| 2008-07-03 |
|
pasky
| [futurama voice] good news, everyone! | 00:00 |
|
| i might get to do a lot of hacking on the repo.or.cz duct-tape as part of my upcoming internship ;) | 00:00 |
|
| => smoother, nicer, more featureful and more up-to-date repo.or.cz | 00:00 |
|
Mikachu
| hooray | 00:00 |
|
| pasky peers at http://www.codeplex.com/ ... intriguing | 00:01 |
| ← vbabiy left | 00:01 |
|
_schulte_
| Eridius|work, did that, got the following output | 00:01 |
|
| | 00:01 |
|
| Counting objects: 35, done. | 00:01 |
|
| Compressing objects: 100% (23/23), done. | 00:01 |
|
| Writing objects: 100% (23/23), 3.40 KiB, done. | 00:01 |
|
| Total 23 (delta 16), reused 0 (delta 0) | 00:01 |
|
| To [email@hidden.address] | 00:01 |
|
| * [new branch] eschulte/rinari -> eschulte/rinari | 00:01 |
|
| but then no change on the actual github page?? | 00:01 |
|
Eridius|work
| the github page shows master by default | 00:01 |
|
| if you reload the page, the branch dropdown should list eschulte/rinari | 00:02 |
| ← insin left | 00:02 |
|
_schulte_
| yup, it does and that up to date | 00:02 |
|
| so if I can impose on your kindness for one more question... | 00:03 |
|
| how can I push these new changes to the master branch? | 00:03 |
|
Eridius|work
| well, my first question is if you want it to be on the master branch, why are you working on eschulte/rinari locally instead of master? | 00:04 |
|
_schulte_
| no idea :) I think I must have branches when I was trying to pull changes from the other repo | 00:05 |
| ← jdahm left | 00:06 |
|
Eridius|work
| my recommendation is you should be using the master branch | 00:07 |
|
| so `git checkout master && git reset --hard eschulte/rinari` | 00:07 |
|
| if you want to remote the eschulte/rinari branch from github you can run `git push origin :eschulte/rinari` | 00:07 |
|
Mikachu
| s/remote/remove/ | 00:08 |
| → huy1 joined | 00:08 |
|
Eridius|work
| err, yeah | 00:08 |
| → markkald_ joined | 00:09 |
|
_schulte_
| wonderfull, back to master with the github matching my local | 00:10 |
| ← timocratic left | 00:10 |
|
_schulte_
| thanks for the help1 | 00:10 |
|
Eridius|work
| np | 00:11 |
| ← _schulte_ left | 00:12 |
| → adante joined | 00:13 |
| dwmw2 → dwmw2_gone | 00:14 |
|
bryanray
| how do i pull a specific tag? | 00:16 |
|
| git pull [tag name] doesn't seem to want to work. | 00:16 |
| → branstrom_ joined | 00:17 |
|
pasky
| aaaaa | 00:17 |
|
Eridius|work
| bryanray: what about `git fetch --tags`? | 00:17 |
|
| or maybe `git fetch remote refs/tags/tagname` | 00:17 |
| → srvpnt joined | 00:17 |
|
pasky
| oh man, don't you hate it when you stop reading a thread just when it seems to go on some unrelated tangent, write your own reply, and seconds after sending it you decide to read the "tangent" and see that it is actually exactly what you wrote? :) | 00:18 |
|
Eridius|work
| heh | 00:18 |
| ← jbunster left | 00:20 |
|
Bombe
| Okay, bug reported to mailing list. | 00:22 |
|
| Man, I sure have been productive today! | 00:22 |
| ← booldog left | 00:22 |
|
srvpnt
| Hi guys, I have poked around a lot but can't seem to put together the right search terms for my problem | 00:26 |
|
| I have a fairly simple repo that takes forever to compress | 00:26 |
| ← markkalderon left | 00:26 |
|
srvpnt
| There are two files that just cause the machine to grind forever | 00:26 |
| → lorandi joined | 00:27 |
|
srvpnt
| my .git/objects/pack is about 150 megs | 00:27 |
| ← aroben left | 00:27 |
|
srvpnt
| but there aren't any giant files anywhere | 00:27 |
|
Eridius|work
| and it's still huge after `git repack -adf`? | 00:28 |
|
| or maybe -Adf | 00:28 |
|
srvpnt
| I will try again. | 00:28 |
|
| It is on a remote server | 00:28 |
|
| is there a good way to set it up to do that after x commits? | 00:28 |
| → doener_ joined | 00:28 |
|
Eridius|work
| no, but you could use cron to do it every X days | 00:29 |
|
Mikachu
| yes there is | 00:29 |
|
Eridius|work
| there is? | 00:29 |
|
Mikachu
| well, almost | 00:29 |
| ← Ademan left | 00:29 |
|
Mikachu
| just call git gc --auto in a post-push hook, whatever it's called | 00:29 |
|
Eridius|work
| ah yes, gc --auto. | 00:29 |
| → Ademan joined | 00:29 |
|
Eridius|work
| of course, I'm never sure exactly what command that invokes | 00:29 |
|
srvpnt
| haha | 00:29 |
|
| thanks guys, I will play around with that and let you know | 00:30 |
|
Eridius|work
| if `git repack -adf` leaves a giant pack behind, and you have nothing large in your repo, you might want to check reflogs and the like | 00:30 |
|
srvpnt
| thanks I will check that too | 00:31 |
| → marc_vw joined | 00:31 |
| ← sergiodj left | 00:31 |
| → albertito joined | 00:31 |
| ← branstrom left | 00:32 |
| ← trochala left | 00:33 |
| ← albertito left | 00:34 |
| ← markkald_ left | 00:35 |
| ← EmilMedve left | 00:36 |
| ← digitalhobbit left | 00:38 |
| ← k0001 left | 00:39 |
| ← doener left | 00:42 |
| ← bryanray left | 00:43 |
| ← cnk left | 00:43 |
| → blowery joined | 00:46 |
| ← mvw left | 00:47 |
| → felipe joined | 00:48 |
| ← pigeon left | 00:49 |
| ← SoreGums left | 00:52 |
| → pigeon joined | 00:55 |
|
dsop
| I wonder why there is a option indexVersion, but a grep over the git repository of git doesen't show any hit but the documentation of git config | 00:56 |
| → albertito joined | 00:59 |
|
Mikachu
| dsop: typo in config.txt, it should be indexversion | 01:05 |
|
| (case) | 01:05 |
| → vbabiy joined | 01:05 |
| → stouset joined | 01:07 |
| → bremner` joined | 01:07 |
| ← sbeyer left | 01:11 |
| ← bremner left | 01:14 |
|
Mikachu
| (sent a patch now) | 01:15 |
| ← joeyh left | 01:18 |
| → igorgue joined | 01:21 |
| ← david_koontz left | 01:23 |
|
vmiklos
| 'The variable names are case-insensitive and only alphanumeric characters are allowed.' | 01:25 |
|
| from man git-config | 01:25 |
| → lea_w joined | 01:25 |
| → joeyadams joined | 01:25 |
| ← njal left | 01:28 |
| → VVSiz_ joined | 01:28 |
|
joeyadams
| If I used git-clone git://anongit.freedesktop.org/git/xorg/xserver to fetch data from a git repository, what's the best way to update to the latest revision (as in, the equivalent to svn up)? | 01:28 |
|
Mikachu
| vmiklos: i was thrown because the code that compares uses strcmp, not strcasecmp | 01:29 |
|
| vmiklos: i suppose they're downcased somewhere else | 01:29 |
|
| hm, how does that work if you have a branch foo and a branch FOO and try to use some branch.foo.remote or so? | 01:30 |
|
vmiklos
| i never use uppercase chars in branch names .) | 01:33 |
|
| i just remember it confused me as well when i read fooBarBaz in the docs and foobarbaz worked fine in the config ;) | 01:34 |
| ← lea_w left | 01:39 |
| ← tvachon left | 01:39 |
| ← bronson_ left | 01:45 |
| ← VVSiz left | 01:46 |
| ← ijcd_ left | 01:46 |
|
cardioid
| joeyadams: Have you made any changes at all? | 01:48 |
| → Leefmc joined | 01:49 |
|
joeyadams
| cardioid> No | 01:49 |
|
cardioid
| joeyadams: git pull | 01:49 |
|
joeyadams
| Okay | 01:49 |
|
| What if I have? | 01:49 |
| → irotas joined | 01:49 |
|
cardioid
| Or if you care to see what's really going on, git fetch and then merge the changes yourself (which is essentially what git pull does) | 01:49 |
| ← Fullmoon left | 01:50 |
|
cardioid
| If you have, than your tree may not be in a good position for pulling, so there may be conflics | 01:50 |
|
| conflicts | 01:50 |
|
| and then you'd most likely want to fetch+merge | 01:50 |
| ← pigeon left | 01:55 |
| → digitalhobbit joined | 01:56 |
| ← pombreda left | 02:00 |
| ← warthog9 left | 02:01 |
| → pigeon joined | 02:01 |
| → jbunster joined | 02:02 |
| → garrin joined | 02:02 |
| → spearce joined | 02:05 |
| ← jbunster left | 02:07 |
| → jbunster joined | 02:07 |
| → bryanray joined | 02:09 |
| ← warlock_mza left | 02:13 |
| → DrNick2 joined | 02:17 |
| ← daaku left | 02:18 |
| ← DrNick left | 02:18 |
| ← wagle left | 02:19 |
| ← jbunster left | 02:24 |
| ← albertito left | 02:25 |
| ← bdiego left | 02:26 |
| → albertito joined | 02:26 |
| → entie joined | 02:27 |
|
bryanray
| how can i view a list of tags in a repository? | 02:28 |
| → pergesu joined | 02:28 |
|
spearce
| git tag -l | 02:29 |
|
bryanray
| Thanks, man. | 02:30 |
|
srvpnt
| thanks Eridius and Mikachu, I got rid of some old reflogs and repacked and now compressing takes a very reasonable amount of time | 02:30 |
|
Eridius|work
| oh good | 02:30 |
|
srvpnt
| have a good night! (or morning, if that's your thing) | 02:32 |
| ← srvpnt left | 02:32 |
| ← lorandi left | 02:33 |
|
cardioid
| git tag without any arguments lists all tags | 02:33 |
|
| -l is if you want to match a pattern | 02:33 |
| → ijcd joined | 02:34 |
| ← ijcd left | 02:35 |
| ← vbabiy left | 02:38 |
| ← albertito left | 02:39 |
| ← Eridius|work left | 02:40 |
| → albertito joined | 02:41 |
| → warthog9 joined | 02:43 |
| ← xinming left | 02:43 |
| ← blowery left | 02:43 |
| ← irotas left | 02:43 |
| → xinming joined | 02:44 |
| → crab joined | 02:44 |
| → aroben joined | 02:47 |
| → dduncan joined | 02:50 |
| ← ciaran left | 02:52 |
| → wagle joined | 02:52 |
| → ciaran joined | 02:53 |
| → daaku joined | 02:54 |
| ← heycarsten left | 02:55 |
| → xinming_ joined | 03:00 |
| ← fade-in left | 03:00 |
| ← pigeon left | 03:01 |
| → pigeon joined | 03:02 |
| ← kanru left | 03:03 |
| → nuge joined | 03:04 |
| ← bobesponja left | 03:04 |
| → bobesponja joined | 03:04 |
| ← xinming_ left | 03:05 |
| ← wagle left | 03:05 |
| → wagle joined | 03:06 |
| → lea_w joined | 03:06 |
| ← joeyadams left | 03:07 |
| ← entie left | 03:07 |
| ← stouset left | 03:08 |
|
crab
| i wonder why i'm not getting list mail. | 03:08 |
|
spearce
| vger was taken offline to do upgrades? | 03:08 |
| → xinming_ joined | 03:09 |
|
spearce
| though i'm pretty sure the mx accepted a message i sent about 30 minutes back. but it hasn't been redistributed yet to the list. | 03:09 |
| → skrit joined | 03:09 |
|
crab
| spearce: oh, is that already? damn, i didn't realise. (yes, it accepted my messages too.) | 03:10 |
|
davidfetter
| hello | 03:10 |
| → ajonat joined | 03:11 |
|
davidfetter
| i have users carping at me that when they run git clone, they get | 03:11 |
|
| > > > fatal: The remote end hung up unexpectedly | 03:11 |
|
| > > > fetch-pack from 'git://git.postgresql.org/git/~davidfetter/postgresql/.git' failed. | 03:11 |
|
| any ideas what might be the problem? | 03:11 |
|
spearce
| is that url really correct? "/git/~davidfetter/" ? not "/~davidfetter/" ? | 03:12 |
|
davidfetter
| yep | 03:12 |
| ← xinming left | 03:12 |
|
davidfetter
| http://git.postgresql.org/?p=~davidfetter/postgresql/.git;a=summary | 03:12 |
|
| whether that repo is set up right is a whole 'nother story | 03:13 |
|
crab
| spearce: according to the message warthog9 posted, the downtime is still 12 hours away (1500UTC, july 3rd). | 03:13 |
|
spearce
| do you have direct access to the repository? if so what does `cd ..pathtorepo; git upload-pack .` say? if it spews SHA-1s and branch names and sits there with "0000" the repository is fine and its just the damn url. | 03:13 |
|
| did you set git-export-ok in the repository? | 03:13 |
|
crab
| davidfetter: did you put a git-daemon-export-ok in the repository? | 03:14 |
|
| yeah, what he said | 03:14 |
|
spearce
| crab: downtime early? unexpected crash? patchbomb on linux list taking a while to distribute? | 03:14 |
|
| usually kernel.org is pretty good about sending stuff out, but if it lags i just wait | 03:14 |
|
davidfetter
| crab, it's not my repo | 03:15 |
|
| i wish it were | 03:15 |
|
crab
| df: did $whoever put a git-daemon-export-ok in the repository? | 03:15 |
|
davidfetter
| git upload-pack | 03:16 |
|
| usage: git-upload-pack [--strict] [--timeout=nn] <dir> | 03:16 |
|
spearce
| you didn't pass the "." argument to mean pwd for <dir> | 03:16 |
|
crab
| df: "git-upload-pack ." | 03:16 |
| ← nuge_ left | 03:16 |
|
davidfetter
| ok, so what might the correct url be? | 03:16 |
|
| i got SHA1s and 000 | 03:16 |
|
| 0000* | 03:16 |
|
spearce
| i think it depends on the configuration settings given to git-daemon when it was starte | 03:16 |
|
| d | 03:16 |
|
davidfetter
| btw, i can pull just fine with that URL | 03:16 |
|
| and unfortunately, i need to run like hell and catch a plane in about a minute :( | 03:17 |
| ← krawek left | 03:17 |
|
Mikachu
| spearce: yw :) | 03:18 |
| ← davidfetter left | 03:19 |
|
warthog9
| spearce | crab - yeah I'm not doing it till tomorrow (PDT) | 03:19 |
| → johnw joined | 03:21 |
| → krawek joined | 03:22 |
| ← pflanze left | 03:25 |
|
| spearce teaches jgit about git 1.6's dislike of dashes in commands | 03:26 |
| ← johnw left | 03:26 |
| → johnw joined | 03:26 |
|
crab
| oh yeah. sigh. my fingers nearly always type the - these days. | 03:27 |
|
spearce
| well it tried to run "git-upload-pack" but that's not in my $PATH anymore. its supposed to be, but its not, and uh, we're running on the local system. "git upload-pack" would work better. over SSH, git-upload-pack is still the accepted form. | 03:27 |
| → jackdempsey joined | 03:37 |
| → l0verb0y joined | 03:37 |
| ← albertito left | 03:38 |
| → kanru joined | 03:39 |
|
crab
| spearce: could you please have a look at builtin-describe.c:207 | 03:39 |
|
spearce
| oh. i posted a patch tonight to make that "n->tag ? n->tag->tagged->sha1 : sha1" | 03:40 |
|
| but vger is sitting on it. | 03:40 |
|
| or taking for a walk. | 03:40 |
|
| i don't know what it is doing. | 03:40 |
|
crab
| that n->tag->tagged->sha1 segfaults when n is tagged with an unannotated tag. but in that case (tagname-0-sha1) isn't it always ok to use n->sha1? | 03:40 |
|
| spearce: heh, i posted a patch to do that too :-) | 03:40 |
| → digitalhobbit_ joined | 03:41 |
| → albertito joined | 03:41 |
| lamont` → lamont | 03:44 |
|
crab
| oh well. | 03:44 |
| ← crab left | 03:47 |
|
| spearce wonders why robinr wanted native smb:// support in jgit | 03:47 |
| → bremner joined | 03:51 |
| ← kanru left | 03:52 |
| → fhobia joined | 03:53 |
| → kanru joined | 03:55 |
| ← kanru left | 03:58 |
| ← digitalhobbit left | 03:58 |
| → imyousuf joined | 03:58 |
| ← bremner` left | 03:58 |
| ← ckoehler left | 03:59 |
| → syscrash joined | 03:59 |
| ← johnw left | 04:00 |
| ← digitalhobbit_ left | 04:00 |
|
syscrash
| i'm having some trouble with a few repositories. i get permission denied errors during push, on files in objects/. the repo was created with git init --shared; however, directories under objects/ are still being created with the wrong group name | 04:01 |
|
| any suggestions? | 04:01 |
| → kanru joined | 04:02 |
| ← spearce left | 04:02 |
| ← Leefmc left | 04:08 |
| → mbuf joined | 04:09 |
| ← ajonat left | 04:13 |
| ← VVSiz_ left | 04:16 |
| → VVSiz_ joined | 04:16 |
| → digitalhobbit joined | 04:17 |
| → Cacher joined | 04:19 |
| ← felipe left | 04:19 |
| ← drizzd left | 04:19 |
| ← biz left | 04:19 |
| ← thijso left | 04:19 |
| ← jpieper left | 04:19 |
| ← zeom left | 04:19 |
| ← mxey left | 04:19 |
| ← Cacheaway left | 04:19 |
| ← jamesjb left | 04:19 |
| ← jmspeex left | 04:19 |
| ← shad_ left | 04:19 |
| ← EgS left | 04:19 |
| ← Sput left | 04:19 |
| ← lnostdal left | 04:19 |
| ← awx left | 04:19 |
| ← MarcWeber left | 04:19 |
| ← bartman left | 04:19 |
| ← IRSeekBot left | 04:19 |
| ← nonix4 left | 04:19 |
| ← pkufrank1 left | 04:19 |
| ← siprbaum left | 04:19 |
| ← dato left | 04:19 |
| ← betagan left | 04:19 |
| ← fil left | 04:19 |
| ← godog left | 04:19 |
| ← exg left | 04:19 |
| ← metellius left | 04:19 |
| ← Illocution left | 04:19 |
| ← Bombe left | 04:19 |
| ← shawn_ left | 04:19 |
| ← Ingmar left | 04:19 |
| ← aeruder left | 04:19 |
| ← drfontus left | 04:19 |
| ← zlin left | 04:19 |
| ← danderson left | 04:19 |
| ← xaiki left | 04:19 |
| ← cliffstah left | 04:19 |
| ← ddollar left | 04:19 |
| ← jones- left | 04:19 |
| → mxey_ joined | 04:19 |
| → Bombe joined | 04:20 |
|
gitster
| "man chgrp(1)"? New directories inherit the group of parent directory. | 04:20 |
| → drizzd joined | 04:20 |
| → biz joined | 04:20 |
| → thijso joined | 04:20 |
| → MarcWeber joined | 04:20 |
| → jpieper joined | 04:20 |
| → zeom joined | 04:20 |
| → Cacheaway joined | 04:20 |
| → jamesjb joined | 04:20 |
| → betagan joined | 04:20 |
| → jmspeex joined | 04:20 |
| → drfontus joined | 04:20 |
| → awx joined | 04:20 |
| → shad_ joined | 04:20 |
| → EgS joined | 04:20 |
| → Sput joined | 04:20 |
| → lnostdal joined | 04:20 |
| → bartman joined | 04:20 |
| → nonix4 joined | 04:20 |
| → pkufrank1 joined | 04:20 |
