| 2008-05-25 |
| ← davi left | 00:02 |
| → zachinglis joined | 00:03 |
| ← bdiego left | 00:04 |
| → dwave- joined | 00:05 |
| → jerbear joined | 00:07 |
| → threeve joined | 00:11 |
| ← schacon_ left | 00:12 |
| ← dwave_ left | 00:18 |
| ← kukks left | 00:20 |
| → GarethAdams joined | 00:31 |
| → igorgue_ joined | 00:33 |
| → schacon joined | 00:38 |
| → zmonx joined | 00:38 |
| ← igorgue left | 00:50 |
| ← zmonx left | 00:51 |
| ← schacon left | 00:52 |
| ← saintdev left | 00:52 |
| igorgue_ → igorgue | 00:52 |
| → saintdev joined | 00:55 |
| → koke joined | 01:03 |
| ← igorgue left | 01:05 |
| → igorgue joined | 01:05 |
| → doener_ joined | 01:05 |
| → brosner joined | 01:09 |
| ← doener left | 01:18 |
| ← radarek left | 01:20 |
| → consolers joined | 01:20 |
|
cout
| how do I undo a git rm? | 01:28 |
| ← BabelO left | 01:28 |
|
tpope
| git checkout HEAD filename | 01:33 |
| ← threeve left | 01:34 |
|
cout
| ok, I remember that now | 01:35 |
|
| feels unintuitive to me, but so did "cvs add" to undo "cvs remove" when I friirst started using cvs yuears ago | 01:36 |
| → nkallen joined | 01:37 |
| → lirit joined | 01:45 |
| ← felipec left | 01:46 |
| → bugfux joined | 01:47 |
| ← goibhniu left | 01:50 |
| ← cypher23 left | 01:53 |
| → kukks joined | 01:55 |
| ← koke left | 02:00 |
| → Xyrio joined | 02:07 |
| → bwwww joined | 02:09 |
| ← saintdev left | 02:18 |
| ← j_cook left | 02:20 |
| ← roylee17 left | 02:21 |
| → threeve joined | 02:29 |
| ← bwwww left | 02:32 |
| ← leachim6 left | 02:32 |
| ← zachinglis left | 02:33 |
| → CodeOfficer joined | 02:39 |
| → zebrahat joined | 02:44 |
| ← mtodd left | 02:47 |
| → mtodd joined | 02:48 |
|
bugfux
| how cam I merge a remote without the message Merge commit 'origin/master' | 02:51 |
| → pergesu joined | 02:55 |
| → FunkeeMonk joined | 02:58 |
|
Mikachu
| bugfux: you can git commit --amend to change the message | 02:59 |
|
bugfux
| i dont want any message | 02:59 |
|
| all im doing is pulling in from remote | 02:59 |
|
| so i dont want anything to say that im merging | 02:59 |
|
Mikachu
| then don't have local changes | 03:00 |
|
bugfux
| hmm | 03:00 |
|
| thats bananas | 03:00 |
| ← nkallen left | 03:00 |
|
Mikachu
| or use rebase instead of pull | 03:00 |
|
| you can't make commits without making commits | 03:00 |
| → nkallen joined | 03:04 |
| ← das-q left | 03:06 |
|
bugfux
| hmmm | 03:06 |
| → G_SabinoMullane joined | 03:09 |
|
Mikachu
| is the format of info/grafts documented anywhere? | 03:10 |
|
| ah, repository-layout.txt | 03:10 |
|
consolers
| "Whatsoever ye shall bind on earth shall be bound in heaven: and whatsoever ye shall loose on earth shall be loosed in heaven." -- kjv10.txt | 03:17 |
|
| sorry. | 03:17 |
|
| misfire. i was giving a ref about the "loose" "lose" misspelling being biblical | 03:18 |
| → Crimea joined | 03:27 |
| ← nkallen left | 03:30 |
| ← bugfux left | 03:35 |
| ← kukks left | 03:41 |
| ← consolers left | 03:47 |
| → saintdev joined | 03:55 |
| ← mtodd left | 04:03 |
| → das-q joined | 04:06 |
| → jones- joined | 04:09 |
| ← threeve left | 04:11 |
| G_SabinoMullane → Z_SabinoMullane | 04:13 |
| → mtodd joined | 04:19 |
| → mtodd__ joined | 04:24 |
| → rubydiamond joined | 04:25 |
| ← tjafk1 left | 04:27 |
| → tjafk1 joined | 04:27 |
| ← CodeOfficer left | 04:30 |
| ← kanru left | 04:33 |
| → mtodd___ joined | 04:38 |
| ← mtodd__ left | 04:40 |
| → ckoehler joined | 04:41 |
| → janm joined | 04:43 |
| ← mtodd left | 04:51 |
| ← lirit left | 04:52 |
| → cmarcelo joined | 04:53 |
|
tflsh
| i dont see the misspelling | 04:57 |
|
| loose is the opposite of bind | 04:57 |
|
| lose isnt | 04:57 |
| ← Crimea left | 05:03 |
| ← jones- left | 05:03 |
| → Debolaz joined | 05:08 |
|
Debolaz
| Hmm.. How would I go about incrementally importing a cvs repository that git-cvsimport can't grok correctly? | 05:09 |
|
| Right now I'm using tailor, but tailor can't figure out branches automatically. | 05:09 |
| ← djwonk left | 05:15 |
| → Crimea joined | 05:18 |
| ← madewokherd left | 05:24 |
| ← bobesponja left | 05:26 |
| ← rubydiamond left | 05:26 |
| → bwwww joined | 05:27 |
| ← mtodd___ left | 05:29 |
| ← ckoehler left | 05:33 |
| → etnt` joined | 05:40 |
| ← Z_SabinoMullane left | 05:43 |
| ← Lucine2 left | 05:46 |
| → Lucine2 joined | 05:50 |
| → mtodd joined | 05:51 |
| ← bwwww left | 05:54 |
| → nkallen joined | 05:55 |
| → maveriick joined | 06:01 |
|
maveriick
| hi erveybody | 06:01 |
|
| I experienced a problem during git commit due to power outage....I would if anyobdy can hint on any help.... | 06:02 |
|
| I am issuing a git commit ..... writing the commit msg in the editor...in that time power outage occured.... | 06:02 |
|
| now when I give the command git status it says: fatal: index file smaller than expected | 06:02 |
| → sgrimm joined | 06:04 |
| ← sgrimm left | 06:05 |
| → sgrimm joined | 06:05 |
|
maveriick
| Anybody faced similar situation? | 06:12 |
| ← csc` left | 06:18 |
| → csc` joined | 06:20 |
| ← nkallen left | 06:37 |
| ← asdx left | 06:52 |
| ← etnt` left | 07:02 |
|
FunkeeMonk
| maveriick: yikes, that sounds scary. Did you manage to recover from the problem yet? | 07:04 |
| ← stni left | 07:06 |
| → stni joined | 07:07 |
| ← cmarcelo left | 07:07 |
| → bwwww joined | 07:13 |
| → rvo joined | 07:14 |
|
thiago_home
| maveriick: it looks like your repository files got corrupted | 07:17 |
|
| try checkout -f HEAD | 07:17 |
| ← rvo left | 07:21 |
| → robinr joined | 07:23 |
| ← brosner left | 07:35 |
| → rubydiamond joined | 07:41 |
|
maveriick
| thiago: i am just back after recovering.... | 07:50 |
|
| thiago_home: what I did is ...sit in a cowrokers PC then add myself as remote....and then fetch my repo from there.... | 07:50 |
|
| thiago_home, it looks clean there.... | 07:51 |
|
| thiago_home: Donno how it works ....but i am happy that it worked ......it consists labor of 3 days in that commit... | 07:51 |
| → anttih joined | 07:52 |
|
maveriick
| brb .....(logging in from another PC) | 07:52 |
| ← maveriick left | 07:52 |
| ← lolage0 left | 07:55 |
| ← jerbear left | 08:00 |
| ← csc` left | 08:00 |
| → devogon joined | 08:05 |
| → lirit joined | 08:06 |
| → priidu joined | 08:06 |
| ← priidu left | 08:07 |
| → priidu joined | 08:11 |
| → ph^ joined | 08:11 |
| ← lirit left | 08:12 |
| ← priidu left | 08:14 |
| → rubydiam_ joined | 08:15 |
| ← rubydiamond left | 08:15 |
| ← bwwww left | 08:16 |
| → priidu joined | 08:17 |
| → deavid joined | 08:21 |
| → Beket joined | 08:26 |
| → d0k joined | 08:34 |
| → hein2 joined | 08:35 |
| ← Sho_ left | 08:35 |
| hein2 → Sho_ | 08:35 |
| ← igorgue left | 08:37 |
| → holizz joined | 08:45 |
| → mjponce joined | 08:51 |
| ← mjponce left | 08:51 |
| → gitte joined | 08:55 |
| → _graham_ joined | 09:00 |
| → jones- joined | 09:02 |
