| 2008-06-28 |
| ← strangy left | 00:00 |
| ← drizzd left | 00:01 |
| ← igorgue left | 00:01 |
| → lea_w joined | 00:01 |
| → strangy joined | 00:01 |
| → igorgue joined | 00:03 |
| ← ph^_ left | 00:04 |
| ← strangy left | 00:11 |
| ← wea0 left | 00:11 |
| ← meyering left | 00:12 |
|
thorny_sun
| do people generally store their git repos as bare on their git servers? | 00:14 |
|
| and does that mean that everytime you push you should include a --bare? or does it know somehow? | 00:15 |
|
bronson_
| thorny_sun, yes | 00:16 |
|
| you never want to push to a non-bare repo. | 00:16 |
|
thorny_sun
| so how do you initially set up the bare repo? | 00:16 |
|
jengelh
| git bare --init in 1.5.6 | 00:16 |
|
bronson_
| git init --bare | 00:16 |
|
jengelh
| git --bare init before 1.5.6; git init --bare from 1.5.6 | 00:16 |
|
bronson_
| right | 00:17 |
|
thorny_sun
| ah-- sweet-- thanks!!! | 00:17 |
|
jengelh
| and GIT_DIR=. git-init for really olds :) | 00:17 |
|
thorny_sun
| there's no way to create a repo remotely through the git-xxx commands? | 00:17 |
|
jengelh
| ssh remote "cd /foobar; git init --bare" | 00:18 |
|
| ;-) | 00:18 |
|
thorny_sun
| well when you put it like that it doesn't seem so bad | 00:18 |
| ← cnighs left | 00:18 |
| queso → queso|away | 00:19 |
| → huyx joined | 00:21 |
| → stouset joined | 00:21 |
| igorgue → Exile | 00:22 |
| Exile → igorgue | 00:23 |
| → stragerLN` joined | 00:23 |
| queso|away → queso | 00:24 |
| igorgue → Exile | 00:24 |
| ← robinr left | 00:26 |
| → doener_ joined | 00:27 |
| ← ijcd left | 00:29 |
| ← careo left | 00:29 |
| ← _slacker_ left | 00:30 |
| → mezcalero joined | 00:31 |
|
mezcalero
| why's there both git-blame and git-annotate? what's the big difference? | 00:31 |
|
jengelh
| the name. | 00:31 |
|
mezcalero
| why are they both shipped with git if one is redundant? | 00:32 |
| → vbgunz joined | 00:32 |
|
jengelh
| because we like the sound of "blame" | 00:32 |
|
| because when code breaks, you do a git-blame and quickly know who to blame for a bug | 00:33 |
|
mezcalero
| sure, but it could be an alias for git-annotate | 00:33 |
|
jengelh
| annotate is the original cvs name | 00:33 |
|
| i'd say it is an alias | 00:33 |
|
mezcalero
| it's not | 00:33 |
|
jengelh
| it is | 00:33 |
|
| 47280694 -rwxr-xr-x 88 root root 947194 Jun 22 12:06 git-annotate | 00:33 |
|
| 47280694 -rwxr-xr-x 88 root root 947194 Jun 22 12:06 git-blame | 00:33 |
|
bronson_
| that's comedy that their manpages are different | 00:34 |
|
jengelh
| heh | 00:34 |
| ← stragerLN left | 00:34 |
|
mezcalero
| jengelh: dude, the link count is 88 | 00:34 |
|
jengelh
| so what | 00:34 |
|
| blame git | 00:34 |
|
mezcalero
| jengelh: it's still a completely different tool | 00:34 |
|
| written by different people | 00:34 |
|
| according to the man page | 00:34 |
|
jengelh
| the only difference I see between annotate and blame is the layout of the text | 00:34 |
|
mezcalero
| with different output | 00:34 |
|
jengelh
| but only marginally | 00:35 |
|
| I see. | 00:35 |
|
| git-blame is the more advanced thing it seems, though. | 00:35 |
|
mezcalero
| they mostly take the same params | 00:35 |
|
DrNick
| git-annotate is return cmd_blame(argc + 1, nargv, prefix); | 00:35 |
| → marc_vw joined | 00:35 |
|
DrNick
| and the argv is annotate -c | 00:36 |
|
| HTH | 00:36 |
|
mezcalero
| the blame man page says -c Use the same output mode as git-annotate(1) (Default: off). | 00:36 |
|
| so, it's basically an alias, however with slightly different output | 00:36 |
|
| i wonder what the point of that is | 00:37 |
| ← digitalhobbit left | 00:37 |
|
thorny_sun
| git-ers -- when i clone a repo i just made on the server, i end up with remote branches: origin/HEAD and origin/master-- is this normal-- all HOWTOs i looked only show the origin/master-- why do i end up with HEAD as well? | 00:37 |
|
cardioid
| There's no git praise :( | 00:37 |
|
thorny_sun
| jengelh: ? | 00:38 |
|
jengelh
| thorny_sun: ? | 00:38 |
| ← doener left | 00:39 |
| ← radarek left | 00:39 |
|
thorny_sun
| jengelh: thought you might know... my previous question two above | 00:39 |
|
jengelh
| my scrolling is like an old man walking | 00:39 |
|
thorny_sun
| is it rude to call out people for questions? | 00:39 |
|
cardioid
| jengelh: slow and annoying? | 00:39 |
| queso → queso|away | 00:39 |
|
jengelh
| just slow | 00:39 |
|
| Sat Jun 28 02:37:01 CEST 2008 | 00:40 |
|
thorny_sun
| jengelh: git-ers -- when i clone a repo i just made on the server, i end up with remote branches: origin/HEAD and origin/master-- is this normal-- all HOWTOs i looked only show the origin/master-- why do i end up with HEAD as well? | 00:40 |
| → freshtonic joined | 00:40 |
| ← mithro left | 00:43 |
| ← felipec left | 00:44 |
| → stragerLN joined | 00:44 |
| → flaguy joined | 00:45 |
| ← spearce left | 00:45 |
|
thorny_sun
| and what does that mean exactly? origin/HEAD? why doesn't HEAD show up in my local repo when i just do 'git remote'? is there something significant about origin/HEAD? | 00:47 |
|
comp
| from what I know - HEAD shouldn't be in bare repos | 00:49 |
|
thorny_sun
| if i create a bare repo on the server can i push my non-bare local repo up to it and expect it to stay bare? | 00:51 |
|
| comp? | 00:51 |
|
comp
| well .. yes | 00:52 |
| → stragerL1` joined | 00:52 |
|
thorny_sun
| or do i need to clone --bare first and then push? | 00:52 |
|
comp
| that's how it is supposed to work | 00:52 |
|
| when you push | 00:52 |
|
| you don't push your working dir | 00:52 |
| ← mvw left | 00:52 |
|
thorny_sun
| i push from my working dir, but it's pushing the latest master checkin-- correct? | 00:53 |
|
| if i do 'git push origin' | 00:53 |
|
comp
| "The special symbol "HEAD" can always be used to refer to the current branch. In fact, git uses a file named "HEAD" in the .git directory to remember which branch is current" | 00:54 |
|
| and if I got it right, there's no "current" branch in bare repo | 00:54 |
|
thorny_sun
| ah i see | 00:55 |
| ← stragerLN` left | 00:55 |
|
thorny_sun
| maybe i need to build to latest version | 00:55 |
| ← SRabbelier left | 00:55 |
| → capisce joined | 01:01 |
| ← warthog9 left | 01:03 |
|
Mantaar
| is there any way to restrict bandwidth usage of git-push? | 01:04 |
|
| (or git-pull for that matter) | 01:04 |
| → up_the_irons2 joined | 01:04 |
|
thorny_sun
| i just did a ./configure, make, make install of latest git, but my version is still the old one-- what gives? any ideas?-- i'm on ubuntu hardy heron. | 01:05 |
| ← stragerLN left | 01:05 |
| ← up_the_irons2 left | 01:05 |
|
mczepiel
| thorny_sun: did you install in the same place as your current binary? | 01:07 |
|
thorny_sun
| mczepiel: well i think i found the old one at /usr/bin and the new one got installed at /usr/local/bin-- how do i fix things so when i type git at command line it uses the one at /usr/local/bin ? | 01:07 |
| ← elight left | 01:07 |
|
thorny_sun
| mczepiel: how do you specify where to install? | 01:08 |
| ← ketralnis left | 01:08 |
|
mczepiel
| thonry_sun ./configure --prefix=/usr would probably do it for you, by default it is /usr/local | 01:08 |
|
| ./configure --help has more information | 01:09 |
|
thorny_sun
| ah ok | 01:09 |
|
| i'll try that | 01:09 |
|
| mczepiel: do you know how do i undo what i just did? | 01:09 |
| ← joydivider left | 01:10 |
|
mczepiel
| thorny_sun: other than just rm'ing a bunch of files no sorry, I know just enough to be dangerous ;-) | 01:10 |
| → jtoy joined | 01:11 |
|
jtoy
| I have fixed the code in the merge, then i do git pull and get "Merge with strategy recursive failed." , how do i fix this? | 01:11 |
| → lorandi joined | 01:11 |
|
lea_w
| Can someone point me to a really large public git repository? Larger than the linux kernel? | 01:13 |
|
thorny_sun
| does such a thing exist? | 01:14 |
|
cnk
| lea_w: I would expect the kernel repo to be pretty big | 01:14 |
| ← SuttoL left | 01:15 |
|
lea_w
| It is (800K object if I'm not miscounting), but I'm looking for something bigger for testing purposes. | 01:15 |
|
| OO.o was apparently thinking about git (and they were talking about a GB-size repo somewhere), but their test repository seems down. | 01:15 |
| → SuttoL joined | 01:15 |
| ← tcoppi left | 01:16 |
|
lea_w
| Ah, haha. | 01:16 |
|
| http://userweb.kernel.org/~hpa/kdegit/ | 01:16 |
|
| Biiiiig. ;-) Apparently they keep every KDE project in a single repo. ^_^ | 01:17 |
| → tcoppi joined | 01:17 |
|
plopix
| lea_w: or try http://gitweb.freedesktop.org/ | 01:17 |
|
| lea_w frees some space on her 18GB hard disk... | 01:17 |
| ← bronson_ left | 01:18 |
| ← hkBst left | 01:21 |
| → stragerLN joined | 01:21 |
| ← tvachon left | 01:23 |
| blowery`afk → blowery | 01:25 |
| ← Eridius|work left | 01:25 |
| → redondos joined | 01:26 |
| → VVSiz_ joined | 01:27 |
| queso|away → queso | 01:27 |
| ← thorny_sun left | 01:27 |
| → tcoppi_ joined | 01:31 |
| → tcoppi_fails joined | 01:31 |
| ← tcoppi_fails left | 01:31 |
| ← stragerL1` left | 01:32 |
| ← tcoppi left | 01:32 |
| tcoppi_ → tcoppi | 01:32 |
| ← ec left | 01:32 |
| → ajonat joined | 01:33 |
| → jhelwig joined | 01:34 |
| ← WALoeIII left | 01:35 |
| → elight joined | 01:36 |
| ← elight left | 01:37 |
| ← ghaskins left | 01:37 |
| → UVSoft joined | 01:41 |
|
UVSoft
| hi there | 01:41 |
|
| i have a lot of deleted files in many directories, how to add these deletitions at once? | 01:41 |
|
| *add to index | 01:41 |
|
| *remove from index) | 01:42 |
| ← VVSiz left | 01:44 |
| → Ramune joined | 01:45 |
|
ReinH
| UVSoft: git ls-files --removed |xargs git rm # is an option, unless someone has a better one | 01:46 |
|
UVSoft
| ReinH, thanks | 01:47 |
|
ReinH
| I think it's --removed | 01:47 |
|
| not sure if git ls-files has a print0 option, if so you should use it too | 01:47 |
|
| with xargs -0 | 01:48 |
|
Ramune
| reinh: -z | 01:49 |
|
UVSoft
| but not --removed, --deleted | 01:49 |
| ← UVSoft left | 01:50 |
| ← huyx left | 01:50 |