| → shawn_ joined | 04:20 |
| → cliffstah joined | 04:20 |
| → fil joined | 04:20 |
| → metellius joined | 04:20 |
| → xaiki joined | 04:20 |
| → aeruder joined | 04:20 |
| → godog joined | 04:20 |
| → dato joined | 04:20 |
| → Illocution joined | 04:20 |
| → siprbaum joined | 04:20 |
| → ddollar joined | 04:20 |
| → Ingmar joined | 04:20 |
| → jones- joined | 04:20 |
| → zlin joined | 04:20 |
| → danderson joined | 04:20 |
| → exg joined | 04:20 |
| ← exg left | 04:20 |
| → exg joined | 04:20 |
|
gitster
| spearce: I see it; thanks. | 04:20 |
| ← danderson left | 04:20 |
| ← cliffstah left | 04:20 |
| → danderson joined | 04:20 |
| ← shad_ left | 04:20 |
| → wagle_home joined | 04:20 |
| ← biz left | 04:21 |
| ← MarcWeber left | 04:21 |
| → shad joined | 04:21 |
| → biz joined | 04:21 |
| ← jamesjb left | 04:21 |
| ← Ingmar left | 04:21 |
| → MarcWeber joined | 04:21 |
| → jamesjb joined | 04:21 |
| ← zlin left | 04:21 |
| ← aeruder left | 04:21 |
| → Ingmar joined | 04:21 |
| → aeruder joined | 04:21 |
| ← Sput left | 04:21 |
| ← thijso left | 04:21 |
| ← Cacheaway left | 04:21 |
| danderson → Guest91257 | 04:21 |
| ← Illocution left | 04:21 |
| → thijso joined | 04:21 |
| → Illocution joined | 04:21 |
|
syscrash
| gitster: i'll look into it, thanks! | 04:22 |
| → nonix4_ joined | 04:22 |
| → IRSeekBot joined | 04:22 |
| → fil_ joined | 04:24 |
| → xaiki` joined | 04:24 |
| ← wagle left | 04:24 |
| DrNick2 → DrNick | 04:25 |
| ← Ingmar left | 04:25 |
| ← biz left | 04:25 |
| ← shad left | 04:25 |
| ← Guest91257 left | 04:25 |
| ← exg left | 04:25 |
| ← siprbaum left | 04:25 |
| ← dato left | 04:25 |
| ← awx left | 04:25 |
| ← pkufrank1 left | 04:25 |
| ← bartman left | 04:25 |
| ← zeom left | 04:25 |
| ← fil left | 04:25 |
| ← jpieper left | 04:25 |
| ← EgS left | 04:25 |
| ← godog left | 04:25 |
| ← metellius left | 04:25 |
| ← nonix4 left | 04:25 |
| ← lnostdal left | 04:25 |
| ← shawn_ left | 04:25 |
| ← drfontus left | 04:25 |
| ← drizzd left | 04:25 |
| ← xaiki left | 04:25 |
| ← ddollar left | 04:25 |
| ← jones- left | 04:25 |
| ← jmspeex left | 04:25 |
| ← betagan left | 04:25 |
| → aroben_ joined | 04:26 |
| ← aroben left | 04:26 |
| → shawn_ joined | 04:27 |
| → bartman joined | 04:27 |
| ← bremner left | 04:27 |
| ← jackdempsey left | 04:27 |
| → metellius joined | 04:27 |
| → pkufranky joined | 04:28 |
| ← tjafk2 left | 04:29 |
| → xinming joined | 04:29 |
| → zeom joined | 04:29 |
| → Ingmar joined | 04:30 |
| → ddollar joined | 04:30 |
| ← ___tortoise left | 04:31 |
| → shad joined | 04:31 |
| → biz joined | 04:32 |
| → Guest91257 joined | 04:32 |
| → exg joined | 04:32 |
| → drizzd joined | 04:32 |
| → jpieper joined | 04:32 |
| → betagan joined | 04:32 |
| → jmspeex joined | 04:32 |
| → drfontus joined | 04:32 |
| → awx joined | 04:32 |
| → EgS joined | 04:32 |
| → lnostdal joined | 04:32 |
| → nonix4 joined | 04:32 |
| → pkufrank1 joined | 04:32 |
| → fil joined | 04:32 |
| → xaiki joined | 04:32 |
| → godog joined | 04:32 |
| → dato joined | 04:32 |
| → siprbaum joined | 04:32 |
| ← EgS left | 04:32 |
| → exg_ joined | 04:32 |
| → EgS joined | 04:32 |
| ← siprbaum left | 04:32 |
| → jones- joined | 04:32 |
| → siprbaum joined | 04:32 |
| ← awx left | 04:32 |
| → tjafk2 joined | 04:32 |
| → biz_ joined | 04:32 |
| → danderson joined | 04:32 |
| → awx_ joined | 04:33 |
| ← lnostdal left | 04:33 |
| ← pkufrank1 left | 04:33 |
| ← godog left | 04:33 |
| ← drizzd left | 04:33 |
| → lnostdal joined | 04:33 |
| ← biz left | 04:35 |
| biz_ → biz | 04:35 |
| → drizzd joined | 04:35 |
| → zlin joined | 04:35 |
| → godog joined | 04:35 |
| ← thiago left | 04:36 |
| ← Guest91257 left | 04:37 |
| ← fil left | 04:37 |
| ← nonix4 left | 04:37 |
| ← kukks left | 04:38 |
| ← xaiki left | 04:40 |
| ← igorgue left | 04:40 |
| ← xinming_ left | 04:41 |
|
bryanray
| Anyone have any suggestions for when a "git submodule update" command will not pull down the submodule? | 04:42 |
|
| the directory does not exist, and it just refuses to go get it | 04:42 |
| → xinming_ joined | 04:42 |
| ← nuge left | 04:43 |
| ← queso left | 04:47 |
| ← l0verb0y left | 04:48 |
| → xinming__ joined | 04:48 |
| ← exg left | 04:51 |
| → swtaarrs joined | 04:55 |
| ddollar → ddollar|away | 04:55 |
| ← xinming left | 04:56 |
| → xinming joined | 05:00 |
| ← catdude left | 05:00 |
| ← xinming_ left | 05:00 |
| → brosner joined | 05:05 |
| → bronson_ joined | 05:06 |
| ← xinming__ left | 05:06 |
| ddollar|away → ddollar | 05:07 |
| ← madewokherd left | 05:07 |
| ← xinming left | 05:08 |
| → xinming joined | 05:08 |
| → foysavas joined | 05:10 |
|
foysavas
| hey guys | 05:11 |
| → nuge joined | 05:11 |
|
foysavas
| what's the best way to clone a specific branch? | 05:11 |
|
syscrash
| you have to clone the whole repository afaik | 05:12 |
|
| then you can checkout the branch you want | 05:12 |
|
foysavas
| git clone ... | 05:12 |
|
| git checkout stable | 05:12 |
|
| would work? | 05:12 |
|
bryanray
| it should | 05:13 |
|
foysavas
| hm.. my only issue is that when i checkout stable it deletes a bunch of directories that aren't in master | 05:13 |
|
| any suggestions? | 05:14 |
|
RandalSchwartz
| then those aren't in stable | 05:15 |
|
| otherwise it wouldn't delete them | 05:15 |
|
| git diff master...stable | 05:15 |
|
lea_w
| I'd like to check when a (bare) repository has last been modified by only stat'ing the directories in the repository (because it's faster than stat'ing every file). When a repository gets pushed to, does git ever modify files in place, or can I rely on the fact that when any file gets changed, the directory containing it gets changed, too? | 05:22 |
|
| (And with "stat'ing the directories" I mean getting their mtimes, in particular.) | 05:24 |
| ← schlort left | 05:24 |
| marc_vw → mvw | 05:24 |
|
RandalSchwartz
| what do you mean "bare repo has been modified"? | 05:26 |
|
| you mean adding new objects? | 05:26 |
| ← dduncan left | 05:29 |
| ← DrNick left | 05:29 |
| → DrNick joined | 05:30 |
|
gitster
| "git push" may (1) leave a new pack in $GIT_DIR/objects/pack/; (2) explode new loose objects in $GIT_DIR/objects/??/?{38}; or (3) transfer _no_ objects whatsoever. | 05:30 |
|
| And in any case it will update $GIT_DIR/refs/{heads,tags}/* somewhere. | 05:30 |
|
| at least, that would be the kinds of change gitweb person would be interested in. | 05:31 |
|
| Now, when you create a new file in $GIT_DIR/objects/aa/a{38}, mtime of $GIT_DIR/objects/ itself would not change ;-) So you are talking about 257 stat in cases (1) and (2). | 05:32 |
| ← eikonos left | 05:32 |
|
gitster
| But because we have case (3), and in any case we will be changing refs anyway, perhaps it is useless to monitor the changed time in objects/ hierarchy at all. | 05:32 |
|
| If push creates (or deletes) refs/heads/lw/gitweb, mtime of refs/heads does not change, so you would need find "refs/ -ls | sortdate" or something, where sortdate script is http://git.pastebin.org/48129 | 05:34 |
|
| A push may also trigger post-update hook that result in gc to pack refs, so you would also need to look at $GIT_DIR/packed-refs. | 05:35 |
| → pombreda joined | 05:36 |
| → schlort joined | 05:36 |
|
gitster
| which would mean "find $GIT_DIR/refs $GIT_DIR/packed-refs -ls | sortdate | head -n 1", but presumably you are operating in Perl so it can be written more reliably and simpler. | 05:36 |
| ← schlort left | 05:37 |
|
lea_w
| gitster: Yup, I thinking about only stat'ing $GIT_DIR and the refs directory and all its subdirectories. | 05:38 |
|
| Do I have to look at packed-refs at all? I mean, if it's updated by push, doesn't push do some kind of create-and-move-in-place operation that touches the directory as a side effect? | 05:39 |
|
gitster
| Are false positives tolerated? | 05:39 |
|
lea_w
| Yup, that's fine. | 05:39 |
|
| When could they happen? | 05:39 |
|
gitster
| That is, if you say "this repository was updated 30 minutes ago" and there are no changes the users care about... | 05:39 |
|
| Cronbot could come and run "repack" or "gc", for example. | 05:40 |
|
| If you monitor $GIT_DIR/ mtime, that is. | 05:40 |
|
lea_w
| Oh yeah, that "last updated" information I'd actually get separately. I just need some kind of "last touched" timestamp to use in cache keys for entries like the HEAD hash, etc. | 05:40 |
|
| So if there's a false positive it just re-reads all refs and re-resolves all non-SHA1 object identifiers (like 'HEAD'). | 05:41 |
|
| So no big deal. | 05:41 |
|
gitster
| good. | 05:42 |
|
lea_w
| gitster: But can I rely on push touching the directory somehow? Let's say push pushes a new HEAD -- will $GIT_DIR's mtime be updated as well? | 05:42 |
|
| That way I wouldn't have to stat each file. | 05:42 |
|
gitster
| Typically push goes to individual branches, not HEAD, so if push updates refs/heads/master, mtime for refs/heads/ may change but that is about it. | 05:43 |
|
lea_w
| OK. | 05:43 |
|
gitster
| Of course, if HEAD points at the master, that push will _affect_ what is visible when the user asks for HEAD, though... | 05:44 |
|
lea_w
| Sure. But what if the HEAD ref actually gets changed, to a different branch say. | 05:44 |
| → schlort joined | 05:44 |
|
lea_w
| Do I have to stat the HEAD file for its mtime, or is it enough to stat $GIT_DIR? | 05:44 |
| ← aroben_ left | 05:45 |
|
gitster
| "stat" of HEAD would give you when the repository's current branch changed, which would also be an interesting change. I think it is updated by creating a new file in $GIT_DIR, writing to it and then renaming it to HEAD, so $GIT_DIR's mtime would change when it happens. | 05:46 |
| → lirit joined | 05:47 |
|
gitster
| ... but you should double check. I am not sitting in front of the source right now ;-) | 05:47 |
| ← stillbourne left | 05:47 |
| ← xinming left | 05:47 |
|
RandalSchwartz
| why not just use a post-recieve hook? | 05:47 |
|
| have it touch a file | 05:48 |
|
| then you'll always know | 05:48 |
| → entie joined | 05:48 |
|
gitster
| she does not have authority over what the repository owners do, unfortunately. | 05:48 |
|
lea_w
| gitster: Thanks! I'll try that and see if it gives any problems. :) | 05:48 |
|
| RandalSchwartz: Yup, it's hard to implement server-wide. | 05:48 |
|
RandalSchwartz
| ahh | 05:48 |
| ← entie left | 05:48 |
|
RandalSchwartz
| ok I'll walk away before I try to understand the problem being solved | 05:49 |
|
| RandalSchwartz walks away | 05:49 |
|
gitster
| k.org gitweb is what we are talking about. | 05:49 |
|
lea_w
| Perhaps Petr will like it, too. ;) | 05:49 |
|
gitster
| repo.or.cz has much better contorl over what hook repo will use, actually. | 05:49 |
|
| you as a repository owner cannot have shell access there, so it is much easier to contorl what happens when users push into their repository from the site owner's point of view. | 05:50 |
| → xinming joined | 05:51 |
| ← krawek left | 05:53 |
| → bremner joined | 05:54 |
| ← \ask left | 05:56 |
| ← bgerber left | 05:57 |
| → bgerber joined | 05:59 |
| VVSiz_ → VVSiz | 06:02 |
| ← aradtke left | 06:02 |
| ← lirit left | 06:05 |
| → deavid joined | 06:12 |
| ← Ademan left | 06:17 |
| → Radar joined | 06:17 |
|
Radar
| I have a SSL certificate I am using for Firefox at the moment to get to an internal site. We're switching our repositories over to git, and the URL for it is https://<our site>/repos... now how do I get git to use that SSL certificate? | 06:18 |
| → Ademan joined | 06:18 |
|
gitster
| "man git-config" and look for http.sslCert, perhaps? | 06:19 |
| → SuttoL joined | 06:19 |
|
skrit
| anyone know if/when the stgit website (procode.org) is coming back? | 06:21 |
|
Radar
| gitster: thanks for that :) Now we're getting... | 06:23 |
|
| Cannot get remote repository information. | 06:23 |
|
| Perhaps git-update-server-info needs to be run there? | 06:23 |
|
| Even after running the git-update-server-info command. | 06:23 |
|
drizzd
| Radar: update-server-info has to be run server-side | 06:25 |
|
Radar
| Yes we ran it server side. | 06:25 |
|
| Does it matter which folder it's ran in? | 06:25 |
|
drizzd
| no | 06:25 |
|
| you mean which subdirectory? | 06:25 |
|
| Radar: I think you also get this error when the URL is wrong or when the host is unreachable. | 06:26 |
| → charon joined | 06:27 |
| ← bryanray left | 06:27 |
|
Radar
| yes drizzd. The server is accessible through firefox. | 06:29 |
| ← brosner left | 06:29 |
|
Radar
| I am using the same SSL certificate in firefox and it works fine. I specified the sslCert as "git config --global http.sslCert /path/to/the/cert" | 06:29 |
| → entie joined | 06:30 |
|
drizzd
| Radar: set the GIT_CURL_VERBOSE environment variable and try again | 06:33 |
|
Radar
| drizzd: set it with anything? | 06:34 |
|
drizzd
| Radar: yes, just not the empty string | 06:34 |
|
Radar
| * unable to use client certificate (no key found or wrong pass phrase?) | 06:34 |
| → xinming_ joined | 06:34 |
|
Radar
| I didn't see anywhere that I had to set a pass phrase. | 06:34 |
| → Masca joined | 06:37 |
| → bentob0x joined | 06:38 |
|
Radar
| drizzd: The error I'm getting is that I have set a wrong pass phrase, but I never set one. | 06:40 |
| ← entie left | 06:40 |
| → Sput joined | 06:41 |
| → entie joined | 06:41 |
|
Radar
| Would it be easier to set up a proper git repository (git://<machine>) and use keys? | 06:47 |
| ← xinming left | 06:47 |
| → aLTer joined | 06:48 |
| ← digitalhobbit left | 06:49 |
| → sverrej_ joined | 06:50 |
|
drizzd
| Radar: not having a passphrase should be ok, it also says 'no key found'. | 06:51 |
|
Radar
| yes. | 06:51 |
|
| The key has a passphrase. | 06:51 |
|
drizzd
| Radar: if you can use git+ssh, use that. it's certainly the more commonly used protocol. And performance is a lot better too. | 06:51 |
| → lea_w1 joined | 06:52 |
| ← lea_w left | 06:52 |
| lea_w1 → lea_w | 06:52 |
|
Radar
| The other developer I'm working with is adamant that we use HTTPS. | 06:52 |
|
drizzd
| you just said you never set a passphrase | 06:52 |
|
| Radar: I'd _only_ use HTTP or HTTPS if there is no other option. | 06:52 |
|
Radar
| Yeah, I meant it never asked for one. | 06:52 |
|
| Apparently there is no other. | 06:53 |
|
| option | 06:53 |
|
drizzd
| well, I gtg. good luck | 06:54 |
|
Radar
| Ok. Thanks. | 06:54 |
| ← bobesponja left | 06:54 |
|
Radar
| I'd love to do it over SSH (in fact, I got it working yesterday myself) but this other developer is all paranoid about security. | 06:54 |
| → marc_vw joined | 06:54 |
|
drizzd
| Radar: use git-shell then | 06:55 |
|
Radar
| drizzd: what? | 06:56 |
| → BabelO joined | 06:56 |
| ← BabelO left | 06:56 |
|
Masca
| does git have database outside of the workspace just like monotone does? (I mean a foo.git or something) | 06:56 |
|
Radar
| drizzd: just go, we'll figure something out. | 06:56 |
| → BabelO joined | 06:57 |
| ← BabelO left | 06:57 |
| → xinming joined | 06:57 |
| → tbf joined | 06:58 |
|
tbf
| well, just want to say thank your for the git-grep command | 06:59 |
|
| no idea how it manages to be that fast | 06:59 |
|
| well, alone the benefit of not having to specify wildcards for selecting interesting files makes it invaluable | 07:00 |
| exg_ → exg | 07:00 |
| ← sverrej left | 07:03 |
| ← Xyrio left | 07:04 |
| ← mvw left | 07:05 |
| ← rrregis left | 07:05 |
| → Xyrio joined | 07:07 |
| → xinming__ joined | 07:09 |
| ← skrit left | 07:14 |
| → volante joined | 07:14 |
| ← xinming__ left | 07:15 |
| → xinming__ joined | 07:15 |
| ← swtaarrs left | 07:17 |
|
volante
| i want to take 1 branch of a remote repository and use it as 'master' in my local one. i will keep pulling from there to keep it up to date and won't do any modifications. should I use clone or fetch for that? | 07:17 |
|
| at the moment i can't work out how to do either | 07:18 |
| ← xinming left | 07:22 |
| ← xinming_ left | 07:22 |
| → Fullmoon joined | 07:26 |
|
bartocc
| is there an easy way to apply a commit that was made on branch A to branch B | 07:27 |
|
| but just one commit | 07:27 |
|
wagle_home
| git-cherry-pick? | 07:27 |
|
beeble
| git cherrypick | 07:27 |
|
| +- | 07:27 |
|
bartocc
| ok, lemme have a look ;-) | 07:27 |
| ← Fullmoon left | 07:28 |
| ← Ademan left | 07:30 |
| → xinming joined | 07:31 |
| ← bronson_ left | 07:33 |
|
bartocc
| awesome !! | 07:33 |
| dwmw2_gone → dwmw2 | 07:33 |
| ← sverrej_ left | 07:34 |
| → priidu joined | 07:38 |
| x-oweb → X-Blade | 07:39 |
| ← aLTer left | 07:40 |
| ← CodeOfficer left | 07:40 |
| ← alegscogs left | 07:40 |