| → scorp007 joined | 09:04 |
|
scorp007
| what is the typical way of releasing a tagged version of the repo? | 09:05 |
|
jast
| depends on the way you want to release it | 09:06 |
|
| one way is to tell people the name of the tag :) | 09:06 |
|
scorp007
| what if I wanted to archive a stable version? | 09:06 |
|
| say, by zipping it up. | 09:07 |
|
jast
| try git archive | 09:08 |
|
scorp007
| ah ok | 09:08 |
|
jast
| you can even give it a tag name iirc | 09:08 |
|
scorp007
| ah, you're right. | 09:09 |
|
jast
| seems to output to stdout | 09:09 |
|
scorp007
| thanks | 09:09 |
|
| there's an example that zips it. | 09:09 |
|
Pieter
| git archive HEAD | gzip > project.tar.gz | 09:10 |
|
| that's how most people use it | 09:10 |
|
scorp007
| yep, that's similar to the example | 09:10 |
| ← mfwitten left | 09:11 |
|
jast
| and to think i wrote my own script to do that some time ago... :C | 09:13 |
|
Pieter
| §9§9 | 09:13 |
|
| oops | 09:13 |
|
scorp007
| what's the pax_global_header files it creates? | 09:15 |
|
| file * | 09:15 |
|
jast
| a tar archive | 09:18 |
|
| (uncompressed) | 09:18 |
|
| use --format=zip for, well, zip | 09:18 |
|
scorp007
| oh ok | 09:20 |
| rubydiam_ → rubydiamond | 09:22 |
| → cypher23 joined | 09:30 |
| → RoboTeddy joined | 09:33 |
| ← RoboTeddy left | 09:33 |
| → consolers joined | 09:34 |
|
consolers
| how to deal with git-cvsimport: request for non-existent rev 1.1 in file XXX | 09:34 |
|
| (its a new file added to the repository) | 09:35 |
| → chris2 joined | 09:36 |
|
consolers
| (its a new file added to the repository in a new branch) | 09:38 |
| → goibhniu joined | 09:42 |
| ← Sho_ left | 09:43 |
| → Sho_ joined | 09:46 |
| ← DrNick left | 09:48 |
| ← dwave- left | 09:51 |
| ← Husio left | 09:51 |
| → DrNick joined | 09:54 |
| ← DV|HearthCore left | 09:55 |
| → DV|HearthCore joined | 09:55 |
| ← DV|HearthCore left | 09:55 |
| → _DV|HearthCore joined | 09:56 |
| ← chris2 left | 09:57 |
| → Yuuhi joined | 09:59 |
| ← gitte left | 10:00 |
| ← MarcWeber left | 10:02 |
| → felipec joined | 10:06 |
|
consolers
| Why does git even choose to use cvsps | 10:15 |
|
| sheesh | 10:15 |
| ← priidu left | 10:19 |
| → Xyri0 joined | 10:23 |
| → jmspeex joined | 10:27 |
|
jast
| well, you need *some* way of reading stuff from the repo | 10:28 |
|
Beket
| Hi people. Is this mem leak a false alarm http://rafb.net/p/tXlCyR86.html ? | 10:30 |
|
jast
| it doesn't actually run the script, does it? | 10:31 |
|
| also, git-cvsimport is a perl script and perl has a garbage collector | 10:31 |
|
Beket
| mm, thanks jast | 10:32 |
|
jast
| try the advice from the last line if you want to see what exactly causes the leak | 10:33 |
|
| (i haven't used valgrind yet so i can't give any specific advice) | 10:33 |
|
AlexB
| When you do, don't remove anything called MD_Update ;-) | 10:35 |
|
jengelh
| only debian users would do that | 10:35 |
|
jast
| bashing debian never gets old, does it :D | 10:35 |
|
jengelh
| debian itself is old | 10:35 |
|
jast
| not after i spent four hours updating ssh keys everywhere | 10:35 |
| ← Xyrio left | 10:36 |
| Xyri0 → Xyrio | 10:36 |
| ← _DV|HearthCore left | 10:37 |
|
consolers
| I've no idea how i'm going to fix this. turning on debugging in cvsps gives gobs | 10:39 |
| → dirk1296 joined | 10:41 |
|
consolers
| the files are under Attic/ | 10:43 |
|
| branch files go under Attic/ ? | 10:43 |
|
jast
| normally, "old" files go under Attic/, i.e. deleted files (iirc) | 10:48 |
| → holizz_ joined | 10:49 |
|
scorp007
| errm, how do I undo a change in my working copy to what is in the HEAD? | 10:50 |
|
| on 1 file | 10:50 |
|
jast
| git checkout HEAD -- filename | 10:50 |
|
| if it's still untouched in the index, you can omit the HEAD | 10:50 |
|
scorp007
| ah, cool | 10:50 |
|
| thanks | 10:50 |
|
| how do I verify a repo is non-corrupt, if I may have accidently interrupted a push? | 10:53 |
|
consolers
| jast, yet when i checkout the branch with -r, the files are correct. theyre in Attic/ though | 10:53 |
|
| cvsps fails horribly because it cant find the files | 10:54 |
|
| this is on the CVSROOT repo | 10:55 |
| → shaftyy joined | 10:56 |
|
consolers
| so i am to infer that new files in a new branch go to Attic/ | 10:57 |
| ← rubydiamond left | 11:01 |
| ← holizz left | 11:04 |
| → PferdOne joined | 11:06 |
|
consolers
| (its documented in the CVS manual under Attic/) | 11:11 |
| → priidu joined | 11:12 |
|
Ilari
| scorp007: 'git fsck --full' on remote end. Be aware that lots of 'dangling' messages are to be expected... | 11:12 |
|
scorp007
| Ilari, ok cool | 11:14 |
|
| so those danglings messages are safe? | 11:15 |
| ← ferdy left | 11:15 |
|
scorp007
| wierd, no output. | 11:15 |
| → ferdy joined | 11:15 |
|
Ilari
| scorp007: If that repo does not have working tree, and has never had branch rewound, then no output is expected... In any case, no output means it passed the checks. | 11:16 |
|
scorp007
| oh ok | 11:16 |
|
| it seems to be missing a tag though.. | 11:17 |
|
| ah, nevermind. | 11:17 |
| ← shermanl left | 11:18 |
|
Ilari
| scorp007: For pushes over ssh:// (and git:// if repository owner doesn't care about security), updating the remote repository is done by a program running on remote host, not by direct writing. | 11:19 |
|
| scorp007: For HTTP WebDAV push (yuck), the files are written directly, but it is possible to write files in correct order (packfile, packfile index, packfile list, refs, refs list), so the repository stays valid at any point during the update... | 11:26 |
| ← sdiz left | 11:29 |
| → Judofyr joined | 11:30 |
| ← GarethAdams left | 11:32 |
| ← priidu left | 11:32 |
| → GarethAdams joined | 11:33 |
| ← Fullmoon left | 11:34 |
| → robin joined | 11:42 |
| → Fullmoon joined | 11:43 |
|
jast
| does http-push really write pack files? | 11:47 |
| → bentob0x joined | 11:49 |
| → ebel joined | 11:52 |
| → G_SabinoMullane joined | 11:52 |
| ← robinr left | 11:54 |
|
Ilari
| jast: Even if it writes loose objects, the operations can be still be done in correct order not to corrupt the repo... | 11:55 |
| → BabelO joined | 11:59 |
| PferdOne → PferdOne`TV | 12:01 |
| → anttih__ joined | 12:03 |
| ← Beket left | 12:04 |
|
jast
| yeah, except if it writes a partial object | 12:04 |
|
Mikachu
| it doesn't write to a temp file and rename? | 12:06 |
|
consolers
| anyone has hacked cvsps before? | 12:06 |
|
jast
| i don't know. i haven't looked at http-push. i'm not sure i want to. | 12:07 |
|
| i know even less about cvs internals and, by extension, about cvsps | 12:09 |
| → chris2 joined | 12:11 |
| ← dirk1296 left | 12:13 |
| → perg joined | 12:14 |
| ← pergesu left | 12:14 |
| → rubydiamond joined | 12:15 |
| ← cypher23 left | 12:17 |
|
consolers
| from reading the comments in cvsps.c:412 assign_pre_revision (33357c69) it looks like I'm fucked | 12:18 |