| ← bgerber left | 01:50 |
| → huyx joined | 01:51 |
|
ReinH
| Ramune: does print0? | 01:51 |
|
Ramune
| yup -- \0 terminates output | 01:52 |
| ← lea_w left | 01:57 |
| → stragerLN` joined | 01:58 |
| → javashin joined | 01:59 |
|
javashin
| hello | 01:59 |
|
ReinH
| Ramune: sorry, does that mean it separates items by \0 ? | 02:00 |
|
| ReinH can just try it | 02:00 |
|
ReinH
| never mind. | 02:00 |
|
| ty | 02:00 |
| → Leefmc joined | 02:00 |
|
javashin
| im trying to clone mesa and i get this | 02:00 |
|
| remote: aborting due to possible repository corruption on the remote side. | 02:00 |
|
| error: waitpid (async) failed | 02:00 |
|
| fatal: error in sideband demultiplexer | 02:00 |
|
ReinH
| any idea why -z ? | 02:00 |
|
| javashin: that is hot. | 02:01 |
|
| o_O | 02:01 |
|
javashin
| and git is downloading really slow | 02:01 |
|
| im using gentoo | 02:01 |
|
ReinH
| javashin: what version of git? | 02:01 |
|
javashin
| git version 1.5.6.1 | 02:02 |
| ← sergiodj left | 02:02 |
|
ReinH
| hmm | 02:02 |
|
| can't say I've seen that, sorry | 02:02 |
|
javashin
| dev-util/git-1.5.6.1 USE="bash-completion gtk iconv perl -cgi -curl -cvs -doc -emacs -mozsha1 (-ppcsha1) -subversion -threads -tk -vim-syntax -webdav -xinetd" | 02:03 |
| → elight joined | 02:03 |
|
javashin
| ReinH, please prove git clone git://anongit.freedesktop.org/git/mesa/mesa | 02:04 |
| ← david_koontz left | 02:04 |
| → stragerL1` joined | 02:06 |
| ← javashin left | 02:07 |
|
csc`
| Can I push all of the branches/tags all at once or do I have to push them individually? | 02:09 |
| ← \ask_ left | 02:10 |
| ← stragerLN left | 02:11 |
| ← message145 left | 02:11 |
| → message145 joined | 02:11 |
| ← plediii left | 02:14 |
| ← lorandi left | 02:14 |
| ← stragerLN` left | 02:17 |
| → mithro joined | 02:17 |
| ← pygi left | 02:18 |
| → threeve joined | 02:21 |
| ← Leefmc left | 02:22 |
| ← stragerL1` left | 02:22 |
| → Leefmc joined | 02:22 |
| ← blowery left | 02:23 |
| ← Leefmc left | 02:39 |
| → Leefmc joined | 02:39 |
| ← Leefmc left | 02:43 |
| → webmat joined | 02:43 |
| → eliot_ joined | 02:43 |
| ← alegscogs left | 02:46 |
| → zachinglis joined | 02:46 |
| → alegscogs joined | 02:48 |
| ← redondos left | 02:50 |
| ← eliot_ left | 02:54 |
| ← pombreda left | 03:00 |
| → vbabiy_ joined | 03:02 |
| ← vbabiy_ left | 03:04 |
| → \ask_ joined | 03:04 |
| ← freshtonic left | 03:12 |
| ← jhbrown left | 03:17 |
| → aurous joined | 03:17 |
| ← aurous left | 03:19 |
| ← zachinglis left | 03:19 |
| → redondos joined | 03:21 |
| → bizb joined | 03:22 |
| ← markkalderon left | 03:22 |
|
bizb
| OK I just migrated to git and modified a file that I want back out of and restore the committed file. if i remove the file it marks it for deletion. How do I restore from the local repo? | 03:23 |
| ← schacon left | 03:23 |
| → vbabiy joined | 03:24 |
| → aurous joined | 03:26 |
| ← aurous left | 03:27 |
| → aurous joined | 03:28 |
| ← mithro left | 03:29 |
| → surial joined | 03:29 |
| → geckosenator joined | 03:30 |
|
surial
| In his google talk, torvalds suggested that you should definitely use multiple git repositories if your project consists of multiple mostly separate units, but, can you still tag a set of separate repositories, or do atomic commits and the like? | 03:30 |
|
geckosenator
| I have two git repositories, the are currently separate, how do I merge them together and keep both histories? | 03:30 |
| ← janm left | 03:30 |
|
mczepiel
| bizb: git checkout modified_file_name will return it back to the pristine state iirc | 03:32 |
| ← aurous left | 03:32 |
| → Beket joined | 03:35 |
| → Ori_B joined | 03:36 |
|
Ori_B
| hi. | 03:36 |
|
| $ git-pull --rebase | 03:37 |
|
| First, rewinding head to replay your work on top of it... | 03:37 |
|
| could not detach HEAD | 03:37 |
|
| ^-- what does that mean? | 03:37 |
|
| and how do I fix it? | 03:37 |
|
| $ git --version | 03:37 |
|
| git version 1.5.6 | 03:37 |
| → aurous joined | 03:38 |
| ← aurous left | 03:39 |
|
mczepiel
| Ori_B: bit of a noob myself but why the --rebase on your pull? | 03:39 |
|
bizb
| mczepiel: thanks | 03:39 |
|
Ori_B
| mczepiel: because I don't want to merge, I just want to grab the latest changes from the repository | 03:39 |
|
mczepiel
| Ori_B: wouldn't 't git fetch do that? | 03:40 |
|
Ori_B
| no, git-fetch only fetches | 03:40 |
|
| git-fetch; git-rebase | 03:40 |
|
| should be the same | 03:40 |
|
mczepiel
| i thought the git pull performed a git-fetch and then git-merge | 03:41 |
|
Ori_B
| yes, the --rebase option changes it to rebase. | 03:42 |
| → zachinglis joined | 03:42 |
|
Ori_B
| --rebase | 03:42 |
|
| Instead of a merge, perform a rebase after fetching. | 03:42 |
|
mczepiel
| yeah i see the man entry | 03:42 |
|
| guess I have the wrong impression of what git fetch actually does | 03:43 |
| ← csc` left | 03:45 |
| ← alegscogs left | 03:48 |
| → pombreda joined | 03:50 |
| → fhobia joined | 03:52 |
| Exile → igorgue | 03:55 |
| ← mczepiel left | 03:55 |
| ← dasil003 left | 03:56 |
| kewark → krawek | 03:58 |
| ← elight left | 04:01 |
| → digitalhobbit joined | 04:02 |
| stouset → stouset|away | 04:02 |
| stouset|away → stouset | 04:04 |
| → alenoosh joined | 04:06 |
| ← shehjart left | 04:08 |
| → dddino joined | 04:10 |
|
dddino
| Hi guys, I'm new to git and was wondering if anyone could explain a few things about gitweb. Mainly, was wondering about its limitations and behavior. | 04:12 |
|
| I believe that I have it installed correctly, etc.. it does run fine on apache, but now I am kind of lost as to its web capabilities. Are users able to 'upload' files via web? | 04:15 |
| ← bizb left | 04:20 |
| ← kukks left | 04:20 |
| ← \ask_ left | 04:23 |
| ← digitalhobbit left | 04:26 |
| ← webmat left | 04:29 |
| ← tjafk1 left | 04:30 |
| → tjafk1 joined | 04:30 |
| queso → queso|away | 04:33 |
| → pergesu joined | 04:35 |
| queso|away → queso | 04:38 |
| → sbahra joined | 04:47 |
|
sbahra
| git roxmasox | 04:47 |
| → thorny_sun joined | 04:52 |
| → stragerLN joined | 04:53 |
| ← queso left | 04:54 |
| ← vbabiy left | 04:54 |
| → foca joined | 04:55 |
|
thorny_sun
| when i do a git clone from a public repo, and then do a 'git branch -r', i get origin/master and origin/HEAD. is this right? origin/HEAD does not make sense to me-- anyone able to explain? | 04:59 |
| ← bdiego left | 05:00 |
| → imyousuf joined | 05:02 |
| ← surial left | 05:07 |
| → surial joined | 05:08 |
| ← surial left | 05:09 |
| → elight joined | 05:10 |
| ← tcoppi left | 05:10 |
| → jordi_bunster joined | 05:13 |
| → mithro joined | 05:22 |
| ← huyx left | 05:23 |
| → aurous joined | 05:27 |
| ← zachinglis left | 05:27 |
| ← igorgue left | 05:28 |
| → digitalhobbit joined | 05:28 |
| ← alenoosh left | 05:33 |
| → alenoosh joined | 05:36 |
| ← elight left | 05:37 |
| ← aurous left | 05:41 |
| ← threeve left | 05:41 |
| → tcoppi joined | 05:45 |
| ← thorny_sun left | 05:46 |
| ← Beket left | 05:48 |
| → swtaarrs joined | 05:54 |
| ← foca left | 05:56 |
| saintdev_ → saintdev | 05:59 |
| ← stouset left | 06:02 |
| ← jordi_bunster left | 06:07 |
| ← dddino left | 06:07 |
| → nud joined | 06:13 |
| ← madewokherd left | 06:21 |
| ← bobesponja left | 06:23 |
| → naeu joined | 06:24 |
| → alegscogs joined | 06:29 |
| → shehjart joined | 06:31 |
| ← digitalhobbit left | 06:34 |
| → ezyang joined | 06:36 |
| ← krawek left | 06:57 |
| → pygi joined | 06:59 |
| → cncfanatics joined | 07:01 |
|
ezyang
| How do I rebase a set of merges? I tried the regular method with -p, but I got an error saying there was no -m flag. Then, when I tried to cherry-pick the commits, I lost my parent information | 07:06 |
|
Mikachu
| you can't do that easily | 07:07 |
|
| -p only "works" with -i, but it doesn't work that well | 07:07 |
| ← swtaarrs left | 07:07 |
|
ezyang
| hmm | 07:07 |
|
Mikachu
| (in my experience) | 07:07 |
|
ezyang
| So, I guess my options are: 1. leave my splotchy merge history as is, 2. reset and then redo each merge (don't want to do that), 3. cherry-pick, and then somehow amend the commit to have parent info | 07:08 |
|
| How would I do 3? | 07:09 |
|
| I know there's a magic MERGE_HEAD file or something in .git that specifies parents | 07:09 |
|
| but since cherry-pick automatically commits I don't know how to use that | 07:09 |
|
Mikachu
| well, there's --no-commit | 07:10 |
|
ezyang
| yeah, I'm gonna try that | 07:11 |
|
| what's the format of MERGE_HEAD? | 07:11 |
|
| (never actually seen it) | 07:11 |
|
Mikachu
| no idea | 07:11 |
|
| you can cheat, do the cherry-pick, record the sha1, do the merge, read-tree the first sha1 and amend :) | 07:11 |
|
| (untested) | 07:11 |
| ← nud left | 07:12 |
|
ezyang
| oh, easier way; use merge -s ours | 07:13 |
|
| hehehe | 07:13 |
|
| let me hope that this sorry hack works | 07:15 |
| ← broonie left | 07:16 |
| → broonie joined | 07:16 |
|
Mikachu
| if you use -s ours, you'll not get any data from the other branch | 07:16 |
|
| if you're doing my amend trick that's useful though | 07:16 |
| ← quabbles left | 07:19 |
| ← unreal left | 07:19 |
| ← anders_ left | 07:19 |
|
ezyang
| Mikachu: I think that's what I want (not 100% sure). I'm going to get the data from the cherry pick | 07:20 |
| → anders_ joined | 07:21 |
|
ezyang
| ruh oh, looks like it's merge first, then cherry-pick | 07:22 |
|
| first merge looks like it worked :-) | 07:24 |
| ← alegscogs left | 07:24 |
|
ezyang
| maybe this'll be the way to get rebase working with merges | 07:24 |
|
Mikachu
| oh, you probably want to do a reset --hard after the read-tree+amend too, or your working tree won't be matching | 07:26 |
|
| (not before the amend or it will undo the read-tree) | 07:26 |
|
ezyang
| Mikachu: As far as I can tell, it's working without the read-tree | 07:27 |
|
Mikachu
| what are you doing exactly? | 07:27 |
|
ezyang
| I'm running git merge -s ours --no-commit Q && git cherry-pick -n -m 1 S && git commit | 07:27 |
|
Mikachu
| ah, that's much smarter :) | 07:27 |