| → alegscogs joined | 07:42 |
| → Ademan joined | 07:45 |
| ← alegscogs left | 07:46 |
| → reel_ joined | 07:49 |
| ← entie left | 07:49 |
| ← imyousuf left | 07:55 |
| ← xinming__ left | 07:56 |
| → redondos_ joined | 07:58 |
| ← DrNick left | 08:02 |
| ← redondos left | 08:02 |
| redondos_ → redondos | 08:02 |
| → DrNick joined | 08:02 |
| → peritus_ joined | 08:03 |
| → ph^ joined | 08:04 |
| → sverrej joined | 08:06 |
| → mcella joined | 08:10 |
| ← Radar left | 08:13 |
| ← huy1 left | 08:16 |
| → huy1 joined | 08:21 |
| → yofel joined | 08:23 |
| → trochala joined | 08:24 |
| nonix4_ → nonix4 | 08:28 |
| → drizzd_ joined | 08:28 |
| ← decostop left | 08:32 |
| → _zack joined | 08:34 |
| ← eMBee left | 08:40 |
| → eMBee joined | 08:40 |
| ← drizzd left | 08:41 |
| ← yofel left | 08:41 |
| → nud joined | 08:43 |
| ← metellius left | 08:43 |
| ← ciaran left | 08:43 |
| → redondos_ joined | 08:44 |
| ← redondos_ left | 08:45 |
| → xinming_ joined | 08:49 |
| → rrregis joined | 08:49 |
| ← _zack left | 08:51 |
| → metellius joined | 08:52 |
| → masm joined | 08:53 |
| → hliusv561 joined | 08:54 |
| ← eternaleye left | 08:55 |
| → _zack joined | 08:55 |
| → koke joined | 08:56 |
| → xinming__ joined | 08:57 |
| → bremner` joined | 08:57 |
| ← bremner left | 08:57 |
| → koke_ joined | 08:58 |
| → sbodo_w joined | 08:58 |
| ← CrnkyGee1 left | 09:01 |
| ← xinming left | 09:02 |
| → _slacker_ joined | 09:02 |
|
masm
| I have this branch where I want to have a patch from another branch applied, but I want to have exactly the same code in HEAD. Right now I'm using git-merge followed by git-revert. But sometimes I get conflits, etc., and it is a lot of work. Is there some way to do what I want without the hard work? | 09:02 |
| ← peritus_ left | 09:03 |
|
jengelh
| git-cherry-pick | 09:03 |
| → CrnkyGeek joined | 09:04 |
| ← rrregis left | 09:07 |
| → rrregis joined | 09:07 |
| ← xinming_ left | 09:11 |
| → Misha joined | 09:13 |
| → bremner joined | 09:13 |
| Misha → Mishakun | 09:14 |
| → radarek joined | 09:15 |
| ← koke left | 09:16 |
| ← daaku left | 09:16 |
| ← bremner` left | 09:19 |
| → fisons joined | 09:24 |
|
| gitster wonders if we are better off installing a 'bot that periodically says "git-cherry-pick" ;-) | 09:26 |
|
pasky
| :)) | 09:27 |
|
gitster
| hi, long time no talk ;-) | 09:28 |
| → chris2 joined | 09:29 |
|
hacim
| if I am using git-svn to interact with an upstream svn repository, should I not be publishing my git repository? it seems that the git-svn dcommit results in some wacked out stuff due to the rebase it does | 09:31 |
|
hliusv561
| hacim: dcommit first, push only after that? | 09:31 |
| ← ash_ left | 09:32 |
| → ash_ joined | 09:32 |
|
hacim
| hliusv561: perhaps, I'll have to try | 09:32 |
| → ebel joined | 09:33 |
|
hacim
| hliusv561: as soon as i dcommit it breaks the push | 09:35 |
| → d0k joined | 09:37 |
| ← fhobia left | 09:38 |
|
hliusv561
| hacim: The commits get their final IDs in dcommit... | 09:38 |
|
jengelh
| give up, svn is lossy when it is to store git information | 09:39 |
| ← priidu left | 09:39 |
| → priidu joined | 09:40 |
| ← priidu left | 09:40 |
| → priidu joined | 09:40 |
| ← _zack left | 09:45 |
| ← priidu left | 09:45 |
|
lea_w
| Hi everyone (and in particular warthog9, gitster, and pasky): I now have a gitweb with memory caching running here: | 09:48 |
|
| http://odin3.kernel.org/git-lewiemann/ | 09:48 |
|
| It's the same as http://git.kernel.org/ -- just (hopefully) faster. ;) The only things that's still slow is the project list, I'm working on that. | 09:49 |
| → priidu joined | 09:49 |
|
lea_w
| If anything doesn't work there (or is unexpectedly slow), please feel free to send email to [email@hidden.address] or ping me in IRC. | 09:50 |
|
jengelh
| pasky: | 09:51 |
| → peritus_ joined | 09:53 |
| → priidu_ joined | 09:54 |
| → markkalderon joined | 09:57 |
|
hliusv561
| lea_w: That gitweb seems to behave in bizarre manner: Clicking some commits lead to page for a different commit (and usually it then gives that page for all commits). | 09:58 |
|
lea_w
| hliusv561: Right, thanks. | 09:59 |
|
hliusv561
| lea_w: It almost seems like first page displays properly, but subsequent ones don't. | 09:59 |
|
lea_w
| Yeah, it doesn't generate the cache key properly and hence reuses the same commit from cache. | 09:59 |
|
| I'll take a look at it and send a note here when it's fixed. | 09:59 |
|
gitster
| Yeah, you almost gave me a heart attack. | 10:01 |
| → mvo joined | 10:01 |
|
hliusv561
| And how much of that speedup is due to better scalability and how much just due to reduced load? | 10:02 |
|
gitster
| The tip "fast-export" one claims to touch builtin-fast-export, decorate _AND_ git-send-email.perl, and has log message that talks about send-email (on commit page). | 10:02 |
|
pasky
| gitster: hi :) (sorry, I'm a bit busy flatkeeping ;) | 10:02 |
|
| lea_w: I'm very likely to actually get to play with it from next week on | 10:02 |
| → _zack joined | 10:03 |
|
jengelh
| pasky: where to send my repoorcz changes? | 10:03 |
|
lea_w
| It's fixed now, hliusv561, gitster, and pasky. :) | 10:03 |
|
pasky
| to me | 10:03 |
|
lea_w
| Cool, thanks Petr! | 10:03 |
|
jengelh
| pasky: address. | 10:03 |
|
| pasky @suse.cz | 10:03 |
|
lea_w
| hliusv561: Most-to-all of it is due to reduced load right now. | 10:04 |
|
| hliusv561: I'll replay the actual traffic against this one and see how it performs. | 10:04 |
| → snailrails joined | 10:10 |
| → gitte joined | 10:10 |
|
snailrails
| Hi, I've change my repository location. When I try clone the new repo it still tries with my old username, how do I change this ? | 10:11 |
|
| actually this could be a github problem.. | 10:11 |
| ← priidu_ left | 10:12 |
| → jomz joined | 10:14 |
| ← snailrails left | 10:14 |
| ← koke_ left | 10:14 |
| → Xyri0 joined | 10:15 |
| → kez joined | 10:15 |
| → koke joined | 10:15 |
| mxey_ → mxey | 10:17 |
|
kez
| Hullo - playing with Git, seeing if it's for me. I have a central "hub" of which I want to push stuff to from multiple locations - is Git suitable for this? | 10:17 |
|
| I played around with clone/push, but could never get the file sent from my computer to the server | 10:18 |
| ← priidu left | 10:19 |
|
telmich
| kez: that's possible with git | 10:22 |
| → priidu joined | 10:22 |
|
kez
| Ok - could you give a brief list of commands I should look into? | 10:22 |
|
telmich
| kez: git-init, git-remote, git-fetch, git-push, git-merge | 10:22 |
|
| more or less in that order :-) | 10:23 |
|
| it's really eas | 10:23 |
|
| y | 10:23 |
|
kez
| Ok, thanks. And would the server need to "know" about the clients connecting to it | 10:23 |
|
telmich
| no | 10:23 |
|
kez
| Splendid. Thank you : ) | 10:23 |
| → petteri_ joined | 10:23 |
|
pasky
| really, setting up the hub, then cloning and pushing from the sattelite repositories should work smoothly | 10:24 |
| ← Xyrio left | 10:24 |
| Xyri0 → Xyrio | 10:24 |
|
telmich
| kez: I think there are plenty of howtos in the net for doing what you want | 10:24 |
|
kez
| Jup - just wanted to check it was doable | 10:25 |
|
telmich
| kez: I mostly use standard unix permissions (as in ssh, umask, group rights) with git, which avoids having an additional authentication daemon | 10:26 |
|
kez
| Ran it over SSH earlier and seemed to talk ok | 10:27 |
| ← petteri left | 10:36 |
| ← priidu left | 10:36 |
| → priidu joined | 10:38 |
| ← SRabbelier left | 10:43 |
| ← huy1 left | 10:43 |
| → SRabbelier1 joined | 10:43 |
| SRabbelier1 → SRabbelier | 10:44 |
| → janm joined | 10:48 |
| → Fullmoon joined | 10:49 |
| ← priidu left | 10:51 |
| ← wagle_home left | 10:52 |
| → wagle_home joined | 10:52 |
| → zedr0n joined | 10:54 |
|
kez
| Git pull is run from the satellite repositories, or the central server? | 10:55 |
| → priidu joined | 10:56 |
|
hacim
| kez: repositories | 10:56 |
|
kez
| Ah, good - thanks hacim | 10:56 |
| ← Gitzilla left | 10:56 |
|
jengelh
| there is no "central" server in git | 10:57 |
|
kez
| For the time being I'd rather think there was : ) | 10:57 |
| ← SuttoL left | 10:58 |
| ← koke left | 11:06 |
|
jomz
| how can i fix the git-svn & git-submodules issue? i removed the submodule, but when i try to dcommit, it's still failing on the commit where the submodule was added.. can i somehow tell dcommit to use HEAD right away or something? | 11:11 |
|
Tv
| jomz: well you can rebase -i to remove the submodule-related commits | 11:12 |
| ← trochala left | 11:12 |
|
jomz
| what does that do exactly? | 11:13 |
| → sbeyer joined | 11:13 |
| → trochala joined | 11:13 |
|
Tv
| jomz: why don't you read the man page first? | 11:13 |
|
jomz
| yeah.. on it :) | 11:14 |
|
| thanks | 11:14 |
| ← zedr0n left | 11:21 |
| ← _zack left | 11:25 |
| ← priidu left | 11:26 |
| → _zack joined | 11:26 |
| → priidu joined | 11:28 |
|
jomz
| so.. i looked up the git_svn_id from before the submodule stuff, then did "git svn rebase -i git_svn_id" | 11:30 |
|
| this returned Current branch master is up to date | 11:30 |
|
| when i try to dcommit again, i still get the same error | 11:30 |
|
| i also tried --id i.o. -i | 11:30 |
|
Mikachu
| rebase, not svn rebase | 11:34 |
| ← mithro left | 11:35 |
|
jomz
| that gives me "fatal: Needed a single revision | 11:36 |
|
| Invalid base" | 11:36 |
| → pflanze joined | 11:36 |
|
jomz
| -i needs a commit sha1 then or something? | 11:36 |
|
| git rebase -i | 11:37 |
|
| d'oh | 11:37 |
| ← _zack left | 11:46 |
| ← peritus_ left | 11:47 |
| → msh joined | 11:47 |
|
msh
| Hi! | 11:47 |
|
| Sometimes, some backup files are created in all the direcotries by some editors like vim. For example, product.rb~ How do I ignore these files altogather in git? | 11:48 |
|
| Do I need to include .gitignore in each of the directory? | 11:48 |
|
hliusv561
| msh: Put '*~' in .gitignore (located at top of repository)? | 11:49 |
| ← ctennis left | 11:49 |
|
msh
| Yes its on top of repo | 11:49 |
|
| But I think this will NOT ignore in the levels below | 11:49 |
| → joydivider joined | 11:51 |
| → vbabiy joined | 11:52 |
|
hliusv561
| msh: If pattern has no '/' in it, it appiles to all directories below the directory it is in. | 11:53 |
|
msh
| Oh thanks! | 11:53 |
| ← hliusv561 left | 11:53 |
| ← msh left | 11:54 |
| ← deavid left | 12:02 |
| ← quabbl left | 12:03 |
| → _zack joined | 12:04 |
| → felipe joined | 12:05 |
| → emachado joined | 12:06 |
|
pflanze
| What could this mean?: git-cat-file tag v2.6.13.4 => "fatal: git-cat-file v2.6.13.4: bad file" | 12:09 |
|
| git-tag -l *does* show that tag name. | 12:09 |
|
Mikachu
| sure it's an annotated tag? | 12:09 |
|
| % git tag hello | 12:10 |
|
| fatal: git-cat-file hello: bad file | 12:10 |
|
pflanze
| hm | 12:10 |
|
| I see. | 12:10 |
|
| strange things. | 12:10 |
|
Mikachu
| % git tag -a hello -m foo | 12:10 |
|
| % git cat-file tag hello | 12:10 |
|
| object fdb397548ad3e8372954a2bcdf9235a6e024727f | 12:10 |
|
pflanze
| Is there a better way to get the [eventual] tag message than git-cat-file? | 12:11 |
|
| or should I just trust that exit code 128 will mean it's an un-annotated tag? | 12:12 |
|
Mikachu
| you can use git cat-file -t | 12:12 |
| → grahal joined | 12:12 |
|
pflanze
| this will show 'commit' then, I see. | 12:13 |
|
Mikachu
| it could show any of the five types i think | 12:14 |
|
pflanze
| hm I see | 12:14 |
|
Mikachu
| er, four | 12:14 |
| → lorandi joined | 12:14 |
|
pflanze
| But there's an issue: I couldn't check whether it is a tag, or a branch, right? | 12:15 |
|
Mikachu
| if it's a tag it says tag, if it's a commit it says commit | 12:15 |
|
pflanze
| it only says it is a tag if it's annotated | 12:15 |
|
Mikachu
| a branch ref always points to a commit | 12:15 |
|
| yes, otherwise it isn't a tag object | 12:15 |
|
| it's a tag if it's in refs/tags/ | 12:16 |
|
pflanze
| but somehow git still knows it's a tag since git-tag -l will list it. | 12:16 |
|
| git cat-file -p hello will not show tag in your example | 12:16 |
|
| at least, git cat-file -t v2.6.13.4 => commit | 12:17 |
| ← vbabiy left | 12:17 |
|
Mikachu
| is there anything unexpected about that? | 12:18 |
|
pflanze
| my point is, given a string, I couldn't confirm whether that is in fact recorded as a tag or not. | 12:18 |
|
Mikachu
| you know it's a tag if it's in refs/tags/, like i said 3 minutes ago | 12:19 |
|
pflanze
| short of running "git-tag -l" and check for the string. | 12:19 |
|
| ah; well I hope implementation doesn't change then. | 12:19 |
| ← grahal left | 12:19 |
| ← anders_ left | 12:19 |
| → ctennis joined | 12:20 |
|
pflanze
| or, how do I check for refs/tags/ without accessing the filesystem directly? | 12:20 |
| ← garrin left | 12:22 |
| → thiago joined | 12:22 |
| ← joydivider left | 12:22 |
| ← _zack left | 12:22 |
|
| pflanze parses stderr for "bad file" as a solution for now | 12:23 |
| → DrNick2 joined | 12:24 |
| ← DrNick left | 12:25 |
| → deavid joined | 12:28 |
| ← lorandi left | 12:28 |
| → lorandi joined | 12:29 |
| → _zack joined | 12:29 |
| ← trochala left | 12:30 |
| → grahal joined | 12:33 |
| → seanstickle joined | 12:33 |
| ← SRabbelier left | 12:35 |
|
lea_w
| pflanze: git-for-each-ref, if you don't mind some inefficiency (it's reasonably fast on warm cache though). | 12:35 |
|
| pflanze: well, i guess git tag -l actually amounts to the same thing ^^ never mind. | 12:36 |
| → Yuuhi joined | 12:37 |
|
pigeon
| how can i show the changes of a merge? | 12:37 |
| → kukks joined | 12:37 |
| → aroben joined | 12:37 |
|
Mishakun
| with diff I presume | 12:38 |
| → juanjoc joined | 12:39 |
|
Mishakun
| git diff HEAD^ HEAD | 12:39 |
|
pigeon
| thanks | 12:39 |
|
Mishakun
| or you can just look at the diff patch up on gitk | 12:39 |
|
charon
| pflanze: 'git rev-parse --symbolic-full-name tagname' starts with refs/tags/ if it's a tag, saves you the effort of finding .git (not that it's much harder) | 12:40 |
|
pigeon
| i was using git show, which doesn't show anything useful for a merge commit. | 12:40 |
|
pflanze
| thanks charon | 12:41 |
| ← jomz left | 12:41 |
|
Mishakun
| show only shows contents afaik | 12:41 |
| → EmilMedve joined | 12:41 |
| → jomz joined | 12:43 |
| → sergiodj joined | 12:48 |
| ← grahal left | 12:52 |
| → grahal joined | 12:52 |
| → highandwild joined | 12:54 |
|
highandwild
| hi all | 12:54 |
|
| has anyone got git to work with plink/putty ? | 12:54 |
| ← Sho_ left | 12:54 |
|
highandwild
| I wrote a bash script that calls a pre-saved putty session, and set GIT_SSH=/path/to/script, but | 12:54 |
|
| when the script prompts me for a password, git returns an error and terminates | 12:55 |
|
| I'm using plink/putty on _linux_ (don't ask why - its a long story) | 12:55 |
| → Sho_ joined | 12:57 |
| ← lu_zero left | 12:58 |
|
highandwild
| anyone ? | 12:58 |
| ← highandwild left | 12:59 |
| ← Fullmoon left | 13:00 |
| ← pygi left | 13:01 |
| → pygi joined | 13:04 |
| → unreal_ joined | 13:04 |
| → njal joined | 13:07 |
| ← dp_ left | 13:07 |
| → vbabiy joined | 13:07 |
| → anders_ joined | 13:07 |
| ← unreal left | 13:08 |
| unreal_ → unreal | 13:08 |
| ← lea_w left | 13:08 |
| → cmarcelo joined | 13:15 |
|
Mishakun
| say, does anyone know what the git-daemon-run package in ubuntu do? | 13:18 |
|
| git-daemon is already installed. :S | 13:18 |
| ← BugeyeD left | 13:18 |
|
Mishakun
| installed with git-core I mean | 13:18 |
|
Ilari
| Mishakun: Probably installs some init.d files for git-daemon (or something like that)... | 13:20 |
| → BugeyeD joined | 13:20 |
|
Tv
| Mishakun: umm, why don't you read the package description | 13:21 |
|
| Mishakun: like, third paragraph | 13:21 |
|
| it will actually have the service running | 13:21 |
|
Mishakun
| Yeah, though it doesn't say where it'll put all the files. :3 | 13:22 |
|
Tv
| no package does | 13:22 |
|
Mishakun
| Yeah... | 13:22 |
|
Tv
| dpkg -L etc for that | 13:22 |
|
Mishakun
| anyhow, I wanted to figure out how to run a service of git-daemon to work with gitosis | 13:23 |
|
| but I found the documentation to be "circular" in a way | 13:24 |
| → alegscogs joined | 13:28 |
| → redondos_ joined | 13:28 |
|
Mishakun
| ohkay... installing git-daemon-run package does not work. :[ | 13:32 |
| → blowery joined | 13:33 |
| → queso joined | 13:34 |
| freaky[t] → c1|freaky | 13:38 |
| ← redondos left | 13:40 |
| redondos_ → redondos | 13:40 |
| → Fullmoon joined | 13:41 |
| → jerbear joined | 13:41 |
| ← blowery left | 13:42 |
| → blowery joined | 13:42 |
| ← aroben left | 13:46 |