| → hatsffilc joined | 12:19 |
| ← anttih left | 12:20 |
|
consolers
| the problem isnt that it is looking for a file which is in Attic, it is looking for wrong revision | 12:20 |
| ← _graham_ left | 12:21 |
| → _graham_ joined | 12:21 |
| ← d0k left | 12:21 |
| ← devogon left | 12:21 |
| ← mithro left | 12:21 |
| ← niki left | 12:21 |
| ← johan-s_ left | 12:21 |
| ← statim left | 12:21 |
| ← jdq left | 12:21 |
| ← Pistahh left | 12:21 |
| ← albertito left | 12:21 |
| ← EvanCarroll left | 12:21 |
| ← majoh left | 12:21 |
| ← bartman left | 12:21 |
| ← noz left | 12:21 |
| ← shawn__ left | 12:21 |
| ← Mutru_ left | 12:21 |
| ← kergoth left | 12:21 |
| ← mojombo left | 12:21 |
| ← bryanl left | 12:21 |
| ← benp- left | 12:21 |
| ← Illocution left | 12:21 |
| ← Thumper_ left | 12:21 |
| ← robbyonrails left | 12:21 |
| ← henr_k left | 12:21 |
| ← frsk left | 12:21 |
| ← Ted left | 12:21 |
| ← Tv left | 12:21 |
| ← CIA-53 left | 12:21 |
| hatsffilc → cliffstah | 12:21 |
| → Ted joined | 12:23 |
| → d0k joined | 12:23 |
| → devogon joined | 12:23 |
| → mithro joined | 12:23 |
| → niki joined | 12:23 |
| → johan-s_ joined | 12:23 |
| → statim joined | 12:23 |
| → EvanCarroll joined | 12:23 |
| → jdq joined | 12:23 |
| → Illocution joined | 12:23 |
| → Mutru_ joined | 12:23 |
| → CIA-53 joined | 12:23 |
| → benp- joined | 12:23 |
| → albertito joined | 12:23 |
| → Tv joined | 12:23 |
| → noz joined | 12:23 |
| → bryanl joined | 12:23 |
| → frsk joined | 12:23 |
| → henr_k joined | 12:23 |
| → bartman joined | 12:23 |
| → shawn__ joined | 12:23 |
| → Thumper_ joined | 12:23 |
| → majoh joined | 12:23 |
| → kergoth joined | 12:23 |
| → robbyonrails joined | 12:23 |
| → Pistahh joined | 12:23 |
| → mojombo joined | 12:23 |
| ← noz left | 12:25 |
| ← felipec left | 12:25 |
|
consolers
| wow this may be fixed in cvsps 2.2b1 which git repo is unaware of | 12:26 |
|
Mikachu
| how many git repos and home pages does that program have anyway? | 12:31 |
| → dirk1296 joined | 12:31 |
|
consolers
| I counted 2 git repos and 3-4 locations for the sources :) | 12:31 |
|
Mikachu
| heh | 12:32 |
|
consolers
| problem is on upadate, i'll bet it'll change the patchsets it finds, thereby giving ALL commits new shas. | 12:33 |
|
| it is _impossible_ to duplicate two imports | 12:33 |
|
| to give the same git repo | 12:33 |
|
Mikachu
| update compares timestamps i think though? | 12:33 |
|
consolers
| cvsps does, yes | 12:34 |
|
| but if i had wanted to rebuid the git repo and get consistent shas from the CVS repo from scratch, i'd have to keep the cvsps binary and snapshots of aCVSROOT | 12:35 |
|
| heh | 12:35 |
| ← PferdOne`TV left | 12:37 |
|
Mikachu
| what's the stuff in ~/.cvsps for? | 12:39 |
|
dirk1296
| does anyone here have experience with "git filter-branch --subdirectory-filter" under msysGit? | 12:51 |
| → kukks joined | 12:53 |
|
consolers
| cache | 12:55 |
|
| I'm facing a problem after merging: | 12:55 |
|
| I fix the conflict and do a commit, i get error: Error building trees | 12:56 |
|
| that means i'm screwed/ | 12:58 |
|
| ? | 12:58 |
| → kixx joined | 12:59 |
|
jast
| it sure sounds so. tried git fsck? | 12:59 |
|
consolers
| i'm in the middle of a merge | 12:59 |
|
| let me try anyway | 12:59 |
|
| one dangling blob | 13:00 |
|
| which on cat-file proves to be the last conflicting file (with conflicts in it) | 13:01 |
|
jast
| that doesn't explain the error message | 13:04 |
|
| dangling blobs are pretty normal, after all | 13:05 |
|
| perhaps try the commit with GIT_TRACE=1 for more details | 13:05 |
|
consolers
| bah. anyway 2.2b1 doesnt fix my problem | 13:06 |
| → felipec joined | 13:10 |
| ← perg left | 13:14 |
| ← ec left | 13:15 |
|
consolers
| nope. trace: built-in: git 'commit | 13:19 |
|
| easy enough to find it in builtin-commit.c as an error in cache_tree_update | 13:20 |
|
| git-diff is non-nil | 13:23 |
|
| how do i cancel the merge? | 13:23 |
|
| is there a canned command? | 13:23 |
|
chris2
| i'm making an tar with git-archive, but i'd like to add additional files, is there a nice trick? | 13:29 |
|
jast
| you can use tar to append files to an existing archive | 13:30 |
|
chris2
| i could "tar r" them, but i want a directory prefix | 13:30 |
|
jast
| i.e. pipe the stuff through another instance of tar before gzipping it (or whatever you do) | 13:30 |
|
chris2
| and i dont know how to make tar add to a subdirectory :/ | 13:30 |
| ← statim left | 13:32 |
|
robin
| chris2: there is a --transform option | 13:33 |
|
chris2
| not in my version on osx here | 13:33 |
|
| but that's what i would need | 13:33 |
|
robin
| ah, cripplix | 13:34 |
|
chris2
| it's a gnutar | 13:34 |
|
consolers
| you can use -r with a list of files to append to a tar archive, or create a diff.tar contaning the new files and and use -A to cat | 13:34 |
|
chris2
| but oldish | 13:34 |
|
| that doesnt allow me to put them into a subdirectory | 13:34 |
|
| will try using pax | 13:35 |
|
consolers
| i dont understand. you create the diff.tar with the subdir structure you want | 13:35 |
|
| or you use -r subdir/filename | 13:35 |
|
chris2
| i want to avoid making the prefix directory in my fs | 13:35 |
|
consolers
| i dont understand | 13:36 |
|
chris2
| if tar simply had a --prefix option it would be fine | 13:36 |
|
| i use git-describe as prefix | 13:36 |
|
| for making releases | 13:36 |
| → koke joined | 13:37 |
|
consolers
| how would a --prefix option work with tar? | 13:37 |
|
chris2
| tar --prefix pre -r filename => tar -r pre/filename | 13:38 |
|
jast
| one thing you could do, though it's ugly, is create a temporary symlink to the dir and pass -h to tar | 13:38 |
|
chris2
| where pre only exists in the archive | 13:38 |
|
| i'll use pax | 13:38 |
|
consolers
| so just temporarily mkdir pre; mv filename pre; tar -r pre/filename | 13:38 |
|
| whats your problem? | 13:38 |
| ← holizz_ left | 13:39 |
|
chris2
| it makes the script more complicated :) | 13:39 |
|
jast
| that's what scripts are for ;) | 13:39 |
|
consolers
| youre going to rm -rf pre anyway | 13:39 |
|
| heh | 13:39 |
|
chris2
| i dont have to if it doesnt exist :) | 13:39 |
|
consolers
| I still have the setup where my git-commit is failing if someone wants to help me debug the cache issue | 13:40 |
|
jast
| consolers, as i said, try with GIT_TRACE=1 to get more information that's possibly more suitable for debugging | 13:41 |
|
consolers
| no, TRACE=1 just shows that builtin-commit is being used | 13:41 |
|
jast
| ah | 13:41 |
|
consolers
| the problem is that `cache_tree_update' fails and that can fail for 2 reasons | 13:42 |
| → charon joined | 13:42 |
|
consolers
| but if someone with git-fu could spot it at a glance .... | 13:42 |
|
jast
| to cancel merges, by the way: git reset --hard (the usual warnings apply) | 13:43 |