|
| then you can disregard me :) | 07:28 |
|
ezyang
| From gitk, it looks like it's working | 07:28 |
|
| nah, you put me in the right direction :-) | 07:28 |
|
| thanks a bunch | 07:28 |
| ← broonie left | 07:28 |
|
Mikachu
| i mean the stuff i said about reset | 07:28 |
| → broonie joined | 07:28 |
| → unreal joined | 07:33 |
|
ezyang
| and done. I'm not sure if using git merge with backports is abusive, but wth | 07:35 |
| ← rleigh left | 07:36 |
| → Ademan joined | 07:37 |
|
ezyang
| I suppose the pretty gitk graphs make up for the ugly shortlog | 07:40 |
| ← fade-in left | 07:40 |
| → daaku joined | 07:42 |
| → nuge joined | 07:44 |
| ← nuge_ left | 07:47 |
| → naelyn joined | 07:48 |
| ← mithro left | 07:48 |
| ← ezyang left | 07:49 |
| → robinr joined | 07:57 |
| → crab joined | 07:58 |
| → ph^ joined | 08:00 |
| → warthog9 joined | 08:11 |
| → nud joined | 08:20 |
|
robinr
| could merge be "smart" and reverse the roles of the branches merged when reversing would result in a fast-forward? | 08:21 |
|
| i.e. when git diff HEAD^2 would be empty | 08:22 |
|
gitster
| Like merging 'master' into 'maint' after every major release? | 08:23 |
| → d0k joined | 08:29 |
|
crab
| hi. | 08:35 |
| ← SuttoL left | 08:38 |
| ← imyousuf left | 08:44 |
|
| crab trawls through the todo. | 08:46 |
| marc_vw → mvw | 08:48 |
| → elliottcable joined | 08:49 |
| → radarek joined | 08:49 |
| ← elliottcable left | 08:50 |
| → elliottcable joined | 08:51 |
| ← sverrej left | 08:52 |
| → sverrej joined | 08:53 |
|
robinr
| gitster: sounds like that | 08:53 |
|
| would be nice if merge could do it automatically | 08:54 |
| → deavid joined | 08:59 |
| → meyering joined | 09:00 |
| → priidu joined | 09:01 |
| ← fhobia left | 09:02 |
| → _zack joined | 09:03 |
| ← jtoy left | 09:14 |
| ← deavid left | 09:15 |
| → _zack1 joined | 09:17 |
| ← _zack left | 09:17 |
| → shodges joined | 09:17 |
| → BabelO joined | 09:18 |
| ← BabelO left | 09:18 |
|
shodges
| hey all | 09:19 |
|
sbahra
| Hi man! | 09:19 |
|
shodges
| Got a question, if I do a git-rm on a file and change my mind before I commit - how do I unstage that removal? | 09:19 |
|
| if I do "git reset -- myfile" it returns "myfile: needs update" - i'm not sure how to perform that update | 09:19 |
|
| s/unstage/undo | 09:20 |
| ← Mantaar left | 09:21 |
|
Mikachu
| shodges: that means you did unstage the removal | 09:23 |
|
| shodges: you didn't undo the removal of the actual file though, so git warns you that you have a change in your worktree that isn't staged | 09:23 |
|
| shodges: if you want the file back you can say git checkout -- thefile | 09:23 |
| ← redondos left | 09:23 |
|
Mikachu
| (that will also stage it btw, even if you didn't do the reset) | 09:24 |
| → BabelO joined | 09:24 |
| → ph^_ joined | 09:24 |
|
sbahra
| shodges, git rocks! You can also do git rm --cached | 09:24 |
|
Mikachu
| he wants to _un_do the removal :) | 09:25 |
|
sbahra
| Oh yeah! | 09:25 |
|
| Well, you answered that. :-P | 09:25 |
|
shodges
| cool, thanks Mikachu that makes sense | 09:26 |
|
Mikachu
| if you want to restore the file but not undo the removal in the index, you have to do some magic incantation like git cat-file blob HEAD:thefile > thefile | 09:26 |
|
| of course it's much easier to do the checkout + rm --cached | 09:27 |
|
shodges
| yeah I added an alias for "git rm --cached" today, which leads me nicely into my last question ;) | 09:27 |
|
| is it impossible to over-ride a built-in command with an alias? I wanted to do "alias.rm = rm --cached" | 09:28 |
|
Mikachu
| yes | 09:28 |
|
| since that would break any script | 09:29 |
| → redondos joined | 09:30 |
|
shodges
| I thought so, ok well long as I've got a way to revert my mistake of doing "git rm" I'll be fine | 09:30 |
|
| thanks a lot Mikachu, and sbahra for your suggestion | 09:30 |
|
Mikachu
| btw, if you make some changes, git-add, and then accidentally git-rm, you can't get them back without lots of trouble | 09:30 |
|
| ie, the object will be in .git/objects, but it won't be referenced anywhere | 09:31 |
|
| but if you do changes, git-add and accidentally /bin/rm, you can get it back with git checkout -- thefile as before | 09:31 |
|
shodges
| ok, worth bearing in mind. I guess "git checkout" should be the first thing to consider when resolving file removal issues | 09:33 |
| → janm joined | 09:34 |
| ← janm left | 09:34 |
| ← _zack1 left | 09:34 |
| → _zack joined | 09:34 |
| → janm joined | 09:35 |
| → yofel joined | 09:36 |
| → drizzd joined | 09:37 |
| ← ph^ left | 09:40 |
| → garrin joined | 09:41 |
| → SRabbelier joined | 09:42 |
| ← _zack left | 09:48 |
| → _zack1 joined | 09:48 |
| ← _zack1 left | 09:48 |
| → _zack joined | 09:49 |
| ← redondos left | 09:51 |
| → Eludias joined | 09:53 |
| → SuttoL joined | 09:58 |
|
biz
| is there a way to add submodules and force git to get always the HEAD of them later on? Eg. git submodule update should not get the commit / head as the submodule was added but always the current head? | 09:59 |
|
Tv
| that would make tree contents unreliable, which is against git philosophy | 10:00 |
|
| maybe you just want a makefile | 10:00 |
|
biz
| Tv: the philosophy is correct, but in this case I'd just add submodules to ease the process / just include a certain project within another without linking against it nor the need for a specific state of the "submodule" | 10:02 |
|
bremner
| apropos of submodules, is it obvious how to merge submodules into one repo? If so, please tell :-) | 10:03 |
|
biz
| so a Makefile with eg. git pull for the specific submodules is the better way? | 10:03 |
| → babbage joined | 10:06 |
|
babbage
| I have a branch in my git repository called rel-4-4-fixes that I normally update with "git fetch origin rel-4-4-fixes:rel-4-4-fixes". Can I set that up as the default effect of "git pull" on that branch? | 10:07 |
|
Tv
| babbage: you fetch straight to a local branch? that's not really meant to be done like that | 10:08 |
|
biz
| bremner: since a submodule is usually localed in its own subdir, you could probably just remove the submodule entry and clone it into this subdir, then remove .git and add it within the containing repo? | 10:08 |
|
Tv
| babbage: normal git fetch would fetch remote's branch foo to remotes/origin/foo | 10:09 |
| ← nud left | 10:09 |
|
bremner
| biz: err yes, but I hope to preserve some history | 10:10 |
|
Tv
| bremner: see how gitk and gitweb got merged to git | 10:10 |
|
| bremner: it's been discussed on the mailing list many times | 10:11 |
|
bremner
| Tv: ahh, RTFML :-) | 10:11 |
|
biz
| Heh :-) | 10:12 |
|
babbage
| Sorry I mean "git pull origin rel-4-4-fixes:rel-4-4-fixes". | 10:12 |
|
Tv
| bremner: Documentation/howto/using-merge-subtree.txt | 10:12 |
|
bremner
| Tv: thank you! | 10:13 |
|
Tv
| babbage: umm, please don't use pull like that | 10:14 |
|
babbage
| Tv, so what should I be doing? | 10:14 |
|
Tv
| babbage: i'd use git fetch && git merge origin/rel-4-4-fixes | 10:15 |
| → \ask_ joined | 10:15 |
|
babbage
| So AIUI all you are really saying is that I have no need of a local rel-4-4-fixes, yes? | 10:15 |
| ← daaku left | 10:15 |
|
Tv
| you could use config branch.rel-4-4-fixes.merge to automate that with git pull | 10:16 |
|
| frankly, i don't want git pull myself | 10:16 |
| → chris2 joined | 10:16 |
|
Tv
| babbage: err, you need local branches if you want to work on them | 10:16 |
|
| i don't know what you want | 10:16 |
|
| but you most likely should not update local branches in the fetch operation | 10:17 |
|
| that probably does not work how you think it would | 10:17 |
|
babbage
| I don't ever commit into the local rel-4-4-fixes branch, though | 10:19 |
|
Tv
| do you check it out? | 10:19 |
|
babbage
| Rarely, and then mainly to conveniently analyse problems or figure out what changed. | 10:19 |
| → mbuf joined | 10:38 |
|
BeerSerc
| I have committed two little debugging changes, which I now have reverted. Is there a way to remove those 4 lines from the log? | 10:45 |
|
babbage
| you could rebase the branch interactively and just drop those changes. | 10:46 |
|
| but that is not ideal if other people are pulling from it/ | 10:46 |
|
BeerSerc
| babbage: I tried, but when I do so, I cannot push afterwards | 10:46 |
|
babbage
| because it's no longer fast-forward | 10:46 |
|
BeerSerc
| this is not much more than a test repo, where no real action happens | 10:46 |
|
| only one other person is working on that, and I know he hasnt commited since | 10:47 |
| ← \ask_ left | 10:47 |
|
babbage
| You could use "push -f" but be careful; the restriction against not losing remote commits is there for your protection | 10:47 |
|
| Frankly in this scenario I always leave the commits in there;. | 10:48 |
|
BeerSerc
| babbage: thanks. As I said, no important code there anyway, so it should not be a problem | 10:48 |
|
babbage
| I don't see the point in trying to erase the traces. | 10:48 |
|
| Why do you want to remove the commits? | 10:48 |
|
BeerSerc
| we want to use the repository later, and have just set it up. And I have tried some things, because I am kind of new to git. and those commits have nothing to do with the project itself, so... | 10:49 |
|
babbage
| You could always delete the entire branch | 10:50 |
|
BeerSerc
| parts were in the master brahcn | 10:51 |
|
| s/brahcn/branch/ | 10:51 |
| ← tenpaiyomi left | 10:52 |
| → Xyri0 joined | 10:57 |
| ← mbuf left | 10:58 |
| → \ask_ joined | 10:59 |
| → scrooloose_ joined | 11:03 |
| ← Xyrio left | 11:06 |
| Xyri0 → Xyrio | 11:06 |
| ← babbage left | 11:10 |
| → lea_w joined | 11:11 |
| ← pergesu left | 11:14 |
| ← madduck left | 11:16 |
| ← scrooloose left | 11:21 |
| → capisce_ joined | 11:29 |
| ← \ask_ left | 11:34 |
| → Yuuhi joined | 11:36 |
| ← Gitzilla left | 11:40 |
| ← chris2 left | 11:44 |
| ← Hobbes left | 11:44 |
| ← beeble left | 11:44 |
| ← KwisatzHaderach left | 11:44 |
| → chris2 joined | 11:44 |
| → Hobbes joined | 11:44 |