| → warlock_mza joined | 13:50 |
| → cannonball joined | 13:50 |
| → bryanray joined | 13:54 |
| → corneliuss joined | 13:59 |
| → koke joined | 13:59 |
| → mariusSO joined | 14:00 |
| ← _zack left | 14:03 |
| → _zack joined | 14:04 |
| ← mstormo left | 14:05 |
| → cncfanatics joined | 14:05 |
| mariusSO → mstormo | 14:06 |
| → rsk joined | 14:06 |
|
rsk
| hm could i first fetch 2.6.25 in git, then update to current git pretty easily ? | 14:07 |
| ← tbf left | 14:07 |
|
jengelh
| you just fetch it all at once. | 14:07 |
|
rsk
| but i dont need the history before 2.6.25 | 14:08 |
| → heycarsten joined | 14:08 |
| → bilboed-pi joined | 14:08 |
|
bilboed-pi
| is there any documentation somewhere about the exact formatting of the parameters of git-mktag ? | 14:09 |
|
| there seems to have been plenty of extra checks added | 14:09 |
|
| but ... they don't really make sense | 14:09 |
|
| like the timestamp/timezone/... | 14:09 |
|
jengelh
| mktag is lowlevel... you usually don't use it | 14:09 |
|
bilboed-pi
| when you're converting a cvs repository to git... you do :) | 14:10 |
| cncfanatics → cnc|work | 14:10 |
|
jengelh
| oh? | 14:10 |
|
| git-cvsimport creates tags, so you don't have to | 14:10 |
|
bilboed-pi
| I'm using cvsparse | 14:10 |
|
| sorry, parsecvs | 14:10 |
|
| which produces better results | 14:10 |
|
| except that with the latest git versions... it's broken because of the changes in mktag.c | 14:11 |
|
_slacker_
| Hi guys, I have a few doubts about working with nested repositories, can anyone help me? I pasted a description here: http://pastie.org/227086 | 14:11 |
| ← masm left | 14:11 |
|
Mishakun
| why would you want to nest repositories? | 14:13 |
|
| Mishakun asks puzzled yet curiously intrigued. | 14:13 |
|
rsk
| mkay | 14:14 |
|
| what part of the manual shoudl i look at then? | 14:14 |
|
broonie
| _slacker_: Perhaps you want to look at submodules? | 14:14 |
| → aroben joined | 14:14 |
|
_slacker_
| broonie: submodules? i never heard about that, what they are? | 14:15 |
|
| broonie: any reference? | 14:15 |
| → jbunster joined | 14:16 |
|
broonie
| _slacker_: look at the git-submodule manpage (google will have copies if you've not got it locally) | 14:16 |
| ← bryanray left | 14:16 |
| → grd joined | 14:17 |
|
rsk
| hello :( | 14:18 |
|
_slacker_
| broonie: will read it, but it sounds like a way to control nested repos, i do not want to do that, i haven't commit access to the Dir2 repo, it's just a clone of a public remote repo, so i want to know i can track it with the parent repo Dir1 without problems | 14:19 |
| → lea_w joined | 14:19 |
|
rsk
| i can't spare the bandwith but i want the kernel source | 14:19 |
|
broonie
| _slacker_: If you want to modify it just point the submodule at something you do control? | 14:20 |
| → ckoehler joined | 14:21 |
| ← sbodo_w left | 14:23 |
|
_slacker_
| broonie: sorry, i have control on the Dir1, the parent repo, i just want to know if i can track the modifications in the Dir2 files without problems... or if it's the right way to do it | 14:23 |
|
| currently i'm only commitin on the Dir1 repo (my own) and everything's working fine, but don't know if it's the right way to deal with this | 14:24 |
|
broonie
| _slacker_: Like I say, you can clone Dir2 and then point a submodule of Dir1 at your cloned copy of Dir2. | 14:24 |
|
| If you want to modify the second directory, this is. | 14:25 |
|
grd
| where do I report git bugs? | 14:25 |
|
jengelh
| git ml | 14:25 |
|
grd
| I'm trying to git pull and get "fatal: Not a git repository" thrice, whilst git status from the same directory works | 14:26 |
|
| I suppose that counts as a bug? | 14:26 |
|
hacim
| why can't I find any documentation on how to setup a centralized git repository? | 14:26 |
|
corneliuss
| grd: Did you set up your remote repo correctly? | 14:27 |
|
_slacker_
| broonie: ok, will take a deep look at the submodules, thx! | 14:27 |
|
jast
| hacim, try searching for gitosis | 14:27 |
|
hacim
| jast: yeah, I dont want to use gitosis, because I am on debian etch which does not have it | 14:27 |
|
grd
| corneliuss: I think it's caused by the fact that I'm accessing a subdirectory through a symlink | 14:27 |
|
jast
| hacim, well, you'll have to write your own hook scripts then, i guess | 14:28 |
|
grd
| if I go to the "real" directory, it works as expected | 14:28 |
|
jast
| hacim, unless you don't need push access control, of course :) | 14:28 |
|
hacim
| jast: you mean to say that nobody has done this? | 14:28 |
|
corneliuss
| grd, what do you mean by "go to the real directory"? | 14:29 |
|
jast
| hacim, i guess most people just use gitosis because it's less work | 14:29 |
|
| iirc the gitosis code is debianized | 14:29 |
|
grd
| corneliuss: lets say repo is at test/repo, and there is a folder sub in the repo. | 14:29 |
|
hacim
| jast: yeah it is, but there is not a backport to etch | 14:29 |
|
| i must be using the wrong search terms | 14:30 |
|
jast
| so i built the package myself. took two minutes. | 14:30 |
| ← pergesu left | 14:30 |
|
grd
| I have a symlink from test/sub to test/repo/sub | 14:30 |
|
jast
| grd, well, if you're in test/sub, you're not inside the repo. i don't see that as a bug then. | 14:30 |
|
hacim
| jast: thanks, I'll do that | 14:31 |
|
grd
| jast: I disagree, and git status seems to agree with me | 14:31 |
|
jast
| anyway, ask on the mailing list | 14:31 |
|
corneliuss
| grd, why do you use this symlink in the first place? | 14:32 |
|
grd
| corneliuss: to save some typing | 14:32 |
|
corneliuss
| you only have to type it once, when you set up the remote repo, or am I mistaken? | 14:34 |
|
grd
| I'm referring to my local copy | 14:34 |
|
corneliuss
| grd, okay. Maybe I can't help you then, because I'm new to git, too. | 14:38 |
| → johan-s joined | 14:41 |
| → foca joined | 14:42 |
| → dkagedal joined | 14:47 |
| ← SamB left | 14:47 |
| comp → Guest18401 | 14:49 |
| → lu_zero joined | 14:50 |
| → comp joined | 14:51 |
| → kumbayo joined | 14:55 |
| → freenotfree joined | 14:56 |
|
freenotfree
| hey - i started editing a file in the wrong branch, and now when i try to switch branches im told the "Entry ... is not uptodate. Cannot merge." -- what can i do to resolve this? | 14:56 |
|
doener_
| try using "checkout -m", that should try to merge the file eventually leaving conflict markers | 14:57 |
| ← anders_ left | 14:57 |
| → crab joined | 14:57 |
| ← jelmer left | 14:58 |
|
freenotfree
| doener_: will that merge the new edits i dont want onto the file that i did want to edit? | 14:58 |
| → rdavila joined | 14:59 |
| ← nud left | 14:59 |
| ← mbuf left | 15:00 |
| → nud joined | 15:00 |
|
doener_
| freenotfree: That will try to merge your uncommitted changes into the files from the other branch and leave the results in your working tree | 15:01 |
|
freenotfree
| doener_: how can i roll back those changes so they aren't merged at all? | 15:02 |
|
doener_
| freenotfree: "git checkout -- file" gets the file as it is in your index into your working tree, and "git checkout HEAD -- file" will get the version from HEAD into your index and working tree | 15:03 |
|
| freenotfree: "file" can also be multiple files, like "git checkout HEAD -- fileA fileB" | 15:03 |
|
| freenotfree: if you just want to drop all uncommitted changes, "git reset --hard" it is | 15:04 |
|
freenotfree
| pretty sure ``reset --hard`` is what i want. do i need to commit or push anything after that? | 15:04 |
|
doener_
| no | 15:04 |
|
corneliuss
| http://www.kernel.org/pub/software/scm/git/docs/git-reset.html | 15:05 |
| ← lea_w left | 15:05 |
|
doener_
| you can also use "git stash" to put changes aside | 15:05 |
| ← cnc|work left | 15:05 |
|
freenotfree
| cool - didnt think so, but wanted to check. | 15:05 |
|
| corneliuss: thanks for the ref. | 15:05 |
| → jelmer joined | 15:06 |
|
corneliuss
| np | 15:06 |
| ← fisons left | 15:08 |
| ← charon left | 15:08 |
| ← queso left | 15:11 |
|
nutrimatt
| Hi folks. I and a fellow developer are pushing to a remote branch on GitHub and we're getting commits we don't expect. We get the commit we made locally plus a "Merge branch 'master'" commit that seems to repeat whatever commit the last developer made on that branch. | 15:12 |
|
| Can anyone either explain why? Or tell us what we're doing wrong? | 15:12 |
| ← rphillips left | 15:14 |
| → charon joined | 15:14 |
|
Ilari
| nutrimatt: Those commits are created by 'git pull' or 'git merge'... To avoid that sort of stuff, you need to use rebase (which OTOH can cause mess if used incorrectly). | 15:15 |
| → rphillips joined | 15:16 |
|
nutrimatt
| Ilari: thanks | 15:16 |
|
Ilari
| nutrimatt: Basically, each time 'git pull' or 'git merge' is called and both local and upstream sides have new commits, it will create new merge commit that combines the two... | 15:18 |
|
nutrimatt
| Ilari: that makes sense... someone in another chan suggested working from another branch, updating master, rebasing the branch from master, then doing the merge and push | 15:20 |
| → ajsharp joined | 15:21 |
| → brosner joined | 15:21 |
|
ajsharp
| I just started noticing very odd behavior with git and a new rails project | 15:22 |
| → bartocc_ joined | 15:23 |
|
ajsharp
| If when I add new files, and I just want to do git add . instead of doing git add '...' for every file, it tries to delete some of my plugins from my local repo | 15:23 |
| ← reel_ left | 15:23 |
| ← bartocc left | 15:23 |
|
ajsharp
| am i missing something here or doing something wrong? | 15:23 |
|
jengelh
| orly? | 15:23 |
| ← charon left | 15:23 |
| ← dsop left | 15:24 |
| → les joined | 15:24 |
| → dsop joined | 15:25 |
| → charon joined | 15:25 |
| → d0k_ joined | 15:26 |
|
ajsharp
| For example, I just added the director ./vendor/rails to my project | 15:27 |
|
| and if i do "git add ." it tries to delete "vendor/plugins/validates_as_email_address" | 15:27 |
|
| but not any other of my plugins | 15:27 |
|
jengelh
| strange though | 15:27 |
|
ajsharp
| ever had anything like this happen? | 15:28 |
|
jengelh
| nope | 15:28 |
|
ajsharp
| so strange | 15:28 |
| → lea_w joined | 15:28 |
|
jengelh
| quick, mail the mailing list before anyone else does! | 15:28 |
| ← redondos left | 15:28 |
|
ajsharp
| why? | 15:28 |
|
jengelh
| because it ought to be fixed? :) | 15:29 |
|
ajsharp
| well im not sure its a git problem | 15:29 |
|
biz
| actually delete or just ignore at the next commit? | 15:29 |
|
ajsharp
| maybe its a problem with this guys plugin or something | 15:29 |
|
jengelh
| strace -e git add | 15:29 |
|
| strace -fe unlink git add | 15:29 |
| → kraymer joined | 15:29 |
| → digitalhobbit joined | 15:29 |
| → scroogemcduck joined | 15:30 |
|
scroogemcduck
| hello | 15:30 |
|
ajsharp
| i dont have strace, what is that? | 15:30 |
|
thiago
| a tool that shows system calls | 15:30 |
| → krawek joined | 15:30 |
|
thiago
| very useful | 15:31 |
|
ajsharp
| k hold up, lets see if i can get er to compile on os x | 15:31 |
|
thiago
| oh, OSX doesn't have it | 15:31 |
|
EgS
| use dtruss on mac os x (10.5) | 15:31 |
|
jengelh
| oh i feel sorry for you :p | 15:31 |
|
nutrimatt
| EgS: ahhh that's where truss went to ;-) | 15:32 |
|
ajsharp
| haha, whys that? | 15:32 |
|
| linux man r ya? | 15:32 |
|
scroogemcduck
| I am using gitosis and I have currently have two branches (master and stable-0.1) | 15:32 |
|
| if i want to do the initial push | 15:32 |
|
| should i do it twice? | 15:32 |
|
| git push origin master:refs/heads/master | 15:32 |
|
| and | 15:32 |
| → bdiego joined | 15:32 |
|
Mikachu
| scroogemcduck: you know too much and too little, "git push origin master" is enough | 15:32 |
|
scroogemcduck
| git push origin stable-0.1/heads/stable-0.1 | 15:32 |
| ← d0k left | 15:33 |
|
scroogemcduck
| * [new branch] master -> master | 15:34 |
| ← EmilMedve left | 15:34 |
|
scroogemcduck
| now it didnt create the stable-0.1 | 15:34 |
|
Mikachu
| you can specify multiple branches on one push line | 15:34 |
| → EmilMedve joined | 15:34 |
|
jengelh
| scroogemcduck: because you did not push that | 15:34 |
|
Mikachu
| you can also use --all if you want to push all local branches | 15:34 |
| ← W|cked left | 15:34 |
|
Mikachu
| also note i gave you the command 3 seconds before you even mentioned the stable branch :) | 15:34 |
|
scroogemcduck
| aha oki | 15:35 |
|
| and now i just need to add those track ? | 15:35 |
|
jengelh
| git push origin stable-0-1 | 15:36 |
|
scroogemcduck
| i did that | 15:36 |
|
jengelh
| then it's up | 15:36 |
|
scroogemcduck
| but wouldn't i need to track remote branch? | 15:36 |
|
Mikachu
| why would you track your own branches? :) | 15:36 |
|
jengelh
| wut? | 15:36 |
|
| well sometimes that makes sense | 15:36 |
|
| set it up and be happy | 15:36 |
|
scroogemcduck
| :) | 15:37 |
|
| oh | 15:37 |
|
| oki | 15:37 |
| → koot joined | 15:37 |
|
jengelh
| Mikachu: it's not unreasonable | 15:37 |
|
| Mikachu: (local) git init; (work work work); git push centralserver; (then from another machine clone&work&push), then you want to do a pull on 'local' too | 15:38 |
|
scroogemcduck
| yes | 15:38 |
|
| i want to switch computers :) | 15:38 |
|
| notebook <> desktop computer | 15:38 |
|
jengelh
| run git pull | 15:38 |
|
| follow the instructions | 15:38 |
|
ajsharp
| is the mailing list [email@hidden.address] | 15:38 |
|
koot
| is there a way to get diff-tree (with -r) to indicate renames of trees and not every file within the renamed tree? | 15:39 |
| ← les left | 15:39 |
|
Mikachu
| ajsharp: no, [email@hidden.address] :) | 15:40 |
| → gberdyshev joined | 15:41 |
| → queso joined | 15:42 |
| → Ramblurr joined | 15:42 |
| gberdyshev → gberdyshev_ | 15:42 |
|
Ramblurr
| is it possible to git reset a remote repo? :\ | 15:42 |
|
koot
| Ramblurr: I suppose you could reset locally and force push the ref | 15:43 |
|
Tv
| or just force push the ref, just like that | 15:43 |
| ← charon left | 15:44 |
|
Ramblurr
| thanks :) | 15:44 |
| → anders_ joined | 15:44 |
|
scroogemcduck
| working guys | 15:45 |
|
| i think, when i do a new clone it wont get those local branches of course :) | 15:46 |
| ← vbabiy left | 15:48 |
| ← seanstickle left | 15:49 |
| ← nicholas_ left | 15:50 |
| → nicholas_ joined | 15:51 |
| → foca_ joined | 15:51 |
| ← bilboed-pi left | 15:52 |
| → jiva joined | 15:52 |
| ← sverrej left | 15:53 |
|
jiva
| Hey, I'm using git-svn... and I have committed a local change that I actually didn't mean to... so I want to revert my change without ever pushing it up to the svn repo... does anyone know how to do this? | 15:53 |
|
crab
| git-reset --hard HEAD^ | 15:53 |
|
| that'll make your HEAD point to whatever commit was its parent, i.e .the one before you committed anything. | 15:54 |
|
Mikachu
| jiva: it will also delete the change from your working tree | 15:54 |
|
jiva
| that's ok, I want it deleted from the local tree | 15:54 |
|
ckoehler
| what would the command be for just redoing the commit but leaving all the changes intact? | 15:54 |
|
| i.e. keeping the changes as uncommitted | 15:54 |
|
| reset soft? | 15:54 |
|
Mikachu
| ckoehler: redoing how? not commit --amend? | 15:55 |
|
ckoehler
| i mean undoing the commit, not redoing | 15:55 |
| → danlarkin joined | 15:55 |
|
jiva
| ahhh you guys are my heroes... thanks... this is doin the job. | 15:55 |
| ← kanru left | 15:55 |
|
Mikachu
| ckoehler: then yes, --soft | 15:55 |
|
ckoehler
| coo | 15:55 |
|
| l | 15:55 |
|
| thanks | 15:55 |
| → vbabiy joined | 15:56 |
| → patientfox joined | 15:57 |
|
jengelh
| --soft, --hard, and --medium :-) | 15:58 |
|
| er, mixed. | 15:58 |
| ← ajsharp left | 15:58 |
| → charon joined | 15:58 |
|
patientfox
| hey.. so i'm trying to do a rebase of branch a to branch b from a commond divergence of commit c.. the classic case, i guess... and i'm having problems with branch a not being able to patch the binaries for branch b.. which is fine and common | 15:59 |
|
| but how do i resolve the conflict in this case? | 15:59 |
|
| i need git to let me choose which branches binaries i want to pick and resolve the conflict.. | 15:59 |
|
| branch's binaries, even | 15:59 |
|
Mikachu
| git checkout a_or_b -- thebinaries | 16:00 |
| ← Sho_ left | 16:00 |
|
danlarkin
| Hi there git experts! On my master branch I have a bunch of "changed but not updated" and "untracked" files. I want to basically move these files into a local branch. So then my master will be a clean checkout and the new branch will be master plus a commit of all these new files. Can I do this? | 16:01 |