| → edward joined | 13:43 |
| → madewokherd joined | 13:43 |
| → imyousuf joined | 13:44 |
| → bdiego joined | 13:45 |
|
jast
| will probably fix the problem too | 13:45 |
|
edward
| anybody using git to store a collection of photographs? | 13:45 |
|
| edward is wondering if there is any best practice | 13:45 |
|
imyousuf
| I want to see all the commits in which a file (object) has been changed, what would be the easiest way to do it? | 13:46 |
| → priidu joined | 13:46 |
| → cobrapg joined | 13:49 |
|
consolers
| hhmm | 13:49 |
|
| git-log --name-only --follow file.c gives the full log | 13:49 |
|
| bah | 13:49 |
|
| i mean it doesnt give "name only" | 13:49 |
|
imyousuf
| I see | 13:50 |
|
| let me check | 13:51 |
|
jast
| edward, what makes it useful to use git for that? | 13:51 |
|
consolers
| thats of course because i misread the documentation for --name-only | 13:53 |
|
jast
| imyousuf, --pretty=oneline may be helpful too to get a better overview. you can even leave out --name-only then because you'll only get the id and subject | 13:53 |
|
imyousuf
| consolers: thanks git-log --follow file.c this worked | 13:53 |
|
| I was also reading man of git-log | 13:53 |
|
| jast tring the --pretty option | 13:54 |
|
cobrapg
| I have a problem with Git... is asking for help allowed here? | 13:54 |
|
jast
| cobrapg, definitely, yes | 13:54 |
|
imyousuf
| :) yeah cobrapg | 13:54 |
|
cobrapg
| cool! | 13:54 |
|
imyousuf
| jast thanks, that shortcut helps as well :-D | 13:55 |
|
jast
| one of the main points of this channel, really | 13:55 |
|
cobrapg
| I am trying to do a rebase with Git and I get a SSL not supported error | 13:55 |
|
jast
| git-svn rebase, i assume? | 13:55 |
|
cobrapg
| yes | 13:55 |
| → Illocuti1n joined | 13:55 |
|
jast
| did you initially clone the repo from the same system? | 13:56 |
|
edward
| jast: so i can track changes | 13:56 |
|
jast
| *on* the same system, i mean | 13:56 |
|
cobrapg
| I get the same error | 13:57 |
|
imyousuf
| cobrapg: One policy I follow in git svn development is I keep by SVN branch clean and work in another branch | 13:58 |
|
jast
| edward, well. most photo collections i know don't change, but makes sense. what sort of best practices do you mean? | 13:58 |
|
cobrapg
| This is the error I get | 13:58 |
|
| RA layer failed to init socket layer: SSL is not supported at /usr/bin/git-svn line 1359 | 13:58 |
|
imyousuf
| so a simple svn fetch, git merge, git dcommit, git rebase is fine | 13:58 |
|
| and did not get any error | 13:58 |
|
jast
| no, that's not a git-specific error | 13:58 |
|
| it's either ssl support missing on your system | 13:58 |
|
imyousuf
| I develop projects in google code | 13:58 |
|
jast
| or on the other side, if it's seriously broken... can't tell which | 13:58 |
|
cobrapg
| hmmmm | 13:59 |
|
edward
| jast: just wondering if i should like jpegs unchanged, store a file with a list of changes, like rotate and crop. i just wondered if anybody here was doing that kind of thing | 13:59 |
| ← anttih__ left | 13:59 |
|
jast
| i can't figure out right now what ssl lib is used | 13:59 |
|
cobrapg
| jast: what is interesting is that I can do a svn checkout from a https enabled repo | 14:00 |
|
jast
| that's weird | 14:00 |
| → lirit joined | 14:00 |
|
jast
| i dunno how git-svn works there. i'll try to check in a couple of minutes | 14:00 |
| ← Illocuti1n left | 14:00 |
|
cobrapg
| jast: OK | 14:00 |
| ← priidu left | 14:00 |
|
cobrapg
| jast: I thought git-svn did an internal pipe, but it seems to be done differently | 14:01 |
|
| if I do a svn --version | 14:01 |
|
jast
| edward, problem is jpegs don't delta-compress well (because they're already compressed). therefore keeping several versions of your pictures will consume approximately as much space in the repo as all versions of all files. | 14:01 |
|
| cobrapg, it uses the perl cpan module SVN (or something like that), but also perl's libwww directly. dunno how that one works. | 14:02 |
|
| (i dunno how either works, really. i *think* perl's svn module uses libsvn) | 14:02 |
|
cobrapg
| yes | 14:02 |
|
| I have all the perl bindings | 14:02 |
|
edward
| jast: thanks | 14:02 |
|
cobrapg
| cause I can checkout with git-svn as long as the repo is not SSL enabled | 14:03 |
|
jast
| libsvn uses neon for webdav (including ssl) | 14:07 |
| ← scorp007 left | 14:07 |
|
jast
| what os/distribution are you using? | 14:07 |
|
cobrapg
| CentOS 4.5 | 14:07 |
| → priidu joined | 14:07 |
|
cobrapg
| so... is there a way to check if neon has SSL enabled in it? | 14:08 |
|
jast
| i think it depends on libssl | 14:08 |
|
| problem is i don't know what all these packages are called in fedora-like systems | 14:08 |
|
cobrapg
| you mean openssl? | 14:08 |
|
jast
| libneon in this case | 14:09 |
|
| in ubuntu it's helpfully called libneon27 | 14:09 |
|
cobrapg
| hmmmm | 14:09 |
| ← Illocution left | 14:10 |
|
jast
| if you know your way around rpm you can probably find out | 14:12 |
|
cobrapg
| right | 14:12 |
|
| now... | 14:12 |
|
| when compiling from source... is ssl an option for neon? | 14:13 |
| ← shaftyy left | 14:13 |
|
jast
| i'm pretty sure it is | 14:13 |
|
| in debian it's a hard dependency, so there's no way around it there | 14:14 |
| → shaftyy joined | 14:14 |
|
cobrapg
| OK | 14:14 |
| ← shaftyy left | 14:15 |
|
jast
| does plain svn work with ssl? because if it does, you can be sure neon has ssl support | 14:15 |
|
cobrapg
| yes | 14:15 |
| → lydgate joined | 14:15 |
|
cobrapg
| it does | 14:15 |
|
jast
| oh, you already said :{ | 14:15 |
| ← koke left | 14:15 |
|
cobrapg
| so... that means neon has SSL then? | 14:16 |
| → athought joined | 14:16 |
|
jast
| i'm pretty sure, yes | 14:16 |
|
| what version of git are you using, by the way? | 14:17 |
|
| oh, sorry, i gotta go for about an hour. | 14:17 |
| → a-priori joined | 14:18 |
| ← cobrapg left | 14:19 |
| ← d0k left | 14:19 |
| → shaftyy joined | 14:19 |
| → cobrapg joined | 14:20 |
|
cobrapg
| sorry I got disconnected | 14:20 |
|
| I am using git 1.5.5.1 | 14:20 |
| ← shaftyy left | 14:21 |
| ← felipec left | 14:22 |
| → koke joined | 14:23 |
| ← koke left | 14:23 |
| ← cobrapg left | 14:24 |
| → cedricv joined | 14:24 |
| ← imyousuf left | 14:25 |
| → csc` joined | 14:26 |
| → [Cobra] joined | 14:26 |
| → priidu_ joined | 14:27 |
| ← [Cobra] left | 14:27 |
| → shaftyy joined | 14:27 |
| ← athought left | 14:28 |
| ← priidu left | 14:28 |
| ← kixx left | 14:31 |
| ← shaftyy left | 14:33 |
| → djwonk_ joined | 14:34 |
| → fadec joined | 14:35 |
| ← priidu_ left | 14:37 |
| → felipec joined | 14:37 |
| → statim joined | 14:41 |
| → priidu joined | 14:43 |
|
chris2
| pax -waf tarball -s ':^:prefix/:' new files works | 14:43 |
|
| but not with the id-tagged tarballs, so you need to archive HEAD^{tree} | 14:44 |