| → KwisatzHaderach joined | 11:44 |
| → beeble joined | 11:44 |
| ← capisce left | 11:45 |
| → Arme[N] joined | 11:48 |
| ← ciaran_ left | 12:06 |
| → ciaran joined | 12:06 |
| ← drizzd left | 12:06 |
| → JanK joined | 12:09 |
|
JanK
| how would i do a change to two branches? | 12:10 |
| ← ciaran left | 12:10 |
| → ciaran joined | 12:10 |
|
JanK
| say, i have a release and master branch, and i want a fix to be in both | 12:10 |
| → ciaran_ joined | 12:12 |
| ← ciaran left | 12:12 |
| → nud joined | 12:16 |
| ← priidu left | 12:17 |
| → quabbles joined | 12:17 |
| ph^_ → ph^ | 12:20 |
|
sbahra
| JanK, you have to merge them into both yourself | 12:23 |
|
shodges
| JanK, you'll could commit it in one, switch to the other branch, and cherry-pick the commit | 12:24 |
|
JanK
| ok, thank you both | 12:24 |
| ← Ademan left | 12:26 |
| ← kanru left | 12:34 |
| ← Ramblurr left | 12:34 |
| → rmblr joined | 12:35 |
| → Leefmc joined | 12:35 |
| ← alenoosh left | 12:36 |
| → vbabiy joined | 12:36 |
| → ciaran joined | 12:43 |
| ← ciaran_ left | 12:43 |
| → kanru joined | 12:43 |
|
Tv
| JanK: cherry-pick makes those two independent commits | 12:44 |
|
JanK
| yeah, that works for me | 12:44 |
|
Tv
| JanK: if you merge your stable to your devel regularly, you might just commit the fix to stable and merge it to devel | 12:44 |
|
| having duplicate commits is annoying | 12:44 |
| ← ciaran left | 12:45 |
| → ciaran joined | 12:45 |
|
JanK
| is there a good gui to browse a repository? | 12:45 |
| ← ciaran left | 12:45 |
|
Tv
| gitk | 12:45 |
|
| git gui is more than browsing | 12:45 |
|
Mikachu
| it's not browsing at all, i would say :) | 12:46 |
| → ciaran joined | 12:46 |
|
Tv
| well git gui sort of includes gitk, iirc | 12:46 |
|
Mikachu
| yeah you can launch it from the menus | 12:46 |
| → ciaran_ joined | 12:48 |
| ← ciaran left | 12:48 |
| ← ciaran_ left | 12:48 |
| → ciaran joined | 12:49 |
| ← naeu left | 12:54 |
| ← shodges left | 12:56 |
| Arme[N] → Arme[0] | 13:00 |
| → devcow joined | 13:01 |
| → ciaran_ joined | 13:08 |
| ← ciaran left | 13:08 |
| → naeu joined | 13:10 |
| ← ciaran_ left | 13:10 |
| → ciaran joined | 13:11 |
| ← ciaran left | 13:11 |
| → ciaran joined | 13:11 |
| → cpfr joined | 13:13 |
| ← naeu left | 13:16 |
| ← ciaran left | 13:16 |
| → ciaran joined | 13:16 |
| ← doener_ left | 13:17 |
| → doener joined | 13:17 |
| ← ciaran left | 13:19 |
| → ciaran joined | 13:19 |
| → lea_w1 joined | 13:20 |
| ← lea_w left | 13:20 |
| lea_w1 → lea_w | 13:20 |
| → jmalonzo joined | 13:23 |
| ← ciaran left | 13:23 |
| → ciaran joined | 13:23 |
| ← ciaran left | 13:25 |
| → ciaran joined | 13:25 |
| ← ciaran left | 13:26 |
| → mawx joined | 13:26 |
| → ciaran joined | 13:26 |
| ← ciaran left | 13:26 |
| → byronvf joined | 13:27 |
| → ciaran joined | 13:27 |
| ← ciaran left | 13:27 |
| → ciaran joined | 13:27 |
| ← JanK left | 13:28 |
|
mawx
| Hi, I tried to push my local repo to a remote server using "git push server branch; git push --tags". But when i first checked it out i got a "Warning: Remote HEAD refers to nonexistent ref, unable to checkout." | 13:28 |
|
| So i did git push server master | 13:28 |
|
| and now if i checkout i can't find my branch. | 13:28 |
|
| Any idea what i am doing wrong? | 13:29 |
|
Tv
| you mean clone, not checkout? | 13:29 |
|
mawx
| yes. | 13:29 |
|
Tv
| it clones successfully and you see your files? | 13:29 |
|
mawx
| I see the files of the master branch | 13:29 |
|
| but git branch does not list any others. | 13:29 |
|
Tv
| yes, master is the default | 13:29 |
|
| try git branch -a | 13:30 |
|
mawx
| okay. so there i got master and three starting with origin/... including the one i was looking for. | 13:30 |
| → snailrails joined | 13:31 |
|
Tv
| those would be called remote branches | 13:31 |
|
snailrails
| What is the opposite of git add . ? | 13:31 |
|
Tv
| git checkout -b foo origin/foo | 13:31 |
|
| snailrails: read git status output | 13:31 |
|
snailrails
| to commit all deleted files | 13:31 |
|
Tv
| oh | 13:31 |
|
| opposite that way ;) | 13:32 |
|
Mikachu
| git add -u . ? | 13:32 |
|
snailrails
| cool thanks | 13:32 |
|
mawx
| Tv: cool, thank you. | 13:32 |
|
snailrails
| so you need to run git add . and git add -u . every time ? | 13:32 |
|
Mikachu
| every time you do what? | 13:32 |
|
mawx
| Can i make this the default for the remote repo? | 13:32 |
|
snailrails
| commit | 13:32 |
|
Mikachu
| no, only if you delete files | 13:33 |
|
mawx
| probably by git push branch:master ? | 13:33 |
|
Tv
| mawx: well that makes master contain what your local branch has | 13:33 |
|
| mawx: you can change the default to not be "master" but that's pretty rare | 13:33 |
| ← ciaran left | 13:34 |
|
Tv
| mawx: like not calling your svn trunk "trunk" | 13:34 |
|
mawx
| okay. i think i will stick to what i have right now. | 13:34 |
| → ciaran joined | 13:34 |
| ← janm left | 13:34 |
|
mawx
| Or what is the usual way of doing it if i clone someone elses repo, add a branch with my changes and put it online somewhere else? Should i make my branch the master because that is what people will be looking for - or rather leave it in the structure it had before? | 13:35 |
| → stouset joined | 13:35 |
|
Tv
| mawx: what's "before"? | 13:36 |
|
cpfr
| hey is there anyway to permanently remove a file from git | 13:36 |
|
| like say for example i accidentally added and committed a passwd file | 13:36 |
|
mawx
| like the repo i cloned initially plus a branch that is not the default. | 13:36 |
|
Tv
| mawx: if you're publishing just one thing per repo, call it "master", if you're publishing multiple things, name the branches sanely (and maybe leave "master" at "whatever upstream master was") | 13:36 |
|
cpfr
| note i am the sole author on my remote branch | 13:36 |
|
bremner
| cpfr: git-filter-branch | 13:36 |
| ← ciaran left | 13:36 |
|
Tv
| cpfr: if it's the latest commit, git commit --amend | 13:37 |
| → ciaran joined | 13:37 |
|
mawx
| Tv: thanks a lot. Still find it hard to get used to that dvcs way of working. | 13:37 |
| → lolage0 joined | 13:37 |
|
cpfr
| Tv, bremner thanks | 13:37 |
|
Tv
| cpfr: you may also find git rebase -i easier than filter-branch; filter-branch is a batch thingie, rebase -i is interactive | 13:37 |
|
cpfr
| its actually a few in | 13:37 |
| ← mawx left | 13:38 |
|
Mikachu
| rebase only works if there are no merges on top though | 13:39 |
|
cpfr
| Mikachu, there are no merges | 13:40 |
| ← ciaran left | 13:40 |
| → ciaran joined | 13:40 |
| ← ciaran left | 13:41 |
| → ciaran joined | 13:42 |
| ← ciaran left | 13:44 |
| → ciaran joined | 13:44 |
| ← ciaran left | 13:45 |
| → ciaran joined | 13:45 |
|
cpfr
| thanks the man page even had the right thing to run | 13:47 |
| ← ciaran left | 13:47 |
| ← stouset left | 13:47 |
| → ciaran joined | 13:48 |
|
bremner
| I'm following the using-merge-subtree recipe, and the second step I get "merge with strategy ours failed". Any ideas? | 13:48 |
| → gitte joined | 13:49 |
| ← cpfr left | 13:49 |
| ← cncfanatics left | 13:51 |
| ← vbabiy left | 13:51 |
| → niki joined | 13:52 |
|
bremner
| NVM: if forgot to commit a delete. | 13:53 |
| → pozic joined | 13:53 |
| → ajonat_ joined | 13:55 |
|
pozic
| Can I see the diff between the previous version of the repo and the new version? I am not referring to working copy changes. | 13:59 |
|
Ilari
| pozic: 'git diff <commit1> <commit2>'? | 14:00 |
|
| pozic: Commits need to be both in same repo... | 14:00 |
|
pozic
| Ilari: yes, but I don't know what to fill in in <commit1> and <commit2>. | 14:01 |
|
bremner
| in this case git diff HEAD HEAD^ | 14:01 |
| ← lolage0 left | 14:01 |
|
bremner
| ? | 14:01 |
|
| pozic: maybe you want gitk? | 14:01 |
|
pozic
| Where is the syntax documented for that command? The manpage on my system doesn't document it. | 14:01 |
|
bremner
| pozic: you mean HEAD and HEAD^ ? | 14:02 |
|
pozic
| bremner: for example. | 14:02 |
| → lolage0 joined | 14:02 |
|
Ilari
| pozic: Do you even know what versions you want to compare? | 14:02 |
|
pozic
| Ilari: I suppose HEAD^ is the version before HEAD that I had before I pulled. | 14:03 |
|
| Ilari: which is exactly what I was interested in. | 14:03 |
|
bremner
| pozic: I guess HEAD and HEAD^ should be in most tutorials | 14:03 |
|
pozic
| bremner: proper software has a readable manpage, imho. | 14:03 |
|
bremner
| pozic: in general a pull can add more than one commit. That is why gitk might help | 14:04 |
|
Ilari
| pozic: maybe 'HEAD@{1}' and 'HEAD'? | 14:04 |
|
| pozic: HEAD@{1} points to commit that HEAD (current branch pointer) was at before the last operation... | 14:06 |
|
| pozic: See 'git reflog'. | 14:06 |
|
bremner
| pozic: I guess in this case you want user-manual.{html,txt} I don't think it is a man page, but on e.g. debian it is in /usr/share/doc/git-doc | 14:08 |
| ← AlexB left | 14:08 |
|
pozic
| bremner: I looked in share/doc/git. Thanks | 14:09 |
| ← ajonat left | 14:09 |
|
pozic
| That's not really consistent. Some packages have their docs in $program, some need an additional package installed $package-doc | 14:09 |
|
bremner
| pozic: git-doc is a separate package, there are many many such cases in debian | 14:09 |
|
pozic
| Or maybe there is some rule I don't know about like size of documentation etc. | 14:09 |
|
| bremner: yes, I know that. I just said how it worked. | 14:10 |
|
| gitk looks horrible, but is nice after some change of preferences. | 14:10 |
|
| Default is 9mm letters. | 14:10 |
|
crab
| really? gitk looks pretty awful for me no matter what i do to it. | 14:13 |
|
Tv
| worse than the size, it seems the default font ends up horribly bitmapped ragged edges on my ubuntu | 14:14 |
|
Mikachu
| if you have tk 8.5 it gets better | 14:14 |
|
| before that, tk didn't have antialiased fonts | 14:14 |
| ← a-atwood left | 14:14 |
|
Tv
| ah still at 8.4 | 14:14 |
|
pozic
| I changed the font too. | 14:16 |
|
bremner