|
crab
| danlarkin: git-stash; git-checkout -b newbranch; git-stash apply --index; git-checkout master | 16:01 |
|
| er, well, not git-checkout master at the end. | 16:02 |
|
cebewee
| git-stash is not necessary here | 16:02 |
|
lea_w
| I have a question about gitweb -- am I understanding the source code correctly that if the forks feature is enabled, it will not recursively search for projects, but only pick up projects in the project root directory? IOW, it won't pick up foo/bar.git even if foo is not a git project? | 16:03 |
|
cebewee
| "git checkout -b newbranch; [do commit]" suffices | 16:03 |
|
crab
| oh, good point. i'd forgotten about that. | 16:03 |
|
danlarkin
| cebewee: creating a new branch won't lose my uncommitted/untracked changes? | 16:04 |
|
cebewee
| danlarkin: right | 16:04 |
| ← markkalderon left | 16:04 |
|
Mikachu
| git usually won't remove changes unless you say something like -f or --hadr | 16:04 |
|
| ^t | 16:04 |
|
crab
| lea_w: basically yes. | 16:05 |
| ← priidu left | 16:05 |
|
danlarkin
| cebewee crab: thanks guys :D | 16:05 |
|
crab
| lew: but it's not that it won't see them, it's that it'll place them under their top-level parent and not display them right away, afaics. | 16:06 |
|
lea_w
| crab: Thanks! | 16:06 |
|
patientfox
| Mikachu: do i need to just checkout the binaries? or use git add binaries ... as well ? | 16:06 |
| → Ryback_ joined | 16:07 |
| → stillLotR joined | 16:07 |
|
crab
| lea: oh, there's a comment around line 260 which explains it. | 16:07 |
| → priidu joined | 16:07 |
| ← foca left | 16:08 |
|
Mikachu
| patientfox: checkout adds them to the index, so you don't need add | 16:10 |
|
patientfox
| Mikachu: thanks | 16:10 |
| ← LotR left | 16:11 |
| ← vbabiy left | 16:12 |
| ← charon left | 16:14 |
| → Sho_ joined | 16:15 |
| ← kez left | 16:18 |
| ← nud left | 16:20 |
| → kanru joined | 16:20 |
| → k0001 joined | 16:21 |
| → sverrej joined | 16:21 |
| → foca joined | 16:23 |
| ← rdavila left | 16:24 |
| blowery → blowery`afk | 16:28 |
| ← grd left | 16:30 |
| ← foca_ left | 16:33 |
| → naeu joined | 16:34 |
| ← gberdyshev_ left | 16:38 |
| ← EmilMedve left | 16:39 |
| → quassel251 joined | 16:42 |
| ← jomz left | 16:43 |
| → timocratic joined | 16:44 |
| → gberdyshev_ joined | 16:44 |
| ← quassel251 left | 16:44 |
| → jomz joined | 16:44 |
| → freqmod_q joined | 16:44 |
| ← freqmod_q left | 16:45 |
| → freqmod_qu joined | 16:45 |
| → hkBst joined | 16:46 |
|
hkBst
| I removed some file with local mods using "git-rm -f *", but now every operation I tried gives "fatal: Unable to read current working directory". | 16:47 |
|
| s/file/files/ | 16:47 |
| ← patientfox left | 16:47 |
| ← comp left | 16:47 |
|
lea_w
| hkBst: that's got nothing to do with git, do cd .. until you are in an existing directory | 16:48 |
|
hkBst
| oh I see, it removed the current directory too :) | 16:48 |
| → comp joined | 16:50 |
| → tcoppi_ joined | 16:51 |
| ← Fullmoon left | 16:51 |
| ← mcella left | 16:51 |
| ← tcoppi left | 16:52 |
| awx_ → awx | 16:54 |
| ← priidu left | 16:55 |
| → anlek joined | 16:55 |
|
anlek
| Hello everyone | 16:55 |
|
| Is there a way to prevent write access to a branch in a project using gitosis? | 16:55 |
| → CodeOfficer joined | 16:55 |
| ← tcoppi_ left | 16:55 |
| ← anlek left | 16:55 |
|
Mikachu
| patient fellow | 16:56 |
|
cardioid
| heh | 16:56 |
| → anlek joined | 16:56 |
|
anlek
| sorry about that | 16:56 |
|
Tv
| anlek: hooks | 16:56 |
|
cardioid
| Oh, if anyone knows it'd be Tv | 16:56 |
|
anlek
| hooks? anywhere I can read about it online? | 16:56 |
| blowery`afk → blowery | 16:57 |
|
Tv
| anlek: git docs | 16:57 |
|
anlek
| Ok, thanks | 16:57 |
|
Mikachu
| man githooks | 16:57 |
|
cardioid
| anlek: What's the use case there? You want to let someone write to the repo in general but not a specific branch? | 16:57 |
| → priidu joined | 16:57 |
| → charon joined | 16:57 |
| → markkalderon joined | 16:57 |
|
anlek
| cardioid, I want a development branch that allows developers to write and read, but the master branch should only have read access except for my account, I should be able to read and write. | 16:58 |
| → tcoppi joined | 16:58 |
|
Tv
| anlek: to get off easy, either 1) use a separate repo for that 2) use a baseball bat | 16:58 |
| → yofel joined | 16:59 |
|
cardioid
| Failing that, a hook definitely seems the right way to go | 16:59 |
|
crab
| lea_w: what are you working on? | 16:59 |
|
anlek
| neither of those are really easier then learning how to use hooks for me... | 16:59 |
|
cardioid
| Assuming the reason you couldn't do 2 was because all you had was a golf club, in which case you should realize that "baseball bat" is just a general term for a blunt object | 16:59 |
|
lea_w
| crab: caching for gitweb (a GSoC project) | 16:59 |
|
Tv
| anlek: i prefer the baseball method, for people in the same office | 16:59 |
|
lea_w
| crab: check it out here: http://odin3.kernel.org/git-lewiemann/ | 17:00 |
|
anlek
| hehe, different country for some... it gets quite pricey | 17:00 |
|
cardioid
| Tv: Sadly the combination of git's distributed nature and the general trend of distributed teams makes that unlikely | 17:00 |
|
lea_w
| the project list is still very slow, and the owner isn't displayed correctly, but the rest should work. let me know if there are any problems, please :) | 17:00 |
|
cardioid
| anlek: http://www.kernel.org/pub/software/scm/git/docs/v1.5.5.4/hooks.html | 17:01 |
|
Tv
| cardioid: think of it more like the people living in constant fear of vacation travel, from that point on | 17:01 |
|
anlek
| cardioid, thanks! | 17:01 |
|
cardioid
| Tv: Your ideas intrigue me and I would like to subscribe to your newsletter | 17:01 |
|
| I had that link handy from when I was trying to figure out the difference between post-receive and post-update | 17:01 |
| → EmilMedve joined | 17:01 |
| → cncfanatics joined | 17:02 |
|
crab
| lea_w: is there something i can diff? | 17:03 |
|
lea_w
| crab: you mean my code? *pushes* | 17:04 |
|
crab
| yes. | 17:04 |
|
lea_w
| right here: http://repo.or.cz/w/git/gitweb-caching.git | 17:04 |
| ← jbunster left | 17:06 |
| → bobesponja joined | 17:06 |
|
crab
| damn, i didn't notice the 'w' and tried to clone that. | 17:07 |
| ← _zack left | 17:07 |
|
crab
| i must say git-clone tried very hard. | 17:08 |
| ← bgerber left | 17:08 |
| → peritus_ joined | 17:09 |
| → ketralnis joined | 17:09 |
| ← crab left | 17:11 |
| → bronson_ joined | 17:12 |
| → rubydiamond joined | 17:12 |
| ← lu_zero left | 17:13 |
|
scroogemcduck
| anyway to get statistics :) | 17:13 |
|
| like how much people cloned it? | 17:13 |
| ← radarek left | 17:18 |
| → peritus__ joined | 17:19 |
| ← comp left | 17:21 |
| → nud joined | 17:22 |
| ← warlock_mza left | 17:22 |
| ← kraymer left | 17:23 |
| metze → metze_away | 17:23 |
| ← grahal left | 17:23 |
| → grahal joined | 17:24 |
| ← aroben left | 17:25 |
| → anttih joined | 17:26 |
| ← d0k_ left | 17:26 |
|
anttih
| hmm, I'm getting a "Already uptodate" from git merge <branch>. it's not true. | 17:26 |
|
cardioid
| Lies! | 17:26 |
|
anttih
| oops | 17:27 |
|
| sorry :-D | 17:27 |
| ← rphillips left | 17:27 |
|
anttih
| my pull already merged | 17:27 |
|
| I was in a wrong branch | 17:27 |
| ← peritus_ left | 17:29 |
| DrNick2 → DrNick | 17:29 |
| → Jearil joined | 17:30 |
| ← Jearil left | 17:30 |
| → pdherbemont joined | 17:31 |
| ← grahal left | 17:33 |
|
foca
| is there a bash tab-completion script to complete git-add only with modified files? | 17:33 |
|
| or should I try to roll my own? | 17:33 |
| → Zilvinas joined | 17:33 |
| → Jearil joined | 17:33 |
|
Zilvinas
| New Irc Server: Just Write /server -m FreeBsd.serveblog.net | 17:33 |
| → grahal joined | 17:33 |
| → bfields joined | 17:36 |
| Ramblurr → rmblr | 17:36 |
| rmblr → Ramblurr | 17:36 |
| ← corneliuss left | 17:38 |
|
Zilvinas
| New Irc Server: Just Write /server -m FreeBsd.serveblog.net | 17:39 |
|
koot
| Important: FILE1 and FILE2 must be sorted on the join fields. | 17:39 |
|
| oops, sorry | 17:39 |
| → quabbles joined | 17:39 |
| → dejavu joined | 17:41 |
|
scroogemcduck
| anyone here good with regexes :D ? | 17:42 |
| ← dejavu left | 17:42 |
|
vmiklos
| don't ask to ask.. | 17:42 |
|
scroogemcduck
| I have a string like: "http://www.nzherald.co.nz/index.cfm?c_id=4&objectid=10519760&ref=rss" | 17:42 |
|
vmiklos
| Zilvinas: could you please stop spamming? | 17:42 |
|
scroogemcduck
| now I want to use a regex to get the value of objectid | 17:43 |
| → kraymer joined | 17:43 |
| ← pombreda left | 17:43 |
| ← markkalderon left | 17:43 |
|
Jearil
| /objectid=(\d*)/ | 17:44 |
| → markkalderon joined | 17:44 |
|
Ilari
| scroogemcduck: Does the string really have those & things? | 17:44 |
| → sbodo_w joined | 17:44 |
| → selckin joined | 17:44 |
|
selckin
| if i delete a file, can i make git restore it? been trying with git reset, but can't seem to get anywhere | 17:45 |
| ← EmilMedve left | 17:46 |
|
lorandi
| selckin, try checkout the removed file. | 17:47 |
|
ciaran29d
| scroogemcduck: in what language? | 17:48 |
|
Jearil
| if you didn't commit the deletion you should also be able to: git reset --hard | 17:48 |
|
Ilari
| Maybe something like "^[^?]*?(.*&|)objectid=([0-9]*)(|&.*)$" would work (take the second capture)... | 17:49 |
|
selckin
| Jearil: that only seems to work for everything, not individual files | 17:49 |
|
Ilari
| selckin: 'git checkout -- file'? | 17:50 |
|
cardioid
| That regex seems overly complicated | 17:50 |
|
selckin
| lemme try | 17:50 |
| → robinr joined | 17:50 |
|
Jearil
| selckin, yes, that's true. It should reset you to your last commit. | 17:50 |
|
selckin
| that works thanks | 17:50 |
| → eikonos joined | 17:52 |
|
Ilari
| Or actually, if free-anchored matches are allowed, "(?|&)objectid=([0-9]*)" (second capture of that too) could also work... | 17:52 |
|
stillLotR
| scroogemcduck: using what language? In Perl, I'd say use URI and URI::Escape | 17:52 |
| stillLotR → LotR | 17:53 |
| ← anlek left | 17:53 |
|
cardioid
| Are git commit messages stripped of blank lines at either end? (non-internal blank lines) | 17:55 |
| ← dkagedal left | 17:55 |
| ← ebel left | 17:57 |
|
lea_w
| cardioid: Yup. | 17:58 |
|
Ilari
| cardioid: Both blank lines at start and end are stripped... | 17:58 |
| ← Zilvinas left | 17:58 |
| → com4 joined | 17:58 |
| ← robinr left | 17:59 |
|
com4
| Someone sent me some patches made with git-format-patch. I try to apply them using git am 0001....patch but it's complaining about .dotest. How do I apply these patches? | 17:59 |
| → robinr joined | 17:59 |
| ← cncfanatics left | 17:59 |
| ← kraymer left | 17:59 |
|
scroogemcduck
| thanks guys | 17:59 |
|
| got it working _again_ :) | 17:59 |
|
| how nice | 17:59 |
|
ciaran29d
| answer the question so you can do it the right way instead, and it won’t break | 18:00 |
| ← foca left | 18:01 |
|
cardioid
| lea_w / Ilari: thanks for the confirmation | 18:01 |
| → foca joined | 18:01 |
|
cardioid
| I always wondered about that extra line when I did a git commit --amend | 18:01 |
|
| then when I did my first squash and all those blank lines appeared, I started thinking | 18:01 |
|
Ilari
| com4: If there isn't supposed to be rebase or patch apply in progress, delete .dotest and retry the am command? | 18:02 |
| ← naeu left | 18:03 |
| ← peritus__ left | 18:03 |
| ← warthog9 left | 18:03 |
| → cncfanatics joined | 18:03 |
|
com4
| Ilari: k | 18:05 |
|
scroogemcduck
| ciaran29d: responding too me? | 18:06 |
| → ijcd joined | 18:06 |
| ← foca left | 18:07 |
| → robin joined | 18:09 |
| → JaapSuter joined | 18:11 |
|
JaapSuter
| Hello. | 18:11 |
|
Jearil
| hi | 18:11 |
|
JaapSuter
| Given two commit hashes A and B, with A a (possibly indirect) parent of B, how can I get a list of all hashes of all objects that got introduced between A and B? | 18:12 |
|
cardioid
| git log A..B, I think | 18:12 |
| → pombreda joined | 18:13 |
|
JaapSuter
| I looked at build-pack and the pack-objects pipeline, but it only specifies how the objects are bundled and transfered, I don't see how is decided what is put in the pack. | 18:13 |
|
jast
| for just the hashes, try git rev-list A..B | 18:13 |
|
Ilari
| JaapSuter: Also see --objects-edge to git-rev-list(1). | 18:14 |
|
cardioid
| Yeah, I was actually more unsure about the A..B bit, jast ;) | 18:14 |
| ← gberdyshev_ left | 18:15 |
| ← gitte left | 18:15 |
|
JaapSuter
| Hari and jast, that seems to be what I was looking for, thanks! | 18:16 |
|
jast
| yet another exciting misspelling :} | 18:16 |
| ← cncfanatics left | 18:19 |
|
com4
| Ilari: worked, thanks | 18:20 |
| → kraymer joined | 18:20 |
| → _zack joined | 18:22 |
| ← sbodo_w left | 18:23 |
| ← Mutru left | 18:23 |
| ← robinr left | 18:25 |
|
JaapSuter
| I end up in list-objects.c, but I think I'll need to spent some time to fully understand what's going on in here. | 18:26 |
| → warthog9 joined | 18:26 |
|
JaapSuter
| Can anybody say anything about the big-O complexity of figuring out what needs to go into a pack? | 18:26 |
| → EmilMedve joined | 18:26 |
| → cncfanatics joined | 18:26 |
| ← cncfanatics left | 18:27 |
| → jbunster joined | 18:27 |
| ← robin left | 18:28 |
| → eternaleye joined | 18:29 |
| → markkald_ joined | 18:31 |
| ← _slacker_ left | 18:31 |
| ← markkalderon left | 18:31 |
| ← rubydiamond left | 18:35 |
| → athought joined | 18:35 |
| ← branstrom_ left | 18:35 |
| ← koke left | 18:37 |
| → Gitzilla joined | 18:37 |
| ← jomz left | 18:38 |
| → jackdempsey joined | 18:41 |
| → thannoy joined | 18:43 |
| → docwhat joined | 18:46 |
|
JaapSuter
| Also, is rev-list conservative or exact? If I generate a list of all objects necessary if I have commit A, but not commit B (i.e., git rev-list --objects A ^B) it seems to display more hashes than necessary. | 18:46 |
|
Ilari
| JaapSuter: Maybe use --objects-edge instead of --objects? | 18:47 |
|
| JaapSuter: And yes, it is conservative. Being strict would mean walking the entiere history. | 18:47 |
|
JaapSuter
| That's what I figured. So that means packs send between two computers may occasionally send more than is strictly necessary. | 18:48 |
| → BabelO joined | 18:49 |
|
JaapSuter
| Moreover, am I correct in assuming that I can theoretically really mess with Git (i.e., generate horrendous rev-list and send-pack performance) if I do some major directory renaming all over a tree? | 18:49 |
| → gberdyshev_ joined | 18:49 |
|
Ilari
| JaapSuter: Worst case would be that the only commit being transferred is revert of last commit before that. In that case, it trasnfers the revert, but only commit object would be required. | 18:50 |
|
| JaapSuter: I don't think that would make bad case. I think reverts cause the worst cases... | 18:50 |
| → rphillips joined | 18:52 |
|
Ilari
| JaapSuter: And renaming a directory (without touching its contents) only generates tree objects for directories on path to root (which is not very bad)... | 18:52 |
|
JaapSuter
| Imagine a tree with one directory A with a thousand files in it. I commit. Then I rename the directory to B, and commit again. I realize this commit only generates a new object for the tree itself and the actual commit, but not for the thousand files in it. | 18:53 |
|
| Now I call rev-list, and Git needs to figure out what is different between A and B. | 18:53 |
|
| How can it know that the thousand entries in A are the same as the thousand entries in B? | 18:53 |
| ← _zack left | 18:54 |
|
JaapSuter
| In this case, it could peek in those directories. But now imagine a fairly complex directory structure (Linux kernel) and I move a large directory from somewhere in the tree to somewhere completely different. | 18:54 |
|
Ilari
| JaapSuter: The object ID for directory A in first commit and B in second commit is the same. | 18:54 |
|
JaapSuter
| I can't imagine Git would want to search an entire tree to see if it already had those objects. | 18:54 |
|
| Ah yes, the name of the directory is stored in an entry in the parent directory (the root, in this case). | 18:56 |
|
Ilari
| JaapSuter: All that tree objects contain is mode, name and SHA1 for each entry in tree. There are no headers or trailers... | 18:58 |
| → chieh joined | 18:58 |
| ← gberdyshev_ left | 18:58 |
|
chieh
| hello, when I do a clone, is there a way i can set a max transfer speed so I don't suck up all the bandwidth? | 18:59 |
| → DrNick2 joined | 19:00 |
| ← DrNick left | 19:00 |
| → gberdyshev_ joined | 19:02 |
|
JaapSuter