| → cypher23 joined | 14:51 |
| → shaftyy joined | 14:54 |
| → kixx joined | 14:56 |
| ← priidu left | 14:56 |
| → advany joined | 14:59 |
| → bobesponja joined | 15:00 |
| → priidu joined | 15:02 |
| → mtodd__ joined | 15:02 |
| ← priidu left | 15:03 |
| → priidu joined | 15:06 |
|
consolers
| Actually 2.1b1 will fix the problem. | 15:09 |
| → cmarcelo joined | 15:10 |
| ← priidu left | 15:12 |
| ← bdiego left | 15:14 |
| → priidu joined | 15:14 |
| → schacon_ joined | 15:15 |
| ← yann left | 15:18 |
| ← beeble left | 15:18 |
| ← anders_ left | 15:18 |
| ← xkr47 left | 15:18 |
| ← mtodd left | 15:18 |
| ← priidu left | 15:20 |
| → tflsh_ joined | 15:25 |
| ← tflsh left | 15:26 |
| → yann joined | 15:26 |
| → xkr47 joined | 15:26 |
| → beeble joined | 15:26 |
| → priidu joined | 15:26 |
| → anders_ joined | 15:26 |
| ← priidu left | 15:31 |
| ← lirit left | 15:32 |
| → tflsh joined | 15:32 |
| → zachinglis joined | 15:37 |
| → radarek joined | 15:38 |
| ← tflsh left | 15:44 |
| → tflsh joined | 15:45 |
| ← consolers left | 15:47 |
| ← tflsh_ left | 15:47 |
| ← Ramune left | 15:54 |
| → Illocution joined | 15:55 |
| ← zachinglis left | 15:58 |
| → nkallen joined | 15:59 |
| ← cypher23 left | 16:00 |
| → kanru joined | 16:00 |
| ← GarethAdams left | 16:01 |
| → zachinglis joined | 16:02 |
| → GarethAdams joined | 16:03 |
|
jp_tix
| if i have added a submodule to my project, the only thing needed to keep it up-to-date is `cd submodule/ && git pull && cd .. && git commit` ? | 16:03 |
|
jast
| it may be necessary to do a git add submodule (without / at the end) before the commit | 16:05 |
|
Ilari
| jp_tix: Yea. No, '/' at the end, or the result will be a mess. Watch out, bash autocompletion puts that '/' there by default. | 16:08 |
|
jast
| as does zsh's, i guess | 16:08 |
|
Mikachu
| it does not | 16:10 |
|
Judofyr
| as does fish's... | 16:11 |
|
Debolaz
| Hmm.. How would I go about incrementally importing a cvs repository that git-cvsimport can't grok correctly and I don't have direct access to the ,v files for? | 16:12 |
|
jp_tix
| Ilari: for `cd dir` that shouldn't matter, no? | 16:15 |
|
Ilari
| jp_tix: Yes, it doesn't matter for 'cd'. It matters for 'git add'. | 16:15 |
|
jp_tix
| ah, didn't put that on my line | 16:16 |
|
| it seems that if i do git diff after the submodule pull, it references a different submodule commit | 16:16 |
|
| and that seems to be enough | 16:16 |
|
Ilari
| jp_tix: Yea. That's supposed to happen. | 16:17 |
|
jp_tix
| so what is the need for `git add` in this context? | 16:17 |
|
Ilari
| jp_tix: 'git add submodule'. | 16:17 |
|
jp_tix
| i need that after every submodule pull? | 16:18 |
|
Ilari
| jp_tix: Well, if you want to update the version recorded in supermodule. I think that 'git commit <submodule>' or 'git commit -i <submodule>' might also work (see 'git status <submodule>' and 'git status -i <submodule>'). | 16:19 |
|
jast
| doesn't commit simply commit the index? in that case, add is unavoidable (or commit -a, of course) | 16:21 |
| ← zachinglis left | 16:23 |
|
Ilari
| jast: Commit also accepts filesnames. Specify those, and its like 'reset HEAD', 'git add <filenames>' (include -i option to leave the reset out). | 16:24 |
|
jast
| oh, right. i haven't used that yet. | 16:24 |
|
| and the one time i tried i did something wrong and it didn't work | 16:25 |
|
Ilari
| jast: It is very rarely used... | 16:25 |
| ← robin left | 16:25 |
| ← kixx left | 16:25 |
| ← dirk1296 left | 16:25 |
| → bdiego joined | 16:26 |
| → robin joined | 16:27 |
| → consolers joined | 16:32 |
|
consolers
| the `Error building trees' I'm seeing on git-commit, is really a bad merge is it? | 16:33 |
|
Ilari
| consolers: Does 'git status' give anything unusual, like unmerged entries? | 16:38 |
| ← stni left | 16:39 |
| → holizz joined | 16:39 |
| → CodeOfficer joined | 16:40 |
|
consolers
| well git-commit reports unmerged entries before bombing out | 16:40 |
|
| I am comitting after resolving a conflict | 16:40 |
|
| i.e. in the middle of a merge | 16:41 |
|
| so git-status is a little out of date: it shows that file to be unmerged | 16:41 |
|
| and the rest: modified | 16:42 |
| → __dexter__ joined | 16:42 |
|
consolers
| hmm git-commit shows unmerged "hunks" i think in that one file | 16:42 |
| → flaguy joined | 16:42 |
|
jast
| did you call git add on that (now unconflicted) file? | 16:43 |
|
| consolers tests a hunch | 16:43 |
|
jast
| (after making sure there are no conflict markers in it, of course) | 16:43 |
|
| consolers kicks self | 16:44 |
|
consolers
| thanks jast | 16:45 |
|
jast
| sometimes it's the small things ;) | 16:45 |
| ← GarethAdams left | 16:48 |
|
consolers
| git-cvsimport needs to be updated for cvsps-2.1b1 in how branch ancestors are calculated | 16:50 |
|
Mikachu
| you mean 2.2b1 | 16:51 |
|
consolers
| indeed | 16:51 |
| ← Sho_ left | 16:53 |
| → Sho_ joined | 16:54 |
| ← bdiego left | 16:55 |
| ← Judofyr left | 16:56 |
| ← ebel left | 16:57 |
| ← advany left | 16:57 |
| → brosner joined | 16:58 |
| → pergesu joined | 17:02 |
| → SuttoL joined | 17:08 |
| ← holizz left | 17:08 |
| ← __dexter__ left | 17:08 |
| → etnt` joined | 17:09 |
| → manveru_ joined | 17:17 |
| → serfurj joined | 17:18 |
| ← manveru left | 17:19 |
|
serfurj
| I pushed changes to a repo, but forgot to check them out before making changes. How can I merge the changes to the working copy with the committed changes? For some reason "git merge master" says "Already up-to-date". | 17:20 |
|
Mikachu
| that's why you shouldn't push to a non-bare repo | 17:21 |
|
| do you know the sha1 the repo was at before the push? | 17:21 |
|
serfurj
| Mikachu: yes | 17:22 |
|
Mikachu
| i haven't tested this, but i think it should work | 17:22 |
|
| git reset --soft thatsha1 | 17:22 |
|
| git commit yourfiles | 17:22 |
|
| git rebase thenewtipyoupushed | 17:23 |
|
| does that make sense? | 17:23 |
|
| ie, the last sha1 is where the branch is now | 17:23 |
|
| so you want to go back to where it was, make the commit, then rebase it to the new code | 17:23 |
|
pasky
| i think that won't work | 17:23 |
|
Mikachu
| why? | 17:24 |
|
pasky
| since the commit will have thenewtipyoupushed as the parent | 17:24 |
|
| so rebase won't do anything | 17:24 |
|
| no? | 17:24 |
|
Mikachu
| you soft resetted to thatsha1 first | 17:24 |
|
pasky
| oh | 17:24 |
|
| right, that'll change your head | 17:24 |
|
serfurj
| Mikachu: I'll try that. Thanks | 17:25 |
|
Mikachu
| just make sure you don't type reset --hard ;) | 17:25 |
|
serfurj
| Mikachu: I'll test it out on a copy first | 17:26 |
|
Mikachu
| that's probably a good idea :) | 17:26 |
|
| if you do a git branch old first you can do git rebase old instead of copying that sha1 too, then just git branch -d old | 17:27 |
|