| pozic: re: package-splitting. No hard and fast rules, but there are good reasons. #debian-mentors on OFTC can explain if you really care | 14:17 |
| ← gitte left | 14:17 |
|
Mikachu
| those huge .h files | 14:17 |
|
crab
| damn, no tk8.5 packages for ubuntu. | 14:18 |
| → a-atwood joined | 14:18 |
| → igorgue joined | 14:20 |
| → _slacker_ joined | 14:21 |
| → madewokherd joined | 14:21 |
| ← ciaran left | 14:26 |
| → ciaran joined | 14:26 |
| ← _slacker_ left | 14:27 |
| ← devcow left | 14:30 |
| → capisce joined | 14:31 |
| ← capisce_ left | 14:32 |
| → radiomark joined | 14:35 |
| → alegscogs joined | 14:40 |
| → mbuf joined | 14:41 |
| ← SuttoL left | 14:41 |
| ← linash left | 14:44 |
| → krawek joined | 14:46 |
| → rubydiamond joined | 14:47 |
| → SuttoL joined | 14:47 |
| → ForsakenSoul joined | 14:47 |
|
ForsakenSoul
| hey guys ... can someone tell me if there is a way to return a pull ? | 14:47 |
| → gitte joined | 14:48 |
| → kukks joined | 14:49 |
|
bremner
| ForsakenSoul: you mean undo it in your local repo? git reset should work | 14:49 |
|
ForsakenSoul
| yeah in my local repo ... so | 14:49 |
| → elight joined | 14:49 |
|
ForsakenSoul
| git reset should return my clone to the state before the last pull | 14:49 |
|
| ? | 14:49 |
|
bremner
| according to what was discussed today, git reset HEAD@{1} (man git-reflog) | 14:50 |
|
| but I never did this myself | 14:50 |
|
doener
| bremner: reset --hard I'd say | 14:50 |
|
Mikachu
| depends on if the pull succeeded or not | 14:50 |
|
ForsakenSoul
| yeah succeeded | 14:50 |
|
| with no conflicts | 14:50 |
|
doener
| ForsakenSoul: and check if HEAD@{1} is correct with "git reflog show $branchname" | 14:50 |
|
Tv
| mm tk8.5 looks so much better | 14:50 |
|
| anyone with modern debian/ubuntu, you want sudo update-alternatives --config wish | 14:51 |
|
doener
| ForsakenSoul: you should pick the one that matches your pre-pull state and reset --hard to that (uncommitted changes are lost) | 14:51 |
|
ForsakenSoul
| well doener ... I've only done pull | 14:52 |
|
| no other work | 14:52 |
|
| this is a lib I downloaded .... I've actually did only one pull | 14:52 |
|
doener
| then HEAD@{1} is likely right and there's nothing to lose, but I'd check anyway, it's not like it would take much time ;-) | 14:53 |
|
ForsakenSoul
| so first | 14:53 |
|
| git reflog show $master | 14:53 |
|
| and afterwards git reset HEAD@{1} | 14:54 |
|
| ? | 14:54 |
|
doener
| without the $, that was just to signal that you shouldn't type "branchname", but the name of the actual branch | 14:54 |
|
ForsakenSoul
| ok :D | 14:54 |
|
doener
| and it's: git reset --hard HEAD@{1} (replace the 1 with the correct entry from the reflog, though 1 is likely to be correct) | 14:55 |
| → bobesponja joined | 14:55 |
|
doener
| ah, crap, master@{1} | 14:55 |
|
| HEAD@{1} would also do in this case, but you're looking at the master reflog with the above command (which is usually less flooded and easier to handle) | 14:56 |
|
pozic
| Tv: that's default on Debian. | 14:56 |
|
Tv
| maybe on a fresh install | 14:56 |
|
| and ubuntu just lags | 14:56 |
| → kewark joined | 14:56 |
| → priidu joined | 14:57 |
|
pozic
| Tv: I am talking about Unstable | 14:57 |
|
Tv
| pozic: maybe you just no longer have tk8.4 | 14:57 |
|
pozic
| Tv: I checked | 14:57 |
|
| Tv: I have both | 14:57 |
|
| Tv: actually, I have three versions installed. | 14:58 |
|
Tv
| pozic: umm, i believe you already | 14:58 |
|
pozic
| Tv: but update-alternatives only lists two. | 14:58 |
| → capisce_ joined | 14:58 |
|
jast
| hmm, it's still very easy to see it's tk :} | 15:00 |
| → kumbayo joined | 15:00 |
|
jast
| but it's better indeed | 15:00 |
| ← capisce left | 15:02 |
| ← krawek left | 15:11 |
| ← garrin left | 15:11 |
|
SRabbelier
| What's the best way to do a reverse lookup on branchname? (e.g., from hash to branchname) | 15:13 |
|
vmiklos
| git describe? | 15:15 |
|
SRabbelier
| vmiklos: don't that only do tags? | 15:15 |
|
vmiklos
| see the --all option | 15:16 |
|
SRabbelier
| vmiklos: made of awesome, thanks :) | 15:17 |
|
vmiklos
| np | 15:17 |
|
SRabbelier
| :) | 15:17 |
| ← alegscogs left | 15:17 |
| → alegscogs joined | 15:19 |
| ← gitte left | 15:21 |
| → ciaran_ joined | 15:22 |
| ← ciaran left | 15:22 |
| ← pozic left | 15:23 |
|
Bombe
| t9101-git-svn-props.sh from v1.5.6.1 fails tests 25 and 26, is that to be expected? :) | 15:25 |
| ← _zack left | 15:25 |
| ← Sonderblade left | 15:26 |
| ← priidu left | 15:27 |
| ← rubydiamond left | 15:28 |
| ← ciaran_ left | 15:30 |
| → ciaran joined | 15:30 |
| → naeu joined | 15:32 |
| ← ciaran left | 15:32 |
| → garrin joined | 15:32 |
| ← alegscogs left | 15:45 |
| ← elight left | 15:45 |
| → stouset joined | 15:46 |
| → advany joined | 15:52 |
| → rubydiamond joined | 15:57 |
| ← pygi left | 16:01 |
| stouset → stouset|away | 16:04 |
| ← G_SabinoMullane left | 16:08 |
| → ben__ joined | 16:12 |
| → ciaran joined | 16:14 |
| ← crab left | 16:16 |
|
naelyn
| is there a way in git-svn to define multiple branches or tags folders to import git-branches from during a 'git svn fetch' ? | 16:17 |
| ← Ori_B left | 16:20 |
| → tenpaiyomi joined | 16:21 |
| ← ciaran left | 16:21 |
|
Ilari
| naelyn: I think that can be done by editing the configuration file... | 16:23 |
|
naelyn
| i thought that the config prefs only allow one tree mapping per line (tags = myproj/tags/*:refs/remotes/tags/*) | 16:25 |
|
lea_w
| The blame command seems to support an undocumented syntax: | 16:27 |
|
Ilari
| naelyn: Yes, but it might be possible to have multiple mappings of the same type... Just be coreful not to have colliding names... | 16:27 |
|
lea_w
| git-blame -- FILE COMMIT | 16:27 |
|
| is this any different from "git-blame COMMIT -- FILE"? | 16:27 |
| → LotRtests joined | 16:28 |
|
naelyn
| one problem that i also have with this is that the different tags&branches locations are overlapping. like: (svn:/tags/published, svn:/tags/deploy, svn:/tags <-- all contain "branch points") | 16:29 |
| → EmilMedve joined | 16:30 |
| ← EmilMedve left | 16:31 |
| ← Eludias left | 16:31 |
|
Ilari
| Aargh... That's what one gets from no first-class support for branches (and tags)... | 16:32 |
| ← garrin left | 16:33 |
| → cncfanatics joined | 16:36 |
| → Cowmoo joined | 16:36 |
|
naelyn
| i know. i have to manually "tend the garden" by flipping on and off 3-5 different .git/config settings for branch&tag mappings after watching the commit email stream at work | 16:37 |
|
| before i can run "git svn fetch" and expect it not to do the wrong thing | 16:37 |
| ← LotRtests left | 16:37 |
|
Cowmoo
| how do I create a local branch originating from a commit previous to the master one (from a remote repo)? | 16:37 |
|
naelyn
| "git checkout -b newbranch master^" | 16:38 |
|
| lea_w looks at builtin-blame.c and discovers that there is more syntax than documented... | 16:38 |
| → ijcd joined | 16:38 |
|
| lea_w is too lazy/busy to send a note to the list, but if anyone wants to take care of it *hint* *hint*... ;-) | 16:38 |
| stouset|away → stouset | 16:39 |
| ← stouset left | 16:39 |
| → SamB joined | 16:39 |
|
SamB
| whoa, you have almost as many people as #haskell in here | 16:40 |
| ← ijcd left | 16:40 |
|
RandalSchwartz
| and we have more untyped fun! | 16:40 |
|
| untyped, but cryptographically secure | 16:40 |
|
| and distributed! | 16:40 |
|
cardioid
| How many people are in #haskell? | 16:40 |
|
SamB
| um, anyway, I was wondering if anyone knows the motivation for git-svn putting everything under refs/remotes/ by default? | 16:40 |
|
| 449 | 16:41 |
|
| RandalSchwartz averts his eyes at the mention of git-svn | 16:41 |
|
cardioid
| I'm guessing the ratio is 80% asking what monads are and 20% trying to explain | 16:41 |
| → ijcd joined | 16:41 |
|
SamB
| cardioid: actually not so much | 16:41 |
|
cardioid
| 5% trying to explain and 15% idling? | 16:41 |
|
| ;) | 16:41 |
|
jengelh
| and the rest talking about git-svn | 16:41 |
|
SamB
| apparantly to qualify as a Haskell newbie you have to write a monad tutorial or something | 16:42 |
| ← ijcd left | 16:42 |
|
cardioid
| Oh, so #haskell is for the people who've already run the gauntlet | 16:42 |
|
SamB
| I seem to have skipped newbie by writing a compiler first ... at least, I probably wrote a compiler at some point ... and hacked on one for sure ... | 16:42 |
| ← rubydiamond left | 16:42 |
| → Fullmoon joined | 16:42 |
| → rubydiamond joined | 16:43 |
| ← naelyn left | 16:43 |
| → capisce joined | 16:43 |
|
SamB
| more seriously, it seems like about 90% idling and 10% talking about ... anything | 16:43 |
|
| (sometimes we actually talk about Haskell and not just FP in general ;-) | 16:44 |
|
jast
| SamB, probably because it's assumed that combining git-svn with other remotes is usually not very practical | 16:44 |
|
SamB
| well, it seems to make it unnecessarily tricky to share the fruits of a git-svn conversion with others... | 16:44 |
| → ckoehler joined | 16:45 |
|
jast
| that's never very nontricky, really | 16:45 |
|
SamB
| ... any particular reason why the tags aren't imported as tags? | 16:45 |
|
jast
| probably because in svn tags can technically be changed | 16:46 |
|
SamB
| oh. | 16:46 |
|
jast
| and with that i mean they can have history | 16:46 |
|
doener
| SamB: tags are not immutable in svn | 16:46 |
|
| hm, too slow :-( | 16:46 |
|
jast
| you used other words so you still get points | 16:46 |
|
SamB
| wait, you don't get extra points for saying the same exact thing 2 seconds later? | 16:47 |
| → gitte joined | 16:47 |
|
SamB
| anyway ... I guess I'm too much of a darcs user to think of that with SVN's tags ;-) | 16:47 |
|
jast
| if it's EXACTLY the same you might get precognition points | 16:48 |
|
doener
| SamB: in svn, tags are a pure policy thing. It's still a "svn cp something tags/foobar", a plain copy that can get changes as usual | 16:48 |
|
SamB
| true, true | 16:49 |
|
| I don't use SVN very much | 16:49 |
|
jast
