53 |
|
tglx
| its better than a broken link | 22:53 |
|
hpa
| Sure | 22:53 |
|
tglx
| ok, time for bed here. CU! | 22:53 |
|
hpa
| Ciao | 22:53 |
|
hackbert-hh
| is that possible at all? | 22:53 |
|
hpa
| git push --tags | 22:54 |
|
| Amazing, isn't it? | 22:54 |
|
hackbert-hh
| hpa: thanks | 22:54 |
|
hpa
| Who is the maintainer of man(1) these days? | 22:54 |
|
| Federico Lucifredi [email@hidden.address] looks like | 22:55 |
|
| It would be really nice if man understood stuff like "man git push" | 22:55 |
| → z3ro joined | 22:56 |
|
motion
| can i tell clone to NOT put any 'links' to the old repo in anyway | 22:56 |
|
| or | 22:56 |
|
| how do i copy a repo to a server to be served with git+ssh | 22:57 |
|
hpa
| You can either create an empty repo (git db-init) | 22:57 |
|
ooook
| man git-push | 22:57 |
|
segher_
| hpa: install the git manpages in a "mangit" mandir, and you can say just that :-) | 22:57 |
|
hpa
| segher_: Heh | 22:58 |
|
| segher_: Cute hack | 22:58 |
|
segher_
| not really a hack... this fits the "man categories" model quite nicely :-) | 22:59 |
|
| hpa figured that instead of the useless behaviour of catting multiple man pages together, it should take "man git foo" and look for either git-foo or "git foo" | 22:59 |
|
motion
| how do i tell git init to make the repo bare like git clone --bare :/ | 22:59 |
|
| or just mv blah/.git something | 22:59 |
|
| :/ | 22:59 |
|
segher_
| while we're talking manpages... feature request... can we have a makefile that installs the manpages in the "manpages" git branch please? :-) | 22:59 |
|
hpa
| Or env GIT_DIR=blah git init | 22:59 |
| → z3ro joined | 23:00 |
|
hackbert-hh
| what does git-archive do? Does it only generate a tarball from the source tree? | 23:01 |
|
motion
| hpa: and set bare = true in config im guessing, meh | 23:01 |
| → notcartman joined | 23:02 |
|
motion
| hahaha | 23:02 |
|
| git repack -a -d -f and the rpeo grows 12K | 23:02 |
|
| hackbert-hh now understands what git-archive does | 23:04 |
|
motion
| is it possible to 'ls' the available repos over git+ssh/ | 23:06 |
|
| or git:/ | 23:06 |
|
cehteh
| yes | 23:07 |
|
motion
| :/ | 23:07 |
|
cehteh
| .. but i forgot the command .. moment :) | 23:07 |
|
motion
| kk | 23:07 |
|
| git-ls-remote says it lists in a remote repo | 23:08 |
|
| not remote repos | 23:08 |
|
cehteh
| git peek-remote .. but lists branches/heads whatever, not the repos | 23:09 |
|
motion
| kk | 23:09 |
|
cehteh
| listing repos is basically a 'ls' | 23:09 |
|
motion
| is pushing over https:// as performant as git+ssh: | 23:09 |
|
cehteh
| dunno if git:// can do that | 23:09 |
|
| i wont use https | 23:09 |
|
motion
| git+shh | 23:09 |
|
| i use https:// on hg right now | 23:10 |
|
| i prefer that | 23:10 |
|
| with clean urls | 23:10 |
|
cehteh
| heh ok | 23:10 |
|
motion
| then something+ssh://host/some/ugly/path/name | 23:10 |
|
cehteh
| well i dont know | 23:10 |
|
motion
| or symlink /git -> /home/git | 23:10 |
|
| mm that owrked well | 23:11 |
|
| cehteh has /git -> /var/www/git | 23:11 |
|
motion
| cehteh: kk :D | 23:12 |
|
cehteh
| /var/www/git us chgrp git and all git users can place symlinks to their homes there | 23:12 |
|
motion
| yeah | 23:12 |
|
| this box /var isnt that big so i stick trac/hg/svn/git repos in /home | 23:13 |
|
cehteh
| needs little trust that noone touches existing symlinks | 23:13 |
|
| yeah here is all in one vserver .. but having the repos in home dirs turned out to be better anyways | 23:13 |
| notcartman → cartman | 23:16 |
| → DrNick joined | 23:17 |
|
motion
| yeah | 23:18 |
|
| cehteh: is there a simple howto for setting up gitweb | 23:18 |
|
cehteh
| apt-get install gitweb | 23:19 |
|
motion
| O | 23:19 |
|
| thnx | 23:19 |
|
cehteh
| and then just read and edit /etc/gitweb.conf or such | 23:19 |
|
| actually i did it once and forgot how it works, was very easy | 23:19 |
|
motion
| does it go through apache or run on its own ? | 23:19 |
|
cehteh
| dont froget the .css files .. that was the only notably gotcha | 23:20 |
|
| apache perl/cgi | 23:20 |
|
motion
| kk | 23:20 |
|
cehteh
| if you have problems later ask me then i can show you my config ... but its really simple | 23:20 |
|
ooook
| is there a git manual/wiki that just lists example commands and effects... | 23:23 |
|
cehteh
| alot | 23:24 |
|
ooook
| really? | 23:24 |
|
cehteh
| google :) | 23:24 |
|
ooook
| yeah, all the wikis try to verbalize everything | 23:24 |
|
| I prefer "trip-wire" -esc explanations | 23:27 |
|
motion
| hahah good ole gentoo | 23:27 |
|
ooook
| gentoo? | 23:28 |
|
| ? = , how so? | 23:29 |
|
motion
| ? | 23:29 |
|
| its webapp-config tool installed gitweb.cgi in htdocs vs cgi-bin | 23:29 |
|
ooook
| ah, I neglected to unset the variable.... | 23:30 |
|
motion
| umm where in gitweb.cgi do i tell it where the repos are | 23:31 |
|
cehteh
| motion: # head -2 /etc/gitweb.conf | 23:36 |
|
| # path to git projects (<project>.git) | 23:36 |
|
| $projectroot = "/var/www/git"; | 23:36 |
| → johnw joined | 23:37 |
|
motion
| ahh thnx | 23:39 |
|
| was commented out | 23:39 |