| (or a tag) | 17:27 |
|
serfurj
| what is thenewtipyoupushed? | 17:28 |
|
Mikachu
| that's the sha1 you're at before you did any of the steps i wrote | 17:29 |
|
| ie, the one you pushed | 17:29 |
|
serfurj
| oh ok | 17:29 |
|
Mikachu
| it should be HEAD@{1} or so probably :) | 17:29 |
|
| hm, or maybe 2 or 3 ;) | 17:30 |
| ← consolers left | 17:31 |
|
serfurj
| Mikachu: worked, thanks. | 17:36 |
| → d0k joined | 17:38 |
|
Mikachu
| nice, you're welcome :) | 17:42 |
| → drizzd joined | 17:43 |
| ← rubydiamond left | 17:51 |
| → koke joined | 17:52 |
| ← etnt` left | 17:56 |
| ← Crimea left | 17:57 |
| → Crimea joined | 17:59 |
| → ebel joined | 18:00 |
| → shermanl joined | 18:04 |
| → consolers joined | 18:05 |
| ← Crimea left | 18:05 |
| → tagx joined | 18:12 |
|
tagx
| does svn 1.5 break git-svn? | 18:12 |
|
| because i am getting segfaults with git-svn now | 18:12 |
| ← mithro left | 18:12 |
|
jengelh
| sh, don't use svn :p | 18:12 |
|
tagx
| i have to | 18:14 |
|
| but i want to use git on top of it | 18:14 |
|
| so it will be nicer | 18:14 |
|
jengelh
| it can't be nicer than the ugliest link | 18:14 |
|
Ilari
| tagx: In what component it segfaults in? Git-svn itself, libsvn, apr? | 18:15 |
|
tagx
| Ilari: how do i find out? | 18:15 |
|
Ilari
| tagx: No way to get stack backtrace? | 18:16 |
|
consolers
| Mikachu, so git-importsvn will have to be reworked completely I think, calling cvsps without -A ? | 18:17 |
|
tagx
| apr i tihnk | 18:17 |
|
Mikachu
| i have no idea how it works internally, sorry | 18:17 |
|
consolers
| I added a branch to parse the Branches: line and build a hashtable | 18:17 |
|
| ok | 18:17 |
| ← chris2 left | 18:17 |
|
consolers
| linus ancestor finding hack was just wrong anyway | 18:18 |
|
| at least on the repo of my interest: | 18:18 |
|
tagx
| Ilari: could bad interaction with svn 1.5rc5 be causing the segfault? | 18:19 |
|
Ilari
| tagx: With badly written software, doing one small mistake may cause segfault (even if it appeared to work in previous version)... | 18:21 |
|
Pieter
| hi | 18:21 |
| → _manveru joined | 18:24 |
| ← manveru_ left | 18:25 |
|
Debolaz
| tagx: Make sure the svn perl modules are updated on the client side. | 18:27 |
| ← schacon_ left | 18:28 |
|
Debolaz
| A perl script, which git-svn is, should *not* segfault, so the problem is likely to be in the subversion libraries themselves. | 18:28 |
|
jengelh
| well if you can make perl crash, that's scary | 18:28 |
|
tagx
| im downgrading my svn version now | 18:29 |
| → cypher23 joined | 18:29 |
|
Mikachu
| http://bugs.gentoo.org/show_bug.cgi?id=219203 | 18:30 |
| → schacon joined | 18:34 |
| ← schacon left | 18:37 |
| ← G_SabinoMullane left | 18:39 |
| ← pergesu left | 18:42 |
| → kixx joined | 18:49 |
|
tagx
| hmm, now im getting permission denied... | 18:49 |
|
Pieter
| that's better | 18:50 |
|
tagx
| when i try to create a new git-svn repo | 18:50 |
|
| but using an older one | 18:50 |
|
| it works now | 18:50 |
| ← cliffstah left | 18:52 |
| → cliffstah joined | 18:52 |
| ← cliffstah left | 18:52 |
| → cliffstah joined | 18:52 |
| ← cliffstah left | 18:52 |
| ← blackdaemon left | 18:59 |
| ← a-priori left | 19:00 |
| → cliffstah joined | 19:04 |
| ← consolers left | 19:05 |
|
biz
| I've pushed a commit from another machine where I didn't set my realname, is there a way to fix this afterwards? (The log on gitweb shows my login name now...) | 19:11 |
|
drizzd
| biz: is it the last commit? | 19:22 |
| → Crimea joined | 19:23 |
| → bdiego joined | 19:24 |
|
biz
| drizzd: I think doing a git-revert won't remove it from history, only a git-reset --hard HEAD^ && git-push -f origin? Since I'm actually the only active one on this project, there should be no problems with this, or would you suggest some other way? | 19:24 |
|
| And yes, it's the last commit. | 19:24 |
|
Mikachu
| just running git commit --amend should fix it | 19:25 |
|
Pieter
| you can do git reset --soft HEAD^ && git commit to commit again | 19:25 |
|
| I was wondering how ammend does that | 19:26 |
|
| so if you ammend someone else's commit, you'll be the author? | 19:26 |
|
| how about git rebase -i and edit? | 19:26 |
|
Mikachu
| not much point if it's the last commit | 19:26 |
|
Pieter
| yeah, but what does rebase -i do? | 19:26 |
|
| retain the original author? | 19:26 |
|
jengelh
| rebase retains author afair | 19:27 |
|
Mikachu
| both amend and rebase retains it | 19:27 |
|
jengelh
| as does cherry pick | 19:27 |
|
Mikachu
| you can git commit --amend --author 'whatever <a@b>' then | 19:27 |
|
| and -C HEAD to suppress the editor | 19:28 |
|
| but maybe reset --soft HEAD^ && git commit -c HEAD@{1} is easier then :) | 19:28 |
|
Pieter
| --amend is the easiest anyway | 19:29 |
|
Mikachu
| but you have to write the author manually then | 19:29 |
|
Pieter
| oh well | 19:29 |
|
| ;) | 19:29 |
|
| what if you set GIT_AUTHOR_NAME? | 19:29 |
|
Mikachu
| you still have to set it manually then | 19:30 |
|
Pieter
| ah | 19:30 |
|
| I was thinking about GIT_EDITOR=cat GIT_AUTHOR_NAME="Pieter" git rebase -i HEAD~50 :) | 19:30 |
|
Mikachu
| i mean setting that is manual :) | 19:30 |
|
Pieter
| oh :) | 19:30 |
|
biz
| git commit --amend yields: | 19:31 |
|
| ! [rejected] master -> master (non-fast forward) | 19:31 |
|
| (the push afterwards I'm talking about) | 19:31 |
|
Mikachu
| of course | 19:31 |
| ← nkallen left | 19:31 |
|
Mikachu
| but you should continue reading | 19:31 |
|
bentob0x
| what's the difference between git reset --soft HEAD and git reset HEAD? | 19:39 |
|
Mikachu
| soft doesn't change the index, while the other one does (--mixed) | 19:40 |
|
| in fact, git reset --soft HEAD is a no-op | 19:40 |
|
bentob0x
| git reset HEAD = git reset --mixed HEAD? | 19:40 |
|
Mikachu
| yes, --mixed is the default mode | 19:40 |
|
Pieter
| yes | 19:40 |
|
bentob0x
| k | 19:40 |
|
| (was in the man, didn't see it) | 19:41 |
|
Mikachu
| git reset --soft sha1 changes HEAD to sha1 (or rather, the branch HEAD points to when it isn't detached) | 19:41 |
|
| mixed does that, and updates the index to match that commit | 19:41 |
| → nkallen joined | 19:41 |
|
Mikachu
| --hard additionally updates the working tree | 19:41 |
|
bentob0x
| so I have tons of modifications and I started to add files for a commit but I want to restart it all while keeping the modifications, I have to do a git reset --soft HEAD and I'll keep the modifications untouched? | 19:41 |
| → advany joined | 19:42 |
|
Mikachu
| you want --mixed then | 19:42 |
|
| as i said, --soft HEAD is a no-op | 19:42 |
|
bentob0x
| I want to unstage everything basically but keep it all as it is | 19:42 |
|
Mikachu
| indeed | 19:43 |
|
| actually just "git reset" is sufficient | 19:43 |
| → larsl joined | 19:43 |
| → a-priori joined | 19:44 |
|
bentob0x