| yeah. it's actually really annoying, but that's subversion for you. | 16:49 |
|
SamB
| it's so dang inconvenient | 16:49 |
|
doener
| and as you don't have to commit the copy immediately, but can make changes along the way, you don't even have a guarantee that there's a commit on some branch that matches the tag | 16:50 |
|
SamB
| true | 16:50 |
|
doener
| way too many special cases to add sane "treat svn tags as git tags" support, IMHO | 16:50 |
|
SamB
| though I'd probably have done the copy by SVN "URL" | 16:50 |
| ← zeom left | 16:50 |
| → wuxia joined | 16:51 |
|
wuxia
| is there anyway to use git to version my entire filesystem? I relaize there are some recursivenesses are git stores stuff in .git, which will be on the fs itself ... but maybe some one has a colever solution around this | 16:51 |
|
| SamB doesn't suppose there's much chance of git-svn supporting the merge history they added in SVN 1.5.0 | 16:51 |
|
jast
| wuxia, not a good idea, really. git doesn't deal well with very large files, for instance. | 16:51 |
|
SamB
| yeah, doesn't it have a 4GB per-file limit currently? | 16:52 |
|
jast
| not quite | 16:52 |
|
doener
| SamB: IIRC there were some who wanted such support, no idea if anyone stepped up to add that though. | 16:52 |
|
jast
| but files need to fit into the system's address space and virtual memory | 16:52 |
|
SamB
| for me, that's a much tighter constraint ;-P | 16:53 |
|
wuxia
| hmm; I wnat to have a versioning fs on olinux; where do you guys suggest I look? | 16:53 |
|
doener
| SamB: and as far as I've heard, it's the svn stuff is more like cherry-pick tracking, which might turn out quite tricky | 16:53 |
|
Jc2k
| SamB: bzr-svn supports such things, so its just a matter of poking the git-svn people i guess | 16:53 |
|
Ilari
| Even worse, computing delta diffs/appiles needs to have both versions in dirty memory... | 16:53 |
|
jast
| wuxia, perhaps ext3cow is for you | 16:53 |
|
SamB
| where do the git-svn people congregate? | 16:53 |
|
jast
| nowhere :) | 16:53 |
|
| but there are actually people here who use it | 16:54 |
|
| me, for example | 16:54 |
| ← capisce_ left | 16:54 |
|
Jc2k
| wuxia: also, keep an eye out for wizbit. | 16:54 |
|
SamB
| doener: it might be possible to use some of the information, though | 16:54 |
| → zeom joined | 16:54 |
|
wuxia
| Jc2k: wizbit is a 1986 bbc children's television show -- what shouL i be googling for instead? | 16:54 |
|
Jc2k
| :) | 16:55 |
|
| its also a not-ready project that will be at guadec :) | 16:55 |
|
doener
| SamB: normalperson (Eric Wong) is probably the primary contact for git-svn, plus a bunch of random contributors | 16:55 |
| → pygi joined | 16:55 |
|
wuxia
| Jc2K: is there a beta page somewhere? | 16:55 |
|
Jc2k
| wuxia: http://guadec.expectnation.com/guadec08/public/schedule/detail/87 | 16:55 |
|
| wuxia: its not even that far :) | 16:55 |
|
doener
| SamB: best place to get feedback would likely be a mail to Eric with the git ml in cc | 16:55 |
|
jast
| ext3cow actually exists, you know ;) | 16:56 |
|
Jc2k
| wuxia: we are tidying up the git code to make a "libgitcore". this is nearly far enough along for us to start with the FUSE and GVFS stuff. | 16:56 |
|
wuxia
| cool; thanks | 16:56 |
|
SamB
| would I need to sign up first? | 16:56 |
|
jast
| no | 16:56 |
|
doener
| SamB: no subscription needed, free to post | 16:56 |
|
Jc2k
| jast: but its not using the awesome Git pack to store stuff :D | 16:57 |
|
jast
| the awesome Git pack has a number of limitations that make its use for filesystems infeasible | 16:57 |
|
| mostly due to implementation issues | 16:57 |
|
byronvf
| I've got a small fix for the vc-hg.el file for emacs.. is that maintained by mercurial or emacs? | 16:57 |
|
Jc2k
| jast@ yeah i know, my team is trying to fix them :p | 16:57 |
|
jast
| cool | 16:58 |
|
byronvf
| Oops sorry | 16:58 |
|
jast
| byronvf, well, it's almost right here ;) | 16:58 |
|
byronvf
| :), got the software topic correct I guess | 16:58 |
| → _zack joined | 16:58 |
|
Tv
| jast: care to elaborate? | 16:59 |
| ← ckoehler left | 16:59 |
|
jast
| on the limitations? well, mostly the fact that the binary delta algo requires that both files be loaded into memory | 17:00 |
|
Tv
| jast: ah, yes big files would suck | 17:01 |
|
| jast: apart from that, mmap support conflicting with the idea of deltas, git feels like venti backed by a posix fs to me | 17:02 |
|
| i'd be perfectly happy with having a filesystem where i'm not supposed to write files bigger than 100MB | 17:03 |
|
jast
| i wouldn't | 17:04 |
|
Tv
| well not for all uses | 17:04 |
|
| but you wouldn't to put your temp dir in version control anyway | 17:04 |
|
| so it's not an ext3 killer.. | 17:04 |
| → G_SabinoMullane joined | 17:07 |
| ← zeom left | 17:11 |
| ← Cowmoo left | 17:12 |
| → zeom joined | 17:20 |
| ← glommer left | 17:22 |
| → tzafrir_laptop joined | 17:25 |
| ← peeja left | 17:25 |
|
tzafrir_laptop
| On the kernel git tree I switched to a specific tag of a version using 'git checkout v<version_number>' | 17:25 |
|
| Is there a relatively easy way to check after that what was the last tag I checked out? | 17:26 |
|
| That tree is supposed to be basically read-only, with only builds of the kernel | 17:27 |
| → peeja joined | 17:27 |
|
Ilari
| tzafrir_laptop: You can't switch to tags. That command creates temporary scratch branch starting from the tag. | 17:27 |
| → ciaran joined | 17:27 |
|
Ilari
| tzafrir_laptop: 'git describe' maybe? | 17:27 |
| → LotRtests joined | 17:27 |
|
tzafrir_laptop
| Ilari, thanks, that seems to work | 17:27 |
| → sherbang joined | 17:28 |
|
tzafrir_laptop
| What I basically want is to be able to easily "switch between versions" to test build of something with various kernels | 17:28 |
| → lhunath joined | 17:30 |
|
lhunath
| What's the deal with site_perl and vendor_perl directories? git seems to have put itself in site_perl but Error.pm installs in vendor_perl and vendor_perl isn't included in @INC so obviously git can't find it. Should I add vendor_perl to @INC or should I fix git to install in vendor_perl somehow? | 17:31 |
| → felipec joined | 17:32 |
| ← _ditocujo_ left | 17:33 |
| → nud_ joined | 17:33 |
| → _ditocujo_ joined | 17:33 |
| ← zeom left | 17:34 |
| ← wuxia left | 17:35 |
| → ramblasauruswrec joined | 17:39 |
| ← ramblasauruswrec left | 17:40 |
| ← advany left | 17:40 |
|
tzafrir_laptop
| lhunath, what distribution is it? git installed from source or from packages? official packages or third-party? | 17:40 |
| ← pombreda left | 17:41 |
| ← igorgue left | 17:41 |
| → igorgue joined | 17:41 |
| ← LotRtests left | 17:42 |
| → pombreda joined | 17:42 |
|
lhunath
| tzafrir_laptop: git installed from source using port with the macports distribution on intel osx leopard. | 17:43 |
|
| official macports package. | 17:43 |
|
tzafrir_laptop
| hmm... I don't really know what's the policy of macports | 17:43 |
|
| But sounds like a bug of that package, or maybe of their perl | 17:43 |
|
floppyears
| there's a mac os x installer for git | 17:44 |
|
tzafrir_laptop
| vendor_perl is common on RPM-based perl installations | 17:44 |
|
| floppyears, does that installer update your installation? and keeps track of all dependencies? | 17:44 |
|
floppyears
| http://code.google.com/p/git-osx-installer/ | 17:45 |
|
lhunath
| what's the point of having two perl inclusion paths? | 17:45 |
|
floppyears
| download that and it takes a couple of seconds to install, no compiling needed | 17:45 |
|
lhunath
| floppyears: I doubt that installes the perl dependencies. | 17:45 |
|
tzafrir_laptop
| the point was to separate modules provided by the distributor of perl and modules installed from CPAN and such | 17:46 |
|
lhunath
| git itself works just fine; git-svn and others don't. | 17:46 |
| → zeom joined | 17:46 |
|
lhunath
| I see; so shouldn't both paths be included in @INC? | 17:46 |
|
tzafrir_laptop
| That way whatever you install in a non-standard way (which has more chances of being careless) won't step over your local installs | 17:46 |
|
floppyears
| lhunath: I have that installed any command that I can use to figure out what dependencies it has installed ? | 17:46 |
|
tzafrir_laptop
| sounds like a bug of the macports perl . | 17:46 |
|
lhunath
| floppyears: git-svn | 17:46 |
|
| tzafrir_laptop: non-standard perl modules are to go in site_perl ? | 17:47 |
|
tzafrir_laptop
| Speaking of git-svn: I am still unable to use it with client-side ssl certificates | 17:47 |
|
| lhunath, site_perl is like /usr/local . It is your playground . Naturally you know better than the packager, and hence it has a priority over vendor_perl in the include_path | 17:48 |
| → threeve joined | 17:49 |
| ← nud left | 17:49 |
|
tzafrir_laptop
| lhunath, But then again, those mac porters might have different ideas and standards. Ask them. And also look at how other perl modules ports are installed | 17:49 |
| Arme[0] → Arme[N] | 17:49 |
|
lhunath
| tzafrir_laptop: yeah. asking there too. thanks for the info; it helps clear things out. | 17:50 |
|
tzafrir_laptop
| BTW: on debian there are no site_perl and vendor_perl directly under /usr/lib/perl<whatever> . INC is: | 17:51 |
|
| /etc/perl /usr/local/lib/perl/5.10.0 /usr/local/share/perl/5.10.0 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl . | 17:52 |
| ← threeve left | 17:52 |
|
lhunath
| heh. | 17:53 |
|
tzafrir_laptop
| try also: perl -V | 17:53 |
|
| perl -V:sitelib -V:vendorlib | 17:55 |
| → Alan joined | 17:55 |
| → threeve joined | 18:00 |
| → z3om joined | 18:03 |
| ← zeom left | 18:03 |
| z3om → zeom | 18:03 |
| → cncfanatics1 joined | 18:06 |
| ← threeve left | 18:06 |
|
| SamB almost wants an option to have git-svn ask the SVN server things thrice | 18:06 |
| ← cncfanatics left | 18:06 |
|
doener
| SamB: ? | 18:08 |
| ← peeja left | 18:08 |
| ← igorgue left | 18:09 |
|
SamB
| I keep getting some kind of mismatched checksum error | 18:09 |
|
doener
| SamB: which git version? | 18:09 |