| Ilari: http://pastebin.com/m3d6d0778 | 19:03 |
| → insin joined | 19:03 |
|
Tv
| chieh: only by generic mechanisms, e.g. iptables/qos on linux | 19:03 |
|
JaapSuter
| If you have time, that is. Feel free to ignore. Meanwhile, I'll go play with Git and see what happens. | 19:03 |
| ← gberdyshev_ left | 19:04 |
| ← chieh left | 19:06 |
| → \ask_ joined | 19:07 |
|
hacim
| if I start out by cloning a remote repository, then a 'push' will push back to where I cloned from. I want to clone that repository and then set up a different place to push to by default, but still be able to push to the place where I cloned from occasionally... what is the best way to do that? | 19:07 |
| ← kumbayo left | 19:08 |
| ← freenotfree left | 19:09 |
| → guillermo_spain joined | 19:11 |
|
guillermo_spain
| hi | 19:11 |
|
| is there any hook to relaunch a fcgi process? | 19:11 |
|
Jearil
| we did a merge between a CVS branch and CVS head. For some reason now git-cvsimport doesn't get the most up to date version of head (missing files that were added). | 19:11 |
|
guillermo_spain
| (i mean prebuilt, i know you can code it yourself, it could be great save the time :) | 19:12 |
|
hacim
| if I clone something, it is set as origin, whats the best way to change that? | 19:13 |
|
bfields
| hacim: look at .git/config. | 19:13 |
|
JaapSuter
| Mmmmm, git appears to do the right for http://pastebin.com/m3d6d0778, I have no idea how it does that. Looks like I'll have to step through the code. Thanks. | 19:13 |
| ← bdiego left | 19:14 |
| → Eridius|work joined | 19:14 |
|
drizzd_
| hacim: git-push pushes to origin by default | 19:14 |
|
jengelh
| orly | 19:14 |
|
hacim
| drizzd_: yeah, thats why I want to change my origin | 19:14 |
|
jengelh
| hm indeed | 19:14 |
| ← athought left | 19:14 |
|
drizzd_
| hacim: i don't think you can change the default remote name created by git clone | 19:15 |
|
| hacim: so what's the problem? | 19:15 |
|
jast
| well, you can always edit the origin section in .git/config and create a renamed copy of the original | 19:15 |
|
bfields
| hacim: also "git remote" will help manage remote repositories. But mainly just by modifying [remote ".."] stanzas in .git/config, which you can do by hand if you prefer. | 19:15 |
|
hacim
| drizzd_: you can when you run the clone by passing --origin | 19:15 |
|
| drizzd_: I already cloned something and I want to make something else the origin | 19:15 |
| ← kraymer left | 19:16 |
|
jast
| edit the config file. problem solved. | 19:16 |
|
drizzd_
| well that will leave stale tracking branches I believe | 19:16 |
|
jast
| it's just a bunch of refs. you can always remove them manually if you care. | 19:17 |
|
hacim
| how about: git remote add neworigin someurl/here; git config branch.master.remote neworigin; git config branch.master.merge refs/heads/master; git fetch ; git merge master ? | 19:17 |
|
jast
| or use git remote prune | 19:17 |
| → aconbere joined | 19:17 |
| → spearce joined | 19:17 |
|
drizzd_
| well the clean way to do it is using git-remote. | 19:18 |
|
aconbere
| if I want to grab another branch from a remote repository besides master, how can I do that (and can I do it without using clone) | 19:18 |
|
jast
| what's unclean about editing the config file? | 19:18 |
|
hacim
| drizzd_: ok, how should I do that? | 19:18 |
|
jast
| aconbere, typically, all local branches that exist in the remote repo are automatically fetched into remote tracking branches, so you only need to do git checkout -b localname refs/remotes/theremotename/thebranch | 19:19 |
|
drizzd_
| jast: what happens if a remote branch is checked out, then you remove the remote tracking branch and the remote? | 19:20 |
|
jast
| you wouldn't want to checkout a tracking branch, usually, since that detaches HEAD | 19:21 |
|
| but the form i gave automatically creates a local branch from it that follows the tracking branch | 19:21 |
|
| (if you didn't set autosetupmerge to false) | 19:22 |
|
aconbere
| jast: hmm... so those would be different from what appear in git branch? | 19:22 |
|
jast
| what do you mean? | 19:22 |
| → ___tortoise joined | 19:23 |
|
___tortoise
| hey folks, how goes it? | 19:23 |
|
jast
| slowly. if you're a tortoise. ;) | 19:23 |
|
___tortoise
| it's true | 19:23 |
|
jast
| (hi.) | 19:23 |
|
___tortoise
| i was wondering, is there a way to change the default editor GIT opens for commit comments? | 19:23 |
|
| i'm on debian and it's opening VIM, i'd love for it to open in gedit instead | 19:24 |
|
jast
| there are even several ways to do that | 19:24 |
|
___tortoise
| nice | 19:24 |
|
aconbere
| jast: so... I've done a git-clone of a remote repository. I know they have a branch "x" that I need to be working on. it's not in the list of branches in my local copy shown by git branch | 19:24 |
|
Ilari
| aconbere: 'git checkout -b x origin/x'? | 19:24 |
|
jast
| ___tortoise, e.g. you can set core.editor in ~/.gitconfig, or set the environment variable VISUAL. | 19:24 |
| blowery → blowery`afk | 19:25 |
|
jast
| ___tortoise, be careful with GUI editors though, some of them run in the background. that confuses git. | 19:25 |
|
___tortoise
| jast: that's a good point | 19:25 |
|
| jast: i suppose even just changing it to emacs would be an improvement | 19:25 |
|
jast
| gvim, for example, has a switch to keep it from forking into the background | 19:26 |
|
| i don't know gedit well enough, check out the manpage perhaps | 19:26 |
|
| or just try if it forks into the background at all | 19:26 |
|
___tortoise
| makes sense | 19:26 |
|
aconbere
| Ilari: is there a way to list the branches on the remote repository? | 19:26 |
|
lea_w
| pasky: The project list code is heavily optimizable if you can assume that there are no symlinks in the project directory hierarchy (symlinked HEAD files etc. are fine, just the directories leading up to them may not be symlinked). Kernel.org uses symlinks, so it's not usable there, but I thought it might still be interesting for a site like yours. So, is your directory hierarchy on repo.or.cz symlink-free? | 19:26 |
|
jast
| of course, vim is still the best editor ;) | 19:26 |
|
| aconbere, git branch -r | 19:27 |
|
pasky
| lea_w: yes it is | 19:27 |
|
lea_w
| Cool. I just kinda implemented it, so I'll clean it up a little and leave it in. | 19:27 |
|
aconbere
| ahhh | 19:27 |
|
| interesting | 19:27 |
|
| hmmm | 19:27 |
|
| so are there two types of branch structures... remote and local? | 19:28 |
|
jast
| sort of | 19:28 |
|
| technically they're the same | 19:28 |
|
___tortoise
| thanks jast | 19:28 |
|
jast
| but remote tracking branches are meant to be exact mirrors of branches in remote repos | 19:28 |
|
Ilari
| aconbere: 'git branch -r' lists remote branches... | 19:28 |
| → hagna joined | 19:28 |
|
jast
| as such you're discouraged from using them just like local branches in several contexts | 19:28 |
|
| e.g. your HEAD gets detached when you checkout tracking branches | 19:29 |
|
| and, yes, they don't show up in the list of "local" branches | 19:29 |
|
hagna
| can I take a subdirectory out of an existing git repo and commit to it etc just like I would to the main repo? | 19:29 |
|
aconbere
| (sorry if this is explained well elsewhere, I was reading around and not finding anything) | 19:29 |
|
| so if I want to work in a remote branch, or commit to one | 19:30 |
| → venger joined | 19:30 |
|
jast
| hagna, no, sorry. you can clone a repo though and remunge it so that only that subdir is left. the results can't be readily committed back into the original repo then, though. | 19:30 |
| → bdiego joined | 19:30 |
|
hacim
| drizzd_: the problem is that I still have remote branches origin/HEAD and origin/master, and git remote prune wont remove those | 19:31 |
|
hagna
| jast: maybe a symbolic link then | 19:31 |
|
| jast: would solve my problem | 19:31 |
|
jast
| hagna, depends on what exactly you're trying to do | 19:31 |
| → scrooge joined | 19:31 |
|
jast
| slash what your problem is :) | 19:31 |
|
venger
| i'd like to have git-log (or some other command) show one output line per commit, maybe something like all (or part - first 7) of the sha1 and the commit message. just to minimize output. any ideas to try? | 19:32 |
|
Mikachu
| git log --pretty=oneline ? | 19:32 |
|
hagna
| jast: yeah I think it will work I'll try | 19:32 |
|
jast
| aconbere, i was assuming you were going to finish the sentence but apparently i get to do that: ... then you create a local branch from the remote tracking branch, using (for example) the command git checkout -b thelocalname origin/theremotebranch | 19:33 |
|
venger
| yep, thats it Mikachu - thanks | 19:33 |
|
aconbere
| jast: what does the work "track" mean... I see it come up | 19:33 |
|
| "tracking a remote branch" | 19:33 |
|
jast
| it's a bit overused, i think | 19:34 |
|
| for one, there are remote tracking branches, i.e. what you see as origin/foo in your local repo | 19:34 |
|
hacim
| bfields: how do I remove the origin/HEAD and origin/master remotes after I edit .git/config? | 19:34 |
|
aconbere
| jast: sure... so I understand that | 19:34 |
|
jast
| now if you create a local branch from that, it's normally set up so that pull fetches/merges from the proper places. some people might call that "tracking the branch". | 19:34 |
|
aconbere
| k | 19:35 |
|
jast
| (of course, it's not just pull, but also fetch and merge if you use them separately) | 19:35 |
|
aconbere
| git branch --track | 19:35 |
|
| ? | 19:35 |
|
| I guess I could read the man page for that one | 19:35 |
|
| :P | 19:35 |
|
jast
| --track is on per default for this situation | 19:36 |
|
| yeah, reading the manpage certainly can't hurt | 19:36 |
| → johnw joined | 19:36 |
|
bfields
| hacim: can probably just rm .git/refs/remotes/origin/HEAD, etc. Or if they're packed, "git push . :refs/remotes/origin/HEAD", etc. | 19:41 |
|
hacim
| bfields: what does the second packed one mean? I understand the first, but not the second | 19:42 |
| → ssb joined | 19:42 |
|
jast
| branches refs don't get packed automatically if you haven't had them packed manually before | 19:42 |
|
| (iirc) | 19:42 |
| bdiego → _bdiego | 19:43 |
| _bdiego → bdiego | 19:43 |
| ← scroogemcduck left | 19:43 |
|
aconbere
| jast: thanks for the help, I think I get it | 19:43 |
| → rolfb joined | 19:44 |
|
ssb
| hi! is there a way to merge to a branch that is not current? | 19:45 |
|
Eridius|work
| whoa, wtf? I'm in a git repo and `git stash` shows "fatal: Not a git repository" 3 times before saying "You do not have the initial commit yet". WTF? | 19:45 |
|
| ssb: merge uses the working tree to do work, so you can only merge stuff into the active branch | 19:46 |
|
ssb
| Eridius|work, ok, thanks. | 19:47 |
|
bfields
| The second one is just a fancier way of saying "delete refs/remotes/origin/HEAD". Technically git can store branch (and other) refs in a different place than .git/refs/..., so in theory a straight "rm" won't always work, though sounds (from what jast says) that this isn't likely in your case, so the rm should do the job. | 19:48 |
|
jast
| if the rm fails we can always investigate further | 19:48 |
|
| though, of course, the git push . :foo thing works no matter how things are stored | 19:49 |
|
bfields
| hacim: as for what it means, "git push ." means "push to this repository ("." is just a path, so means this repo), and "a:b" in a push just means push local ref "a" to remote ref "b", and if "a" is absent then that's taken to mean "delete b". | 19:50 |
|
| Eridius|work loves `git push . foo:bar` | 19:50 |
|
Eridius|work
| `git fetch && git push . origin/next:next && git rebase next` is my frequent operation for git.git | 19:51 |
| ← JaapSuter left | 19:51 |
|
Mikachu
| why don't you just git rebase origin/next? :) | 19:51 |
|
lea_w
| why not configure it so you can simply run "git pull" and "git push"? | 19:51 |
|
Eridius|work
| Mikachu: I could, but I like to keep my local next branch intact | 19:51 |
|
Mikachu
| why do you then have a local next branch? | 19:52 |
|
Eridius|work
| because I used to sit on that, but right now I'm holding onto a commit from dscho on another branch | 19:52 |
| → name joined | 19:52 |
| → snailrails joined | 19:52 |
|
snailrails
| what do you do when two files conflict ? | 19:52 |
| → robin_ joined | 19:52 |
|
Eridius|work
| I suppose I could just let the local next branch languish until I want to use it again | 19:52 |
|
name
| wow git-svn takes a long time :) | 19:52 |
|
Mikachu
| you don't need it if you don't want to commit on it, you can just checkout origin/next | 19:52 |
|
Eridius|work
| I don't like detached heads | 19:53 |
| ← rolfb left | 19:53 |
|
hacim
| bfields: thanks, that is useful information | 19:53 |
|
Mikachu
| git fetch; git checkout origin/next | 19:53 |
|
Eridius|work
| I suppose I could also just do `git pull --rebase origin next` for the time being | 19:53 |
|
| in any case, anybody know why `git stash` fails if you're behind a symlink in a git repo? | 19:54 |
|
jast
| snailrails, edit the conflicts away and git add the file | 19:54 |
|
hacim
| how do I push just one branch? I did "git push git+ssh://git/url/here branchname" which seemed to push, but then when I tried to clone it elsewhere I got the message "Warning: Remote HEAD refers to nonexistet ref, unable to checkout" | 19:54 |
| ← nuge left | 19:56 |
| → d0k joined | 19:58 |
|
jast
| hacim, can you find out what HEAD (in the remote repo) does point to? is that a bare repo you've been pushing to, anyway? | 19:59 |
|
hacim
| jast: yes, its a bare repo, I just created it | 20:00 |
| ← guillermo_spain left | 20:00 |
|
jast
| so, what's in HEAD? | 20:01 |
|
hacim
| jast: how do I find out? | 20:02 |
|
| jast: oh I see in refs/heads/ there is <branchname> | 20:02 |
|
jast
| if you can, just ssh to the repo server and see what HEAD in the repo dir contains | 20:03 |
| → spearce` joined | 20:03 |
| ← spearce left | 20:03 |
|
hacim
| jast: it has refs/heads/master, but I have not pushed a master ever | 20:03 |
| DrNick2 → DrNick | 20:03 |
|
hacim
| ah, I think I know what I should have done | 20:04 |
|
| I should have instead done: git push git+ssh://git/url/here branchname:refs/heads/master | 20:04 |
|
jast
| that would have created master, yeah | 20:05 |
| spearce` → spearce | 20:05 |
| ← johnw left | 20:05 |
|
jast
| i think it shouldn't be necessary to use the name master, though | 20:05 |
| → johnw joined | 20:05 |
|
jast
| but apparently git init always sets HEAD to master | 20:08 |
| ← snailrails left | 20:08 |
|
hacim
| jast: yeah I think you are right | 20:10 |
|
| thats fine, I learned something :) | 20:10 |
|
jast
| me too | 20:10 |
| ← ketralnis left | 20:14 |
| ckoehler → ckoehler_ | 20:14 |
| → tarbo joined | 20:15 |
| → johnw_ joined | 20:15 |
| ckoehler_ → ckoehler__ | 20:15 |
| → anlek joined | 20:15 |
|
anlek
| hey guys | 20:15 |
| → k0002 joined | 20:15 |
|
anlek
| this may sound like a noob question but how to do revert a file to the version in GIT? | 20:16 |
|
dvorak_
| git checkout <filename> | 20:17 |
|
anlek
| really? that easy? | 20:17 |
| ← lorandi left | 20:17 |
|
dvorak_
| works for me | 20:17 |
|
anlek
| good stuff! :) | 20:17 |
| → SRabbelier joined | 20:17 |
| → kevwil joined | 20:18 |
| ← SRabbelier left | 20:18 |
| ← anlek left | 20:20 |
| → SuttoL joined | 20:21 |
|
hacim
| is there an easy way to re-write commit ids? | 20:22 |
|
| I have two repositories that were checked out via git-svn, and they used different URIs so they have 'no common commits', but they do | 20:23 |
| → radarek joined | 20:23 |
| → ciaran_ joined | 20:23 |
| ← johnw left | 20:23 |
| ← janm left | 20:24 |
| → jli joined | 20:27 |
|
hacim
| how can I make a patch with the commit message intact without creating emails out of them (ie. git-format-patch)? | 20:27 |
|
drizzd_
| git diff | 20:28 |
|
| or rather, git log | 20:28 |
| ← ctennis left | 20:28 |
| → kewark joined | 20:28 |
|
Mikachu
| git show is probably a good fit | 20:30 |
|
| if you use git log -p, they'll all be concatenated | 20:31 |
|
hacim
| yeah I need to only find the ones that were added to this branch | 20:31 |
|
jast
| the output of neither can be processed automatically, though | 20:31 |
| ← k0001 left | 20:32 |
|
cebewee
| looks like an use case for rebase? | 20:32 |
|
Mikachu
| if someone puts a patch in the commit message you mean? | 20:32 |
| ← kevwil left | 20:32 |
| ← markkald_ left | 20:32 |
|
jast
| no, i mean because i don't think either git apply or git am process the output format of log/show | 20:33 |