| I just did git reset HEAD | 19:44 |
|
| seems fine | 19:44 |
|
| git reset == git reset HEAD == git reset --mixed HEAD | 19:44 |
|
| thx guys | 19:44 |
|
Mikachu
| np | 19:44 |
|
felipec
| hmm, "echo foo | git commit" is trying to use my editor | 19:45 |
|
Mikachu
| who said git commit reads commit messages on stdin? | 19:46 |
|
biz
| Thanks Mikachu and Pieter | 19:46 |
|
felipec
| Mikachu: so how can I read a commit message from a file? | 19:46 |
|
bentob0x
| git commit -m "foo" felipec | 19:46 |
|
robin
| -F file | 19:46 |
|
| in tfm | 19:47 |
|
felipec
| robin: cool, can it be -F -? | 19:47 |
|
bentob0x
| be nice with people robin or I call batman | 19:47 |
|
| ^^ | 19:47 |
|
robin
| :) | 19:47 |
|
| felipec: reading man pages doesn't hurt | 19:48 |
|
| even git man pages sometimes contain useful information | 19:48 |
|
Mikachu
| even if it didn't have -, -F /dev/stdin usually works | 19:49 |
|
felipec
| robin: yeah, I just thought git would automatically ready the message from stdin if there was input there | 19:49 |
|
Mikachu
| felipec: that would mean it would read from the terminal instead of the started editor | 19:49 |
| ← SuttoL left | 19:49 |
|
Mikachu
| unless you do weird things like isatty() | 19:49 |
|
robin
| no sane programs does that | 19:53 |
|
jengelh
| actually, git-filter-branch reads it from stdin... in a way. | 19:53 |
|
robin
| sure programs read from stdin, but not by detecting if there in input on stdin or not. They always read from stdin or they don't, depending on optins | 19:54 |
| → jamessharpe joined | 20:00 |
| → priidu joined | 20:04 |
| ← d0k left | 20:05 |
| → shaftyy_ joined | 20:05 |
| ← shaftyy_ left | 20:06 |
| → shaftyy_ joined | 20:06 |
| ← shaftyy_ left | 20:06 |
| ← a-priori left | 20:09 |
| ← Crimea left | 20:10 |
| ← statim left | 20:11 |
| ← larsl left | 20:13 |
| ← koke left | 20:14 |
| ← shaftyy left | 20:20 |
| ← shermanl left | 20:22 |
| ← pygi left | 20:22 |
| → ec joined | 20:27 |
| ← cliffstah left | 20:27 |
| → jjlee joined | 20:31 |
|
jjlee
| after a git-svn clone, git-branch only lists "master", despite the clone operation showing multiple branches being fetched | 20:32 |
|
| gitk seems to show some branches but not others -- but I'm not used to reading gitk display | 20:33 |
| ← Modius_ left | 20:33 |
|
jjlee
| example: Found possible branch point: https://python-nose.googlecode.com/svn/trunk => https://python-nose.googlecode.com/svn/branches/160-jython, 436 | 20:34 |
|
| but I don't have a branch 160-jython, or similar | 20:34 |
|
| why not? | 20:34 |
|
Mikachu
| jjlee: git branch -a | 20:34 |
|
jjlee
| ah, I see - thanks | 20:35 |
|
| I still don't understand why, in gitk, only some branches are displayed in the "Branches:" list as hyperlinks | 20:36 |
|
| for example, in "Branches: master, remotes/0.10.2-stable, remotes/141-collect-only, remotes/tags/0.10.2-release, remotes/trunk" | 20:37 |
|
| "remotes/141-collect-only" is not a hyperlink, but the others are | 20:37 |
|
Mikachu
| probably because the commit it points to isn't shown in the graph | 20:38 |
|
jjlee
| why would it not be shown in the graph? | 20:38 |
|
Mikachu
| you didn't ask for it, try gitk --all | 20:39 |
|
jast
| gitk shows only the current branch by default | 20:40 |
|
| (iirc) | 20:40 |
|
| i rarely use gitk, truth be told | 20:40 |
|
jjlee
| yes, --all just discovered that :-) | 20:40 |
|
| jast: you prefer command line, or some better GUI tool? | 20:41 |
| → threeve joined | 20:41 |
|
jast
| i use command line most of the time | 20:42 |
|
| but the main reason for that is that i mainly use git for smallish projects and svn project (which don't have complicated branch structures) | 20:42 |
|
Mikachu
| i usually use qgit when i want to look at more complicated history, i get confused by gitk's nonstraight lines | 20:42 |
| ← janm left | 20:46 |
|
jjlee
| can anybody explain what "File->Update" and "File-Reread references" in gitk mean? | 20:47 |
|
| I'm guessing "references" here means the same thing as the "refs" in this diagram | 20:47 |
|
| http://utsl.gen.nz/talks/git-svn/git-model.png | 20:47 |
| → ironfroggy joined | 20:47 |
|
jjlee
| (though I don't really know what those "refs" are either ;-)) | 20:48 |
|
| specifically, I'm wondering what the difference is between "refs" and "heads" | 20:48 |
|
Mikachu
| nothing, "branch" is another word | 20:49 |
|
| i guess branch is a subset of refs though, tags are also refs | 20:49 |
| ← niki left | 20:49 |
|
jjlee
| that diagram lists "refs", "heads" and "tags" | 20:50 |
|
| all three | 20:50 |
|
Ilari
| jjlee: They update the display to reflect changes in refs. Not sure about the difference (it could be that one picks new refs, while the other doesn't)... | 20:50 |
|
Mikachu
| it describes refs and heads identically afaics | 20:50 |
|
| jjlee: if you commit something you use those to refresh the display | 20:51 |
|
jjlee
| ok, so "reread references" does not alter the working tree (as I expected) | 20:51 |
|
| does File-Update alter the working tree? | 20:51 |
|
Ilari
| jjlee: Nope. | 20:51 |
| ← bentob0x left | 20:52 |
|
jjlee
| what does it do? | 20:52 |
|
Ilari
| jjlee: It apparently performs full refresh of the drawn graph (picking up any possible new refs created, etc...) | 20:54 |
|
| jjlee: (effect pretty much similar to closing and restarting gitk with the same parameters). | 20:55 |
|
jjlee
| ok. I'm a coward :-) | 20:56 |
|
Mikachu
| i think the only thing gitk can do is cherry-pick? | 20:56 |
|
Ilari
| jjlee: If you want GUI that can actually do a lot to the working tree, use git-gui... | 20:56 |
|
| jjlee: Altough Git-gui AFAIK can't do rebase operations (those must be done from command line), but is otherwise quite capable... | 20:57 |
|
jjlee
| not especially - it's just nice to know what a button labelled "Update" will do before you click it | 20:58 |
| → G_SabinoMullane joined | 20:58 |
|
Ilari
| jjlee: At least it won't do anything destructive. :-) | 20:58 |
| ← jamessharpe left | 20:59 |
| ← threeve left | 20:59 |
|
jjlee
| hmm, I prefer the straight lines in qgit too - thanks for the pointer | 21:00 |
| ← ironfroggy left | 21:04 |
| → sbeyer joined | 21:06 |
|
jjlee
| though there seems no easy way in qgit to browse through patches, which is something I like about gitk | 21:06 |
|
Mikachu
| if you double click a commit it'll show the patch in a new tab | 21:07 |
|
jjlee
| not quite as easy as up/down arrow | 21:07 |
| ← priidu left | 21:09 |
| → angasule_ joined | 21:11 |
| ← angasule left | 21:11 |
| → statim joined | 21:18 |
| ← ph^ left | 21:22 |
| angasule_ → angasule | 21:24 |
| → ec_ joined | 21:27 |
| ← ec left | 21:29 |
| → ferdy_ joined | 21:32 |
| ← ferdy_ left | 21:32 |
| → ferdy_ joined | 21:32 |
| → marze joined | 21:32 |
| ← ferdy left | 21:32 |
| ← ferdy_ left | 21:32 |