|
SamB
| I'm fairly certain it's a flaky SVN server | 18:10 |
| ← yofel left | 18:10 |
|
SamB
| but, right now I'm using version 1:1.5.6-1 of the Debian packages | 18:10 |
|
| and I'd prefer not to end up with a corrupt import just because the SVN server flakes out the same way twice | 18:11 |
| → joydivider joined | 18:11 |
|
doener
| not necessarily. The performance improvements that went into 1.5.6 had at least one reproducable bug (which got fixed before the 1.5.6 release IIRC), and someone else reported a similar issue not to long ago | 18:11 |
|
| I don't remember if he just told about it here or actually reported it to the list | 18:12 |
|
| SamB: the import should just be aborted and not cause your repo to be broken, you can just try again and continue at the last good state with "git svn fetch" | 18:12 |
|
Bombe
| Hmm... | 18:13 |
|
SamB
| doener: well, if the SVN server sends the same bad data several times, git svn fetch will probably think it is good data... | 18:13 |
|
Bombe
| git init; echo hello > file; git commit -m "test" file; git tag 0.1; git tag 0.2; git describe --all | 18:13 |
|
| Should that line write "0.1" or "0.2"? | 18:13 |
| Arme[N] → Arme[0] | 18:14 |
|
SamB
| yes? | 18:14 |
|
doener
| hm... | 18:14 |
|
| is that svn repo pulblic? | 18:14 |
|
| s/pulblic/public/ | 18:14 |
|
SamB
| yeah | 18:15 |
| ← ciaran left | 18:15 |
| → ciaran joined | 18:15 |
| ← sherbang left | 18:15 |
|
doener
| Bombe: depends on how fast you do that ;-) if the creation time of the tags is equal, it's probably random. | 18:15 |
| ← byronvf left | 18:15 |
|
doener
| Bombe: otherwise, I read the docs so that it should say 0.2 | 18:15 |
|
Bombe
| doener, well, put a sleep 60 between the two tags, then. :) | 18:16 |
|
SamB
| svn://scm.gforge.inria.fr/svn/coq | 18:16 |
|
Bombe
| I have a repo that has two tags at the same commit, from two different days. | 18:16 |
|
| git describe shows the earlier tag. | 18:16 |
|
| Contrary to what the documentation says. | 18:16 |
|
SamB
| the documentation should be updated to say something else ;-P | 18:16 |
|
doener
| Bombe: does that command actually work for you? git commit bails out here | 18:17 |
| → igorgue joined | 18:18 |
|
Bombe
| Yeah, forgot a "git add file" before the commit. | 18:18 |
| ← pombreda left | 18:18 |
|
doener
| Bombe: 0.1 here, too | 18:18 |
|
Bombe
| git init; echo hello > file; git add file; git commit -m "test" file; git tag 0.1; sleep 2; git tag 0.2; git describe --all | 18:18 |
|
| This line works, though, and prints 0.1. | 18:18 |
| ← ciaran left | 18:18 |
|
Bombe
| Actually, it prints "tags/0.1" but that's still wrong, I'd say. :) | 18:18 |
| → pombreda joined | 18:18 |
|
doener
| Bombe: yeah, I fixed the "git add" part myself, just wondered if there was some change in behaviour that made your git version accept the original command | 18:19 |
| → priidu joined | 18:19 |
|
Bombe
| doener, no, that was just an error. :) | 18:19 |
|
| Anyway, I have to go. See you later. :) | 18:19 |
| → pergesu joined | 18:20 |
| ← _zack left | 18:23 |
| ← joydivider left | 18:23 |
|
doener
| SamB: http://kerneltrap.org/mailarchive/git/2008/6/25/2221294 | 18:26 |
| ← zeom left | 18:27 |
|
doener
| SamB: does that (approx.) match what you experience? | 18:28 |
| → advany joined | 18:28 |
| → sbeyer joined | 18:30 |
| → elight joined | 18:31 |
| → rdavila joined | 18:33 |
| → _slacker_ joined | 18:34 |
|
SamB
| doener: yeah, sorta | 18:35 |
|
| he doesn't mention what happens if he attempts to resume | 18:35 |
|
| sometimes that works for me, sometimes not... | 18:35 |
|
doener
| I assumed his "ih I run the same command multiple times" to refer to the fetch part only | 18:37 |
|
| which kind of is like resuming, because he never gets past that single revision | 18:37 |
|
SamB
| yes, but I'm not at all familiar with -r | 18:37 |
|
doener
| limits the range of revisions to fetch, in his case, only 3095 | 18:38 |
| ← _ditocujo_ left | 18:38 |
| ← priidu left | 18:38 |
|
doener
| nice to avoid fetching a few thousand revisions just to test something that you can reproduce with a single one :-) | 18:38 |
| → foca joined | 18:38 |
| → _ditocujo_ joined | 18:39 |
|
doener
| or to start somewhere else than revision 1 | 18:39 |
|
| btw, how many revisions is the clone I'm running in the background going to fetch? I'm at ~1720 now | 18:39 |
| ← jhelwig left | 18:40 |
|
SamB
| hitting problems at 2403 crashing at 2403 ri | 18:40 |
|
| er. | 18:40 |
|
| (I should remember to delete stuff that I don't need anymore in my IRC messages...) | 18:41 |
|
doener
| ok, so that's reasonably close to my current state | 18:41 |
| ← ben__ left | 18:42 |
|
SamB
| okay, fetching just one refision is not enough to trigger it... | 18:42 |
|
| er. | 18:42 |
|
| revision... | 18:42 |
| → bdiego joined | 18:43 |
|
doener
| you can specify ranges like -r10-20 | 18:43 |
|
| ok, died at 2320 here | 18:44 |
| → joydivider joined | 18:45 |
|
doener
| and "svn fetch" continuously gets the same wrong hash | 18:46 |
|
SamB
| so the solution is to switch back to the previous version? | 18:46 |
| → jhelwig joined | 18:46 |
|
SamB
| doener: I think that's because it's a PREVIOUS version that's bad | 18:46 |
| ← _slacker_ left | 18:47 |
|
doener
| well, I'll now try to read the perl code without getting a major headache and find out what's going on | 18:47 |
|
| don't hold your breath | 18:47 |
| ← gitte left | 18:47 |
| ← rdavila left | 18:48 |
| ← jezdez left | 18:49 |
|
doener
| SamB: oh, oops, ranges are -r10:20, actually. | 18:49 |
|
SamB
| 'sokay, that's the syntax I used naturally | 18:50 |
|
| I've used SVN enough to know that syntax ;-) | 18:50 |
|
doener
| heh :-) | 18:50 |
| ← mbuf left | 18:54 |
|
doener
| SamB: uh, ouch, crap, my import failed on .depend and that got totally messed up in r2314 here | 19:00 |
| → yofel joined | 19:01 |
|
doener
| SamB: I'm going to send a bug report to the list, if you want to get cc'd let me know your email address please (privmsg is ok) | 19:02 |
| ← kewark left | 19:04 |
| → \ask_ joined | 19:08 |
| → zeom joined | 19:10 |
| → krawek joined | 19:10 |
| ← Arrowmaster left | 19:11 |
| → Arrowmaster` joined | 19:12 |
|
SamB
| also, my name is Samuel Bronson in case you didn't see that already | 19:12 |
|
bronson
| hm, another bronson. | 19:14 |
| ← joydivider left | 19:14 |
|
bronson
| good to see. | 19:14 |
|
| doener is quick with /whois :-) | 19:15 |
| → _slacker_ joined | 19:19 |
| → jezdez joined | 19:21 |
| ← pergesu left | 19:27 |
| ← igorgue left | 19:30 |
| ← jezdez left | 19:30 |
| ← nuge left | 19:30 |
| → joydivider joined | 19:31 |
| → jezdez joined | 19:31 |
| → vbabiy joined | 19:32 |
| ← jezdez left | 19:32 |
| → patrick2 joined | 19:32 |
| ← rubydiamond left | 19:34 |
| → igorgue joined | 19:35 |
| → csc` joined | 19:35 |
| → gass joined | 19:35 |
|
patrick2
| hey there, i got a problem with git-core_1.5.5.1-1~bpo40+1_amd64.deb running with etch --- gitt add my2gigXenDisk.img --> fatal: Out of memory, malloc failed --- any suggestions? | 19:35 |
|
Tv
| patrick2: don't put big files in git, that's gonna suck | 19:36 |
|
gass
| what command does the same that an svn export? | 19:36 |
|
Tv
| patrick2: think "source control", okay? | 19:36 |
|
| gass: nothing, but you may want git archive | 19:36 |
|
gass
| Tv, i want to convert a git local repo to a normal dir | 19:36 |
|
Tv
| gass: rm -rf .git | 19:36 |
|
dato
| gass: I normally do `git-archive | tar xC /path/to/dir` | 19:37 |
|
gass
| from the root dir? | 19:37 |
| → jezdez joined | 19:39 |
|
patrick2
| Tv: argh, i thought it would be good idea to develop my xen-images with git | 19:40 |
| ← jezdez left | 19:48 |
| ← lolage0 left | 19:52 |
| → jezdez joined | 19:52 |
| → alegscogs joined | 19:52 |
|
patrick2
| hmmmm, isnt there any other "version control system" for large files? -- i want sth like me@xen1$ touch file1 file2; me@host$ git commit; .. and so on | 19:56 |
| ← alegscogs left | 19:56 |
| → drizzd joined | 19:58 |
|
Tv
| patrick2: you're confusing block device images with their contents, too | 19:58 |
| ← elight left | 20:01 |
| → ciaran joined | 20:01 |
| ← radiomark left | 20:05 |
| ← naeu left | 20:05 |
| → elight joined | 20:06 |
| ← \ask_ left | 20:07 |
| ← elight left | 20:07 |
| Arrowmaster` → Arrowmaster | 20:09 |
|
patrick2
| Tv: hm i think i got u, it would be better to control the versions of the images content -- like me@xen1$ tourch file; me@xen1$ git commit; -- but there are some problems regarding the permissions and so on, wouldnt it? -- ill do some investigation on gibak (http://eigenclass.org/hiki/gibak-backup-system-introduction) -- ty anyway! | 20:09 |
| → \ask_ joined | 20:10 |
| → stouset joined | 20:10 |
| ← _slacker_ left | 20:12 |
| ← joydivider left | 20:12 |
| ← Sho_ left | 20:13 |
| → Sho_ joined | 20:15 |
|
hagabaka
| is git cherry-pick used to split commits? | 20:19 |
| → naeu joined | 20:19 |
| ← igorgue left | 20:20 |
|
hagabaka
| i have multiple changes in the working tree, is there a way to split the changes into multiple commits? they're not separated by file either | 20:20 |
|
Tv
| hagabaka: man git-add | 20:20 |
|
hagabaka
| they're not splitted by file... | 20:21 |
|
doener
| hagabaka: you can use "git add -i" (or -p) to stage individual changes from your working tree | 20:21 |
|
hagabaka
| oh | 20:21 |
|
doener
| hagabaka: you can also read the huge thread (well, almost flamewar by now) on the ml for different approaches | 20:21 |
|
hagabaka
| is stage a command? | 20:22 |
| → igorgue joined | 20:24 |
| ← skotku left | 20:25 |
| ← pygi left | 20:27 |
|
hagabaka
| ok, add -i seems to work well enough for me | 20:27 |
| → raalkml joined | 20:34 |
| ← stouset left | 20:37 |
| ← raalkml left | 20:42 |
| → daaku joined | 20:44 |
| ← kumbayo left | 20:46 |
| ← d0k left | 20:49 |
| ← BabelO left | 20:50 |
| ← lhunath left | 20:52 |
| ← bdiego left | 20:54 |
| → xenoterracide joined | 20:55 |
|