| → markkalderon joined | 20:33 |
|
Mikachu
| it's just text and a patch | 20:33 |
|
jast
| well, yes, apply will deal with it | 20:33 |
|
| but you don't get the commit message automatically set in any useful way | 20:34 |
|
hacim
| hrm | 20:35 |
|
cebewee
| hacim: You want to change one of the repositories so that they share the same svn commits? | 20:35 |
|
hacim
| I am confused... I want all the diffs between branch 'test' and 'master' with the commit messages | 20:35 |
| ← markkalderon left | 20:35 |
|
jast
| yes, git log -p will help there, but the output can't be automatically converted back into sensible commits | 20:36 |
| ← pflanze left | 20:38 |
|
hacim
| well, the problem is that I have 'no common commits' but they are common, and I just want the changes I've made so I can use the new format | 20:39 |
| ← thannoy left | 20:39 |
|
Mikachu
| sure you don't want rebase --onto? | 20:40 |
|
hacim
| maybe, I'll have to read up on that | 20:40 |
|
jast
| don't mess with rebase in git-svn repos unless you know what you're doing | 20:42 |
| → ffm|cake joined | 20:42 |
|
ffm|cake
| How do I throw away my last commit? | 20:42 |
|
Mikachu
| how far? | 20:42 |
|
| do you want the changes to stay in the working tree? | 20:42 |
|
jengelh
| 57mm | 20:42 |
|
ffm|cake
| Mikachu: The last one I made, it hasn;t been synced with remote yet. | 20:42 |
|
| Mikachu: Yes. | 20:42 |
|
Mikachu
| git reset HEAD^ | 20:42 |
|
| add soft if you want to keep it in the index | 20:43 |
|
jast
| bah, you're faster than me | 20:43 |
|
Mikachu
| but you can just make the changes you want, and use git commit --amend when you're happy | 20:43 |
|
ffm|cake
| Mikachu: Really? | 20:43 |
|
Mikachu
| really | 20:43 |
|
jast
| --amend deserves lots of love | 20:43 |
|
ffm|cake
| Mikachu: --amend in'st in the manpage... | 20:43 |
| → _krawek_ joined | 20:43 |
|
ffm|cake
| Coo. | 20:43 |
|
| l | 20:43 |
|
Mikachu
| yes it is | 20:43 |
|
jast
| (don't forget to git add your changes before you amend) | 20:43 |
|
Mikachu
| let's ask about your version | 20:44 |
|
ffm|cake
| ffm@crank:~/olpc-getone-utils$ git --version | 20:44 |
|
| git version 1.5.3.8 | 20:44 |
| ← krawek left | 20:45 |
|
ffm|cake
| Mikachu: ^^ | 20:45 |
| ← anttih left | 20:46 |
|
Mikachu
| that should work then | 20:46 |
|
ffm|cake
| Mikachu: It does. I'm just wondering why "man git-commit" doesn't show it. | 20:47 |
|
Mikachu
| maybe you have old manpages | 20:47 |
|
| man git-commit|tail -1 | 20:47 |
|
hacim
| ok, I'm totally confused now :o | 20:47 |
| ckoehler__ → ckoehler | 20:47 |
| ← sergiodj left | 20:48 |
|
ffm|cake
| Git 1.5.3.8 01/25/2008 GIT-COMMIT(1) | 20:48 |
|
| Mikachu: ^^^^ | 20:49 |
|
| ffm|cake complains to the ubuntu devels. | 20:50 |
|
Mikachu
| it's been documented since 1.3.0 | 20:50 |
|
doener_
| ffm|cake: --amend is in git-commit(1) since 1.3.0 | 20:50 |
|
Mikachu
| according to git | 20:50 |
|
hacim
| i migrated a subversion repository to a git repository using this procedure: http://www.simplisticcomplexity.com/2008/03/05/cleanly-migrate-your-subversion-repository-to-a-git-repository. Now I've got two repositories: one is a checkout of this clean migration, the second is a repository where I have been doing git-svn, and I have a branch that I now want to convert to this clean repository... but they dont share common commits and I dont know how to rec | 20:50 |
|
Mikachu
| know how to rec | 20:50 |
|
ffm|cake
| kk | 20:50 |
| → wurble` joined | 20:52 |
|
doener_
| hacim: your message was cut off after "know how to rec" | 20:52 |
|
wurble`
| How do I get just the file names modified in a given commit (specifically HEAD)? | 20:52 |
|
hacim
| ... dont know how to reconcile this | 20:52 |
| → z3om joined | 20:52 |
|
hacim
| doener_: that was almost the end :) | 20:53 |
|
cardioid
| I thought Mikachu's response was an answer | 20:53 |
|
doener_
| hacim: but it sounds like you're looking for rebase --onto or want some grafts magic | 20:53 |
| ← zeom left | 20:53 |
| z3om → zeom | 20:53 |
|
cardioid
| "I don't know how to rec" "Then figure out how" | 20:53 |
|
Mikachu
| wurble`: --name-only | 20:53 |
|
doener_
| hacim: does your "other branch" have any merges in it? | 20:53 |
|
| hacim: ... in the part of the history that you want to tack onto the new repo | 20:54 |
|
hacim
| doener_: no, I just branched the git-svn checkout and did some commits there | 20:55 |
| ← docwhat left | 20:55 |
|
doener_
| hacim: then just "git fetch path/to/other/repo branch:refs/heads/tmp" and "git rebase --onto target_branch first_commit_to_get^ tmp" | 20:56 |
|
wurble`
| Mikachu: That still gives me the commit message etc. I want to feed the names to a script | 20:56 |
|
doener_
| hacim: note the ^ in there | 20:56 |
|
name
| can i commit to my svn repo using git-svn? | 20:56 |
|
Mikachu
| wurble`: you didn't say what command you were using | 20:57 |
|
doener_
| name: yes | 20:57 |
|
wurble`
| Mikachu: log | 20:57 |
| johnw_ → johnw | 20:57 |
|
wurble`
| name: dcommit I think | 20:57 |
|
Mikachu
| wurble`: --pretty=format: | 20:58 |
|
hacim
| doener_: is 'tmp' the name of the branch? | 20:58 |
| ← ffm|cake left | 20:58 |
|
wurble`
| Mikachu: Thanks. Is there a better way than log? | 20:58 |
|
| (Just out of curiosity) | 20:59 |
|
Mikachu
| git diff? | 20:59 |
|
doener_
| hacim: "tmp" is what it will be called in the new repo. The branch name in your git-svn repo would be "branch" (left of the colon in the fetch command) | 20:59 |
|
Mishakun
| name: doesn't "git-svn commit" work? | 20:59 |
|
doener_
| Mishakun: it's "dcommit" | 21:00 |
|
Mishakun
| aight | 21:00 |
| ← kewark left | 21:00 |
|
name
| will it also create branches etc the right way? | 21:02 |
|
| i created a new branch, will it go into the branches/ directory? | 21:02 |
|
doener_
| it won't create branches at all | 21:02 |
|
name
| hm crap | 21:03 |
|
| how do i commit my new branch then? | 21:03 |
|
hacim
| doener_: nice, then I can just delete the tmp branch afterwards | 21:03 |
|
doener_
| hacim: yep | 21:03 |
|
hacim
| doener_: what would I do if there were merges? | 21:03 |
|
name
| or can i only commit to trunk by merging into it | 21:04 |
|
doener_
| hacim: either the same thing, but you would get a linear history then, or use grafts and filter-branch to connect the history | 21:04 |
|
name
| not into a new branch.. | 21:04 |
|
doener_
| hacim: you could have done the same now, but the rebase is easier | 21:04 |
|
| name: try "git svn dcommit -n" to see where the commits are going | 21:05 |
|
hacim
| i'm ok with a linear history I think | 21:05 |
|
name
| doener_: to trunk :( | 21:05 |
|
hacim
| its not a lot of changes | 21:05 |
| → peeja joined | 21:05 |
| ← emachado left | 21:05 |
|
name
| damnit | 21:05 |
|
doener_
| name: if you need a new branch in svn, you can create one using svn, then "git svn fetch" to get the new branch and use "git rebase" with --onto to get your commits over there | 21:06 |
| → redondos joined | 21:06 |
| ← Mishakun left | 21:06 |
|
doener_
| name: you do track the svn branches, right? | 21:06 |
|
name
| doener_: sure | 21:06 |
|
doener_
| name: http://git.pastebin.com/m2cb6c24c | 21:06 |
|
| name: that script uses the svn binary to create a new svn branch from the latest svn commit in your current branch | 21:07 |
|
hacim
| doener_: thanks, thats exactly what I was hitting my head over :) | 21:07 |
|
name
| doener_: i hope for you that it works :) | 21:08 |
|
doener_
| name: so if you forked from trunk, you can do: "git-svn-branch my_branch; git svn fetch; git rebase --onto remotes/my_branch trunk my_branch" | 21:09 |
|
| name: and when that's done, "dcommit -n" should tell that you're dcommitting to my_branch now | 21:09 |
|
name
| Permission denied (publickey). | 21:10 |
|
| svn: Connection closed unexpectedly | 21:10 |
|
doener_
| looks like svn isn't correctly setup then... o.O | 21:10 |
|
name
| well it is | 21:10 |
|
| i need my ssh key to commit | 21:10 |
|
doener_
| so do I | 21:11 |
|
| name: you could put an "echo" in front of the svn call in the script to check if the command looks good | 21:12 |
| → ebel joined | 21:12 |
|
name
| svn cp -r 138 svn+ssh://svn.gna.org/svn/pypentago/trunk svn+ssh://svn.gna.org/svn/pypentago/branches/feature-elo -m Create branch feature-elo | 21:13 |
|
| i could just do that in a normal svn checkout | 21:13 |
|
doener_
| name: creating branches doesn't even require a working copy in svn | 21:14 |
| blowery`afk → blowery | 21:14 |
|
doener_
| name: maybe you need --username in there? | 21:15 |
|
name
| probably | 21:15 |
|
| so my_branch needs to contain what's in trunk? | 21:15 |
| ← johnw left | 21:16 |
|
doener_
| name: hm? | 21:16 |
| → johnw joined | 21:16 |
|
name
| the branch i just created | 21:16 |
| ← johnw left | 21:16 |
|
name
| could it be empty and i can still commit into it? | 21:16 |
| → johnw joined | 21:16 |
| → markkalderon joined | 21:17 |
| → koke joined | 21:17 |
| ← jbunster left | 21:17 |
| ← hagna left | 21:18 |
|
doener_
| name: I don't understand the question... Your branch apparently forked off from trunk revision 138, so the script will create the svn branch from that revision | 21:19 |
|
name
| so how do i commit to that | 21:20 |
|
| screw svn branching :S | 21:21 |
|
doener_
| name: "git svn fetch" to get the branch into git, and then create commits in git, for example using the rebase command I gave | 21:21 |
| ← bfields left | 21:21 |
| → omajid joined | 21:24 |
|
name
| it seems to have worked | 21:24 |
|
| thanks :) | 21:25 |
| ← omajid left | 21:25 |
|
name
| git-svn rules :) | 21:25 |
|
| and once i'm sure about the code i can merge it into master, right? | 21:26 |
|
doener_
| yeah, but note that if you want to dcommit merge commits to svn, that's a bit hairy... | 21:27 |
|
name
| can't i "flatten" the commit? | 21:27 |
|
| jengelh flattens name | 21:27 |
|
name
| so svn doesn't know that it was merged | 21:27 |
|
jengelh
| svn does not know merges anyway | 21:27 |
|
name
| and thinks i just changed trunk | 21:27 |
|
doener_
| name: svn doesn't know anyway | 21:27 |
|
name
| why is it hairy? | 21:27 |
|
jengelh
| because it does not know! | 21:27 |
|
name
| ah okay | 21:28 |
|
doener_
| name: because git svn uses rebase internally | 21:28 |
| → b0ef joined | 21:28 |
| ← marc_vw left | 21:28 |
|
name
| that means? | 21:28 |
|
doener_
| name: if you want to dcommit a merge, make sure that you dcommit that merge commit _only_ no other commits | 21:28 |
|
| otherwise, a rebase might kick in and you would get the whole lot of commits that were merged duplicated on your current branch | 21:29 |
|
b0ef
| ehlo | 21:29 |
|
| what command can I use that does the same as "svn info" to get the url of the origin? | 21:29 |
|
Eridius|work
| git remote show -n origin | 21:29 |
|
doener_
| b0ef: vim .git/config ;-) | 21:29 |
|
jengelh
| b0ef: git remote -v | 21:29 |
|
doener_
| or what Eridius|work said | 21:29 |
|
jengelh
| lol | 21:29 |
|
Eridius|work
| the -n means don't bother trying to contact the origin to verify remote branches and availability | 21:29 |
|
jengelh
| THREE ways... that's more than Perl could give ya! | 21:29 |
|
b0ef
| hehe; thanks all | 21:29 |
| → rdavila joined | 21:30 |
|
jengelh
| b0ef: not having the esben-stien.name DNS anymore? | 21:31 |
|
b0ef
| I guess I'm using an old version, cause none of those commands works | 21:31 |
|
jengelh
| (the reverse dns from ip->host) | 21:31 |
|
b0ef
| I'm on 1.4, so I'll grab 1.5 | 21:31 |
|
| jengelh: sure | 21:31 |
|
jengelh
| not so sure, 23:28 -!- b0ef [n=b0ef@062016141231.customer.alfanett.no] | 21:31 |
|
b0ef
| jengelh: yeah, but you see; it's fake;) | 21:32 |
|
jengelh
| it seems quite real | 21:32 |
|
b0ef
| I just point esben-stien.name to my dynamic ip | 21:32 |
|
jengelh
| I said reverse dns :p | 21:32 |
| ← BabelO left | 21:32 |
|
b0ef
| it's a different company | 21:32 |
|
jengelh
| same shit :) | 21:33 |
|
b0ef
| right, but you're contacting the ones responsible for that ip | 21:34 |
| → madewokherd joined | 21:34 |
|
| jengelh forgets about it | 21:34 |
|
b0ef
| alfanett.no | 21:34 |
| ← \ask_ left | 21:37 |
| ← peeja left | 21:40 |
| ← ph^ left | 21:42 |
| ← ckoehler left | 21:44 |
| ← danlarkin left | 21:46 |
| → Mutru joined | 21:54 |
| blowery → blowery`afk | 21:56 |
|
| telmich thinks about indexing $HOME in git.. anyone tried/used that before in here? | 21:56 |
| → carrus85 joined | 21:57 |
|
thiago
| I saw someone who did that | 21:57 |
|
| but you don't want to index your trash or cache files | 21:57 |
|
telmich
| that's true | 21:57 |
|
| it will need some time until I'll have filled .gitignore and added the right parts | 21:57 |
|
| the biggest problem I see is space -- my eeepc has only 16GiB /home | 21:58 |
|
| and my t60 has 44GiB used, of which 11GiB are mails | 21:58 |
|
jengelh
| telmich: purge linux-kernel... | 21:58 |
|
telmich
| jengelh: that's in ~compiler/ already :-) | 21:59 |
|
jengelh
| mails in ~compiler? | 21:59 |
|
telmich
| ahh, no, thought you mean kernel-sources | 21:59 |
|
thiago
| you don't want to archive your emails in Git | 22:00 |
| → orbisvicis joined | 22:01 |
|
thiago
| you will probably want to archive it on an IMAP server | 22:01 |
|
Mikachu
| i think the consensus is it's a pretty bad idea to put ~ or /etc in git | 22:01 |
| ← spearce left | 22:01 |
|
vmiklos
| "unless you know what you are doing" | 22:01 |
|
Mikachu
| even then :) | 22:01 |
|
vmiklos
| i do both. with chmod 700 .git, and i never push, pull, merge, etc. | 22:01 |
|
| erm, for ~, just the config files, not the whole dir :) | 22:02 |
|
telmich
| well, I do not want to put all stuff (like audio) into git | 22:03 |
|
dsop
| you better use backup systems for that | 22:03 |
|
| Mikachu: sorry for confusing you with indexversion | 22:03 |
|
Mikachu
| heh, no problem :) | 22:03 |
|
telmich
| but .zshrc, .profile, browser-configs, dokuments, congress plannings, etc are intesting | 22:03 |
|
dsop
| my fault, nexttime I should default grep to grep -i | 22:03 |
|
Mikachu
| dsop: i did look at the code and it does strcmp | 22:04 |
|
telmich
| but mails would still be interesting... but maybe i really have to give offlineimap or how it was called a chance | 22:04 |
| ← mvo left | 22:05 |
|
dsop
| Mikachu: but git_config provides the key lowercase? | 22:05 |
|
Mikachu
| i suppose it does, since it works | 22:05 |
|
| and almost all config options are listed with camelcase in the manpage anyway | 22:06 |
| → _zack joined | 22:09 |
|
orbisvicis
| im trying to change the location of the gitweb css ... should it be @stylesheets in gitweb.cgi ? | 22:11 |
|
| but changing that does nothing | 22:11 |
| cbrake → cbrake_away | 22:11 |
| ← _zack left | 22:11 |
| → johnw_ joined | 22:11 |
| → _zack joined | 22:12 |
| → WALoeIII joined | 22:12 |
|
orbisvicis
| nvrmind, gitweb.conf, didnt know it existed | 22:14 |
| → markkald_ joined | 22:15 |
| → WALoeIII_ joined | 22:15 |
| ← WALoeIII left | 22:15 |
| ← markkalderon left | 22:16 |
| blowery`afk → blowery | 22:16 |
| ← nud left | 22:18 |
| → W|cked joined | 22:18 |
| ← johnw left | 22:20 |
| ← Ademan left | 22:24 |
| → Ademan joined | 22:25 |
| → brendon joined | 22:25 |
| ← heycarsten left | 22:25 |
| → sergiodj joined | 22:27 |
| ← priidu left | 22:27 |
| ← scrooge left | 22:27 |
| → sverrej_ joined | 22:28 |
| ← Eridius|work left | 22:29 |
| → Eridius|work joined | 22:29 |
| ← jiva left | 22:29 |
| ← charon left | 22:33 |
| ← lea_w left | 22:34 |
|
RandalSchwartz
| "converting from CVS to git" - one day class? | 22:35 |
|
| or less? | 22:35 |
|
| or more? :) | 22:36 |
| → lea_w joined | 22:36 |
| eikonos → eikonoklastes | 22:36 |
| ← lea_w left | 22:38 |
| eikonoklastes → eikowork | 22:38 |
| eikowork → eikowerk | 22:38 |
| → lea_w joined | 22:38 |
| ← bentob0x left | 22:39 |
| ← rsk left | 22:39 |
|
sbeyer
| RandalSchwartz: hmm, imho it's better to not learn what git commands are like what CVS commands... but to just make clear, that you learn something new and show the principles. | 22:41 |
| ← insin left | 22:42 |
|
RandalSchwartz
| Yeah - I'm ttrying to figure out what to teach to existing CVS users | 22:42 |
|
Whtiger
| RandalSchwartz: name it something like "Forgetting CVS: The switch to git" | 22:42 |
|
RandalSchwartz
| and how long it takes | 22:42 |
|
jast
| it's probably best if you start over | 22:42 |
|
| otherwise you have to keep explaining differences | 22:42 |
|
RandalSchwartz
| well - everyone will want a contrast | 22:42 |
| ← _zack left | 22:43 |
|
jast
| what's the target audience like beyond "cvs users"? | 22:43 |
|
RandalSchwartz
| that's all I know | 22:43 |
|
vmiklos
| that depends on how experienced are those users. for example if they never branched and tagged, you can avoid explaining all the git branching/merging commands in the first round :) | 22:43 |
|
RandalSchwartz
| welcome to my world. ) | 22:43 |
|
jast
| is that a phyical class or an online thing? | 22:43 |
|
sbeyer
| RandalSchwartz: I think most CVS users do stuff like "cvs update" and "cvs commit" and nothing else ;) | 22:44 |
|
RandalSchwartz
| in person | 22:44 |
|
jast
| ah, cool | 22:44 |
|
Whtiger
| Are you trying to convince them to switch? | 22:44 |
|
| or just teaching them git | 22:44 |
|
jast
| personally, i would concentrate less on the contrasting thing. dive right into git and structure it so that the differences show up pretty early, but with minimal explicit highlighting of them | 22:44 |
|
vmiklos
| (and cvs diff -u, cvs add. but that's all yes. usually, i think.) | 22:45 |
|
sbeyer
| RandalSchwartz: at my internship there was a 60(?) minute talk about git... the audience was mainly CVS, SVN and quilt users ;) | 22:45 |
| ← sverrej left | 22:45 |
| ← jli left | 22:45 |
|
sbeyer
| RandalSchwartz: but as I saw later, the talk was like yours that can be found on video.google.com | 22:45 |
|
vmiklos
| sbeyer: did you talk then about dvcs in general, or about git? (i remember most of Linus' talk was about blaming svn, hyping dvcs, and it was a very little part of it about git itself) | 22:46 |
|
sbeyer
| vmiklos: it was about git | 22:46 |
| blowery → blowery`afk | 22:46 |
|
vmiklos
| nice:) | 22:47 |
|
sbeyer
| vmiklos: it started with the objects (commits, trees, blobs) and the implications (history etc) | 22:47 |
|
| vmiklos: and it ended up with rebase and interaction with the CVS servers, that are still used | 22:48 |
| → jli joined | 22:48 |
|
vmiklos
| ack | 22:48 |
|
sbeyer
| vmiklos: when the talker said that after merging, the history of *each merged branch* is still available, that was the second where I decided to try git ;) | 22:48 |
|
vmiklos
| hehe | 22:49 |
|
sbeyer
| vmiklos: (I've used darcs privately before and a little hg repo for university homeworks with a friend.) | 22:49 |
|
vmiklos
| :] | 22:49 |
|
| originally the speed was the only reason i decided to play with git, i was happy with darcs anyway. | 22:49 |
|
sbeyer
| I was happy, too, but I never tried branching/merging in it. Is it even possible? ;) | 22:50 |
|
vmiklos
| then i wrote darcs-like porcelain on top of git, to learn a bit more about plumbing, i still happen to use it sometimes. | 22:50 |
|
jast
| i think darcs makes too many assumptions | 22:51 |
|
sbeyer
| hmm the only private rpeo of mine, that is still in darcs, is my homepage | 22:51 |
|
vmiklos
| sbeyer: yes, but only for small repos. for big repos merges never end. | 22:51 |
|
jast
| this whole patch theory thing only really pays of if unrelated patches don't interact | 22:51 |
|
| *off | 22:52 |
| ← EmilMedve left | 22:52 |
|
vmiklos
| for example there is an equivalent of git-cherry-pick in darcs, but it was so slow i never used it, after once i waited that *20 minutes* it completed the first time :) | 22:53 |
|
sbeyer
| hmm | 22:54 |
| ← orbisvicis left | 22:54 |
|
sbeyer
| ok I've never waited 20 minutes, but yes, darcs is slow. | 22:54 |
|
vmiklos
| ah, that repo had about 30k patches | 22:54 |
|
jengelh
| Slow is the good of all coffee breaks. | 22:54 |
|
sbeyer
| or doing swordplay | 22:54 |
|
| .o( xkcd ) | 22:55 |
|
vmiklos
| http://frugalware.org/~vmiklos/docs/dr-git-bench <- this is very incomplete benchmark i did about a year ago. | 22:55 |
|
| but the numbers are still interesting, imho :) | 22:55 |
| ← WALoeIII_ left | 22:56 |
|
vmiklos
| (and the text is project-specific, but oh well.) | 22:56 |
| → paltman joined | 22:57 |
| ← Bryson left | 22:58 |
| → spearce joined | 22:58 |
| ← venger left | 22:59 |
|
paltman
| trying to migrate from svn to git, everything seems to have worked except that when i do a git branch -a, it shows master as selected and all the other branches are listed as remote and I can't switch to them | 22:59 |
|
| any tips for how to suck in my subversion history including my branches? | 22:59 |
|
thiago
| they are there | 22:59 |
|
| you've just said the other branches are listed | 22:59 |
|
paltman
| but i can't checkout any of them | 22:59 |
|
thiago
| why not? | 22:59 |
|
| RandalSchwartz averts his eyes at the mention of SVN | 23:00 |
|
jackdempsey
| hehehe | 23:00 |
|
thiago
| paltman: you have to create a local tracking branch first | 23:00 |
|
paltman
| it tells me that i am moving to a branch that isn't local | 23:00 |
|
thiago
| you shouldn't checkout a remote branch | 23:01 |
|
| aconbere wishes git-svn did everything he wanted safely | 23:01 |
|
paltman
| how do i make all the branches imported by git-svn local? | 23:01 |
|
| thiago wishes git-svn were fast | 23:01 |
|
thiago
| paltman: I don't know. Are you converting the repository? | 23:01 |
|
paltman
| yes | 23:01 |
|
thiago
| then I don't know | 23:01 |
|
aconbere
| thiago: I would just settle with it letting me make my own clones and push them back up | 23:02 |
|
paltman
| i want to do a one time conversion | 23:02 |
|
aconbere
| (without breaking my svn repository) | 23:02 |
|
paltman
| leaving svn | 23:02 |
|
| in the dust | 23:02 |
| → bgerber joined | 23:02 |
|
paltman
| but i don't want to wait until we are ready to merge all the branches into trunk | 23:02 |
| → foca joined | 23:02 |
| ← juanjoc left | 23:03 |
| ← zeom left | 23:03 |
|
jast
| paltman, you sort of do that on demand. whenever you need a branch, you "localize" (my term, not the official one) it | 23:06 |
| ← Yuuhi left | 23:06 |
|
jast
| it can be automated but it's usually not worth the effort | 23:06 |
|
paltman
| how do i localize? | 23:06 |
| ← ebel left | 23:06 |
|
jast
| it's rather easy to create a local branch from a remote one, anyway: | 23:06 |
|
| bah, i don't remember how exactly git-svn calls the remote branches | 23:06 |
| ← spearce left | 23:06 |
|
jast
| anyway: git checkout -b localname remotename | 23:07 |
| ← cmarcelo left | 23:07 |
|
jast
| (you'd best use the full remotename as listed by branch -r/-a) | 23:07 |
| ← d0k left | 23:08 |
|
Jearil
| isn't there a git-svnimport for importing an svn repository? | 23:08 |
|
jast
| nobody uses that anymore. everybody uses git-svn. :} | 23:08 |
|
| and it's shipped with git (both are, actually) | 23:09 |
|
paltman
| i don't have svnimport | 23:09 |
|
sbeyer
| hmm, I wonder why git remote update says " + f4fe2a0...1363959 master -> vmiklos/master (forced update)" ... I haven't forced. Or is it because published history has changed and I don't have a tracking branch? | 23:09 |
|
paltman
| maybe that would work better | 23:09 |
|
jast
| perhaps it isn't anymore. git-svn is better anyway. | 23:09 |
|
sbeyer
| vmiklos: do you rebase in master? ;) | 23:09 |
|
paltman
| doesn't seem to work very well (or i am not feeding it hte right parameters) | 23:09 |
|
jast
| sbeyer, i suppose the upstream branch was reset or rebased or something. | 23:10 |
|
| anything else would be rather unlikely. | 23:10 |
|
| paltman, well, what's the problem? | 23:10 |
|
sbeyer
| jast: yes, that's what I meant with "published history has changed"... but why "forced update"? | 23:10 |
|
paltman
| i specified explicity -t tags -b branches -T trunk (because there where other stuff at the root of my repo) | 23:10 |
|
| it seemed to pull everything down fine | 23:11 |
|
sbeyer
| jast: (I think it is because I don't have a tracking branch on it and I think this is sane behavior, but I'm confused a bit...) | 23:11 |
|
paltman
| but then a git branch -a | 23:11 |
|
| showed master and then all branches (including trunk) as remote | 23:11 |
|
jast
| sbeyer, vmiklos/master is a tracking branch | 23:11 |
|
vmiklos
| sbeyer: sure. | 23:11 |
|
paltman
| most of the branches where deleted a long time ago | 23:11 |
| ← radarek left | 23:11 |
|
paltman
| and don't exist at HEAD anymore | 23:11 |
|
sbeyer
| jast: no, that's a remote branch ;) | 23:11 |
|
paltman
| so was confused to see them there | 23:11 |
|
sbeyer
| vmiklos: oh ok | 23:12 |
|
paltman
| if trunk is listed a remote branch, then what is in master? | 23:12 |
|
jast
| paltman, too many branches is better than too few, isn't it? :) and to get local branches from remote ones, i already suggested something | 23:12 |
|
| master defaults to the last-updated branch in svn at the time you imported the repo. a bit crazy, i admit. | 23:12 |
|
paltman
| right and i see that your suggestion works | 23:12 |
|
| ahh that makes sense | 23:12 |
|
jast
| remote branches are sort of mirrors of the svn branches | 23:12 |
|
| so you never touch them and create local ones from them whenever you want to work on a branch | 23:13 |
|
paltman
| so i should probably just delete the remotes that are no longer valid branches just to clean up a bit and i think i'm all set | 23:13 |
|
| thanks | 23:13 |
|
| is there anyway to change what master is? | 23:14 |
|
Jearil
| can really just delete master and create it again from a different branch | 23:14 |
| ← redondos left | 23:14 |
|
paltman
| that last update in my subversion repository was in an experiemental branch and i would prefer it be another branch | 23:15 |
|
| so i can do git branch -d master | 23:15 |
|
jast
| paltman, you can create a different branch, switch to it and then just delete master. | 23:15 |
| queso → queso|away | 23:15 |
|
jast
| (you can't delete the current branch) | 23:15 |
|
Jearil
| yeah.. the name master doesn't really mean anything | 23:15 |
|
paltman
| and then git checkout -b master my-remote-svn-branch | 23:15 |
|
| ? | 23:15 |
|
Jearil
| it's more of a convention | 23:15 |
| ← koke left | 23:16 |
| → spearce joined | 23:16 |
|
jast
| for example, yes. nothing forces you to call your branch master, anyway. | 23:17 |
|
| it's just a default name. | 23:17 |
|
paltman
| thanks guys, that has cleared a lot of confusion up for me | 23:18 |
|
name
| how do i undo all changes since last commit? | 23:18 |
|
Mikachu
| git reset --hard | 23:20 |
|
| it won't ask for confirmation | 23:20 |
|
name
| thanks | 23:20 |
| → les joined | 23:21 |
|
paltman
| one more thing, in the process of doing the migration, is there a way to clean up what the remote branch names will be (going to use GitHub as a central place to share with team) | 23:22 |
|
vmiklos
| sbeyer: it contains random patches where i was lazy to create a separate topic branch, so i frequently rebase it against junio's master. | 23:22 |
|
name
| how do i tell it to just take the file from the other branch when i get a conflict | 23:23 |
| → mithro joined | 23:23 |
|
jast
| paltman, if you push to somewhere else remote branches aren't included anyway (by default) | 23:23 |
|
vmiklos
| name: git checkout otherbranch -- path/to/file. | 23:23 |
|
paltman
| so if i pushed somewhere to seed a common repo for others to clone, i would need to checkout and create the local branches that i wanted first? | 23:24 |
|
name
| why in gods name does git not auto merge spaces? | 23:24 |
|
paltman
| and that push would include my locals? | 23:24 |
| ← glommer left | 23:24 |
|
vmiklos
| write your own strategy which does :) | 23:24 |
|
jengelh
| name: to annoy people that needlessy fuck up the spacing | 23:25 |
|
name
| jengelh: really? | 23:25 |
|
| i just fixed it so it complies with python coding rules :) can i somehow make it auto merge? | 23:25 |
|
paltman
| this is what i needed.... https://37s.backpackit.com/pub/1465067 | 23:25 |
|
| sorry that i didn't find it before asking you all the simple questions | 23:26 |
|
name
| http://paste.pocoo.org/show/78511/ < this is kinda weird for git having good merge algorithms | 23:26 |
|
jast
| paltman, i would have ignored you had i had any objections to your questions, so never mind | 23:26 |
|
vmiklos
| once you resolved a merge conflict, git remembers about it. at least till you run git rerere gc | 23:27 |
|
name
| but why can't it auto merge spaces :S | 23:27 |
|
jast
| name, because it would have to understand the files in order to know which whitespace is relevant and which isn't | 23:28 |
|
| example: in python indentation matters, in C it doesn't | 23:29 |
|
| also whitespace always matters within strings and it often doesn't outside them | 23:29 |
|
name
| jast: the algorithm should know that ;) | 23:31 |
|
jast
| no. | 23:31 |
|
robin_
| that would require more work | 23:32 |
|
jast
| that's beyond what a generic merge algorithm should do | 23:32 |
|
pygi
| jast, don't tell people indentation doesn't matter in C! | 23:32 |
|
jast
| you can always write your own auto-merger for whitespace-only conflicts in files and run it whenever needed | 23:32 |
|
pygi
| jast, that's why we end up with tons of unreadable code xD | 23:32 |
|
jast
| pygi, code that doesn't lend itself to being submitted at ioccc isn't worth writing in the first place :P | 23:33 |
|
| pygi shoots down jast | 23:33 |
|
| jast dies | 23:34 |
|
pygi
| fine, that should make the world a better place! :) | 23:35 |
|
name
| ah uh. now even newlines? | 23:36 |
|
| *not | 23:36 |
|
noodl
| is there a way to commit all changes to a list of files regardless of them being in the index or not and other changes being in the index or not? | 23:36 |
|
jast
| nope. incidentally, whitespace patches are generally frowned upon. not just in git land, even. | 23:37 |
| → gerberb joined | 23:37 |
|
jast
| noodl, you mean git commit -a? | 23:37 |
|
robin_
| noodl: git commit <list of files> | 23:37 |
|
jast
| oh, list of files? as in, a list of your choice? | 23:37 |
|
noodl
| jast: yes | 23:37 |
|
name
| jast: it's not my fault that some coder got the whitespaces wrong | 23:38 |
|
noodl
| say a is changed but not in the index, b is changed and in the index, i want to commit the cahnges to a but not b | 23:38 |
|
jast
| i'm afraid that involves a few hoops to jump through | 23:38 |
|
noodl
| (git add isn't hard to do. i'm just wondering if there's a granular version of git commit -a) | 23:38 |
|
jast
| hmm. tricky. | 23:38 |
|
| especially if you want to exclude already staged things | 23:39 |
|
noodl
| no's a fine answer :) | 23:39 |
|
sbeyer
| vmiklos: "writing your own strategy"? How does that work? ;) (Or do you have a pointer, e.g. a mail on the list?) | 23:39 |
|
jast
| it's possible for sure | 23:39 |
| ← pombreda left | 23:39 |
|
jast
| perhaps: git diff >unstaged.patch; git diff --cached >staged.patch; git reset --hard; git apply unstaged.patch; git commit -a; git apply staged.patch | 23:39 |
|
robin_
| noodl: yes is the answer | 23:40 |
|
| ^ | 23:40 |
|
jast
| vmiklos, doesn't the rerere dir in .git not get created by default? | 23:40 |
|
noodl
| robin_: heh, so it is. sorry for the noise | 23:40 |
| → ketralnis joined | 23:41 |
|
jast
| hmm, i can see an even easier way, actually... | 23:41 |
|
spearce
| robin_: were you series about dumb transport smb:// support in jgit? | 23:41 |
| → evanlight joined | 23:41 |
|
vmiklos
| sbeyer: not really. at the moment you have to patch git-merge to add a new strategy. but once builtin-merge is in next/master, i will try to remove that limitation :) (along with docs) | 23:41 |
|
jast
| git stash; git stash apply; git commit -a; git stash apply --index; git stash drop | 23:42 |
|
vmiklos
| sbeyer: basically you get the remotes to merge as parameters and you need to do the merge in the working dir, see git-merge-resolve.sh for example, that's pretty simple. | 23:42 |
|
jast
| haven't tried either approach yet | 23:42 |
| evanlight → elight | 23:42 |
|
vmiklos
| jast: afaik it's created automatically. | 23:42 |
|
jast
| oh wait, that doesn't work correctly | 23:42 |
|
| back to the first approach then, i guess | 23:42 |
|
| vmiklos, it isn't in 1.5.4.3 at least. can't check newer versions right now. | 23:43 |
|
vmiklos
| ah, you need git config rerere.enabled true? | 23:44 |
|
| (yes, i just should have read man git-rerere more carefully.) | 23:45 |
|
jast
| dunno. i haven't needed rerere so far. | 23:45 |
| → spearce` joined | 23:45 |
| ← spearce left | 23:45 |
| → heycarsten joined | 23:46 |
| ← bgerber left | 23:47 |
| ← chris2 left | 23:49 |
| → branstrom joined | 23:49 |
| ← les left | 23:53 |
| spearce` → spearce | 23:53 |
| ← deavid left | 23:55 |