xenoterracide
| should I be asking about zsh filecompletion for git here or on #zsh ... | 20:56 |
|
Mikachu
| the zsh mailing list is probably the best bet | 20:58 |
|
xenoterracide
| filename autocompletion only seems to work relative to the root of the repository, not from my current position | 20:58 |
| → lolage0 joined | 20:58 |
|
xenoterracide
| Mikachu: ah | 20:58 |
| ← message145 left | 20:58 |
| → markkalderon joined | 21:03 |
| ← Fullmoon left | 21:08 |
| → priidu joined | 21:09 |
| ← foca left | 21:10 |
| ← jezdez left | 21:14 |
| → jezdez joined | 21:14 |
| ← flaguy left | 21:18 |
| → pergesu joined | 21:22 |
| → rolfb joined | 21:28 |
| ← felipec left | 21:29 |
| scrooloose_ → scrooloose | 21:34 |
| → gitte joined | 21:34 |
| ← advany left | 21:38 |
| → _slacker_ joined | 21:42 |
|
gass
| how can i erase a commit i did not pushed yet? | 21:44 |
| ← igorgue left | 21:44 |
| metze → metze_away | 21:45 |
|
RandalSchwartz
| is it the most recent commit on a branch? | 21:45 |
|
drizzd
| git reset --soft HEAD^; undoes the latest commit | 21:45 |
|
gass
| thanks | 21:45 |
|
RandalSchwartz
| also - git commit --amend if you just want to patch it | 21:46 |
| ← radarek left | 21:46 |
| → powerfox joined | 21:46 |
|
powerfox
| Hi! I get an error: "*** Please tell me who you are." when commit. But I'm really IDed. | 21:47 |
|
| Any ideas? I do commit from my code. It worked before, but now I get errors. | 21:48 |
| ← _slacker_ left | 21:49 |
|
vmiklos
| what's the output of 'git config user.name' and 'git config user.email'? | 21:51 |
| → igorgue joined | 21:54 |
| ← geckosenator left | 21:55 |
|
drizzd
| by default, git does something like getent passwd $USER | cut -f5 -d: | cut -f1 -d, | 21:56 |
|
| If that's empty, git complains | 21:56 |
|
jengelh
| drizzd: uh? | 21:57 |
|
| since when does it getpwuid(getuid())? | 21:57 |
|
drizzd
| jengelh: see ident.c:260 | 21:57 |
|
SamB
| jengelh: only if you didn't configure it | 21:58 |
|
jengelh
| must be new | 21:58 |
|
gitster
| No. | 21:58 |
|
| See Linus's very original | 21:58 |
|
| e83c5163316f89bfbde7d9ab23ca2e25604af290 | 21:58 |
|
jengelh
| line 260 has a closing brace :p | 21:59 |
|
drizzd
| sorry, it's ident.c:79 | 21:59 |
|
gitster
| "git grep -C20 -e getpwuid e83c5163316f89bfbde7d9ab23ca2e25604af290" | 21:59 |
|
powerfox
| drizzd: it's not empty | 21:59 |
|
| git grep -C20 -e getpwuid is empty | 22:00 |
|
jengelh
| I seem to remember git commit barfed at me when I do not have the author/email set in ~/.gitconfig | 22:00 |
|
drizzd
| powerfox: what about vmiklos' question? | 22:00 |
| ← branstrom left | 22:00 |
|
powerfox
| jengelh: I have a proper ~/.gitconfig\ | 22:00 |
| → branstrom joined | 22:00 |
|
powerfox
| I can do git commit, but my application can't. I try to understand where should I start searching the bug | 22:01 |
|
| It's new behaviour: so git or Qt bug ( | 22:02 |
| → strangy joined | 22:04 |
| → thiago_home joined | 22:07 |
| thiago_home → thiago | 22:07 |
| → mnemoc joined | 22:10 |
|
mnemoc
| hi, there is any way more elegant than $( git-branch | grep '^*' | cut -c3- ) to get the name of the current branch ? | 22:11 |
|
cncfanatics1
| is it possible to access a git repos from behind a *very* restrictive firewall ? | 22:12 |
|
Jc2k
| cncfanatics1: how restricted? http? | 22:13 |
|
cncfanatics1
| no app can access the internet except internet explorer basicly | 22:13 |
|
| and there's no proxy settings in IE set so its not a proxy | 22:13 |
|
RandalSchwartz
| it's probably just a proxy that browser-sniffs | 22:14 |
|
SamB
| can you view websites under hh.exe? | 22:14 |
|
RandalSchwartz
| but that would seem dumb | 22:14 |
| → eternaleye joined | 22:14 |
|
mnemoc
| faking headers? | 22:14 |
|
cncfanatics1
| I'm currently not at said place though | 22:14 |
|
Jc2k
| isa transparent proxy maybe | 22:15 |
|
| ewww | 22:15 |
|
SamB
| well, that is consistant with the fact that you are speaking with us via IRC | 22:15 |
|
cncfanatics1
| SamB: not really, enough web clients around | 22:15 |
|
| would I somehow be able to access my git repos ? | 22:16 |
|
| some http based client or something | 22:16 |
| ← daaku left | 22:16 |
|
Ilari
| powerfox: Set GIT_{AUTHOR,COMMITTER}_{NAME,EMAIL}? | 22:17 |
| cncfanatics1 → cncfanatics | 22:17 |
|
SamB
| cncfanatics: you would only be able to read them, I *think* | 22:17 |
| → daaku joined | 22:17 |
|
Jc2k
| tunnels FTW | 22:18 |
|
| can you ping? | 22:18 |
|
| :p | 22:18 |
|
powerfox
| Ilari: what's this? | 22:18 |
|
Ilari
| cncfanatics: If the firewall is on the same machine, somehow uses "trusted" computing or does something very "exotic", you are basically SOL... | 22:18 |
|
SamB
| cncfanatics: at least, that's true of the easy, obvious way | 22:18 |
|
| which is using GIT repository accessed via HTTP... | 22:18 |
|
SRabbelier
| jengelh: ping? | 22:19 |
|
Ilari
| powerfox: Environment variables which control the idents used. IIRC, just setting one NAME, EMAIL pair is enough (but I don't offhand recall which one)... | 22:19 |
| ← Sho_ left | 22:19 |
|
SamB
| I suppose there's also the option of somehow faking an HTTPS connection | 22:19 |
|
Ilari
| Some really paranoid "firewalls" MITM TLS connections... | 22:19 |
|
cncfanatics
| its a hardware solution | 22:20 |
|
| nothing software | 22:20 |
|
| for the record :p | 22:20 |
|
powerfox
| Ilari: for my user everything ok, but I don't have these vars | 22:21 |
| → Sho_ joined | 22:21 |
|
cncfanatics
| more info about it when I'm at work, can't really check much from here | 22:21 |
|
SamB
| or maybe git has some kind of commit-by-POST-to-CGI backend? | 22:21 |
|
cncfanatics
| SamB: would an SSH tunnel to my local machine on port 80 do ? | 22:21 |
|
Ilari
| No, it hasn't... | 22:21 |
|
SamB
| cncfanatics: do't think so | 22:21 |
|
| since you said only IE can get through | 22:22 |
|
| implying they at least look at the HTTP headers | 22:22 |
|
Ilari
| Not even SSH on port 443. SSH-in-TLS has some chance of working (through even it may fail)... | 22:22 |
| → _slacker_ joined | 22:22 |
|
mnemoc
| openvpn is great doing that | 22:22 |
|
Ilari
| Telnets on port 443? :-) | 22:23 |
|
cncfanatics
| hmmm | 22:23 |
|
mnemoc
| openvpn looks "like" https | 22:23 |
|
powerfox
| Ok, it uses name == git_default_name, so some vars are broken. Have to look on Qt ((( | 22:23 |
|
cncfanatics
| mnemoc: any link ? | 22:23 |
| ← chris2 left | 22:24 |
|
Ilari
| http://openvpn.net? | 22:24 |
|
mnemoc
| cncfanatics: http://www.google.com/search?q=openvpn | 22:24 |
|
cncfanatics
| hehe, tnx, I'm in a lazy mood | 22:24 |
| ← priidu left | 22:24 |
|
powerfox
| git-var -l is really ok... | 22:25 |
| → threeve joined | 22:25 |
|
cncfanatics
| can someone remember me what that web backend for remote control was on linux ? | 22:25 |
|
| can't recall the name :s | 22:25 |
| ← xenoterracide left | 22:27 |
| ← \ask_ left | 22:28 |
| ← powerfox left | 22:29 |
| ← SuttoL left | 22:29 |
| ← nud_ left | 22:29 |
| → lll joined | 22:30 |
| ← Yuuhi left | 22:31 |
| → xenoterracide joined | 22:32 |
| → nuge joined | 22:32 |
| ← mtodd left | 22:32 |
| ← vbabiy left | 22:39 |
| → Sonderblade joined | 22:39 |
| → _zack joined | 22:40 |
| ← _zack left | 22:40 |
| ← krawek left | 22:43 |
| → mtodd joined | 22:44 |
| → krawek joined | 22:44 |
| ← strangy left | 22:49 |
| ← robinr left | 22:52 |
| ← Ariella left | 22:53 |
| → unreal_ joined | 22:54 |
| ← unreal left | 22:54 |
| unreal_ → unreal | 22:55 |
| ← krawek left | 22:56 |
|
plopix
| how do I git-push all branches and all tage to a public repository? I thought that --mirror is doing it, but it isn't | 23:02 |
|
vmiklos
| git push --all? | 23:04 |
|
plopix
| --all dosn't push tags | 23:05 |
|
vmiklos
| git push --tags? :) | 23:05 |
|
| (note: --tags pushes tags _only_) | 23:06 |
|
plopix
| aehm, I forgot to mention that I am searching for a command that does it _one_ line | 23:06 |
| ← thiago left | 23:07 |
| → capisce_ joined | 23:08 |
|
mnemoc
| plopix: sh -c "git push --all && git push --tags" :) | 23:08 |
| ← drizzd left | 23:09 |
| → deavid joined | 23:09 |
|
plopix
| why is it seperated that strict? | 23:11 |
|
doener
| hm, the man page says taht --mirror should include tags though | 23:12 |
|
| s/taht/that/ | 23:12 |
| → thiago joined | 23:13 |
|
plopix
| --mirror has deleted all tags in my public repos. I was able to restore all tags using a second push line with option --tags | 23:15 |
|
doener
| that sounds like a bug, which git version? | 23:15 |
|
mnemoc
| if foo has remote=origin,merge=master, and I `git push foo` ... will it go to origin/master or origin/foo ? | 23:16 |
|
doener
| mnemoc: merge is for "pull" only IIRC | 23:17 |
|
plopix
| doener: v1.5.6 | 23:17 |
|
doener
| mnemoc: the branch.$name.merge config entry that is. That is what you meant, right? | 23:17 |
|
mnemoc
| doener: yes, i do want it "pull-only" | 23:18 |
|
doener
| mnemoc: should be fine then. And after all, there's is no merge step during push ;-) | 23:18 |
|
mnemoc
| :) | 23:18 |
| ← naeu left | 23:19 |
| ← capisce left | 23:20 |
|
doener
| plopix: works fine here | 23:22 |
|
plopix
| hmm | 23:22 |
| → Ariella joined | 23:25 |
| ← SamB left | 23:26 |
| → capisce joined | 23:30 |
| ← cncfanatics left | 23:31 |
| ← capisce_ left | 23:38 |
| → SamB joined | 23:41 |
| ← jezdez left | 23:42 |
|
patrick2
| /j #ocaml | 23:44 |
|
davidfetter
| heh | 23:44 |
| ← rikur left | 23:46 |
| → rikur joined | 23:47 |
| → schacon joined | 23:50 |
| ← lea_w left | 23:50 |
|
schacon
| hey all - what is git connecting over the git protocol on this address? localhost[0: ::1] | 23:51 |
|
SamB
| schacon: that's YOU | 23:52 |
| → lea_w joined | 23:52 |
|
schacon
| ohh, i'm an idiot - it's because I added that to my hosts file a long time ago and i'm running the daemon on localhost. sorry, thanks SamB | 23:53 |
|
SamB
| ::1 is the new 127.0.0.1 -- didn't they tell you? | 23:54 |
|
schacon
| i was told, but i wasn't paying attention | 23:55 |