| → ferdy joined | 21:32 |
| ← jroes left | 21:33 |
| → larsl joined | 21:34 |
|
ToxicFrog
| Hmm. How do I push tags? | 21:35 |
|
dato
| push --tags | 21:35 |
|
ToxicFrog
| That was simple. Thanks. | 21:35 |
|
| What if only want to push a specific tag? | 21:35 |
|
larsl
| If I have a CVS working directory containing a subdirectory coming from a separate repository, is there any way to get git-cvsimport to import the whole thing from both repositories as a single Git repository? | 21:35 |
|
vmiklos
| git push origin v1.0? | 21:35 |
|
larsl
| With history intact. | 21:36 |
|
ToxicFrog
| Aha! I was missing the "origin" | 21:36 |
|
| It's "git push <branch> tag <tagname>" | 21:36 |
|
vmiklos
| nope :) | 21:38 |
|
| git push remote tag v1.0 | 21:38 |
|
| <remote>* | 21:38 |
|
| origin is not a branch | 21:38 |
| ← Sonderblade left | 21:39 |
|
mugwump
| larsl: do them independently, then stitch together with git-filter-branch | 21:39 |
|
| you'll probably need to develop your own script for the stitching | 21:39 |
|
larsl
| mugwump: Ah, interesting - hadn't seen that before. | 21:39 |
|
mugwump
| you'll probably want to use a "tree filter", and in that tree filter, put the other project in as a submodule | 21:40 |
|
| get it working for just the first commit or two, and check that you can clone the result | 21:41 |
|
| git-filter-branch *might* need some minor changes to work with that - but it's shell so should be pretty straightforward. | 21:41 |
| → redondos joined | 21:43 |
|
larsl
| mugwump: Thanks, I'll have a look. | 21:43 |
| → kristoffer joined | 21:47 |
|
Mikachu
| ToxicFrog: it's just git push <remote> <refname>, no "tag" | 21:49 |
|
ToxicFrog
| Mikachu: what's the man page on about, then? | 21:50 |
|
Ilari
| ToxicFrog: If you specifically want to refer to tag (say, you got branch by the same name), then its 'git push remotename tags/tagname'. | 21:52 |
|
Mikachu
| oh, i didn't know you could say tag <tag>, hm | 21:55 |
|
jengelh
| kinda ugly though | 21:56 |
|
Mikachu
| what if you have two branches named "tag" and "test"? | 21:56 |
| ← kristoffer left | 21:57 |
| → kristoffer_ joined | 21:57 |
|
mugwump
| then you probably need to use the full syntax to specify them :) | 21:57 |
|
jengelh
| what syntax if there is none | 21:57 |
|
Mikachu
| who came up with that? | 21:57 |
|
jengelh
| oops | 21:57 |
|
mugwump
| you can use git push remotename refs/tags/foo | 21:57 |
|
| "tag foo" is just a shortcut | 21:57 |
|
Mikachu
| i guess you could jsut say git push foo tag instead, but you have to remember to put the "tag" last | 21:58 |
|
| no, that doesn't work either... | 21:59 |
|
| % git branch tag | 21:59 |
|
| % git push . tag master | 21:59 |
|
| error: src refspec refs/tags/master does not match any. | 21:59 |
|
| % git push . master tag | 21:59 |
|
| fatal: tag shorthand without <tag> | 21:59 |
|
| btw, isn't that error message missing a noun? | 21:59 |
|
mugwump
| it's missing a verb, yes | 21:59 |
|
| why not send a grammar patch | 21:59 |
|
| :) | 21:59 |
|
Mikachu
| how is it missing a verb? | 22:00 |
|
| it should have "existing refs." at the end i think | 22:00 |
|
mugwump
| the second one is missing a verb, the first one is a bit awkward | 22:00 |
| ← Yuuhi left | 22:00 |
|
Mikachu
| ah, the second is a fatal message, not an error message :P | 22:00 |
|
mugwump
| fatal is a class of error | 22:01 |
|
Mikachu
| yes, i was just defending my ambiguous referring to the first one earlier :) | 22:01 |
|
| mugwump puts away his Pedant's Own™ grammar book | 22:01 |
|
Mikachu
| it should really say "fatal error:" | 22:02 |
| ← marze left | 22:02 |
|
Mikachu
| lots of things can be fatal | 22:02 |
|
| so i guess what we learned is to not call branches "tag" | 22:02 |
|
| Debolaz ponders how much effort it'd take to write a new cvs to git converter with support for incremental updates. | 22:04 |
| _manveru → manveru | 22:06 |
|
larsl
| Debolaz: git-cvsimport does do incremental updates. | 22:06 |
| → schacon joined | 22:08 |
| ← programmerq left | 22:09 |
| ← drizzd left | 22:09 |
| ← CIA-53 left | 22:10 |
|
larsl
| But it's sloooooow. | 22:11 |
|
mugwump
| I think the problem with a lot of importers is that they muddle what should be separate phases of conversion: loading the information, mining out and converting history, and then exporting it | 22:12 |
| ← charon left | 22:12 |
|
mugwump
| and then people are left trying to paper over those problems by adding switches to try and change the behaviour of the combined thing, rather than just being able to add some mapping data to a database table and then re-exporting | 22:13 |
| ← ec_ left | 22:13 |
| ← schacon left | 22:16 |
|
Debolaz
| larsl: Slowness isn't an issue, it's that it just doesn't work on some repositories. Admittedly, it's hard to make it work right everywhere considering how fucked up CVS really is, so I'm not going to say cvsimport is crap, but it's still a problem for me when it doesn't work where I need it. | 22:16 |
| → CIA-2 joined | 22:17 |
|
jengelh
| well, it's just fubar :) | 22:17 |
| ← deavid left | 22:17 |
|
larsl
| Debolaz: Yeah, it doesn't work when the CVS repo owner has done weird things in CVSROOT/modules to create "virtual modules" for example. | 22:18 |
|
| larsl is trying to work around that at the moment. | 22:18 |
|
jengelh
| start afresh. | 22:19 |
|
larsl
| I really don't want to get that far deep in CVS. | 22:20 |
| ← redondos left | 22:23 |
|
Debolaz
| Luckily, I found that tailor was able to convert the cvs repository to git. But it cannot detect branches automatically, so you must explicitly specify them (With a separate configuration section even) or you only get the mainline. | 22:24 |
|
Ilari
| Doesn't CVS have per-file history (not per-project history like most VCSs do)? :-/ | 22:25 |
|
Debolaz
| Ilari: It has per-file history yes. | 22:25 |
| → redondos joined | 22:26 |
|
larsl
| That's why importing is such a pain. | 22:26 |
| ← johan-s left | 22:26 |
|
larsl
| git-cvsimport checks out every revision of every file in a repository. | 22:26 |
|
Ilari
| Trying to approximate it with per-project history could be quite nasty... | 22:27 |
|
| Add to that some bizarre tags that are sometimes encountered... | 22:27 |
|
Debolaz
| And sometimes, there are a few commits that it just seems to be unable to grok. | 22:28 |
|
| I suspect it might have something to do with character sets used, but that's really just guessing. | 22:28 |
| ← brosner left | 22:29 |
| ← ebel left | 22:31 |
|
larsl
| Hm, git filter-branch looks a bit too complicated for a one-night hack. | 22:33 |
|
| I wonder if it would be easier to hack cvsps to merge the rlogs from two CVS repositories. | 22:33 |
|
mugwump
| that's the problem with conversions isn't it. you think it shouldn't take long | 22:33 |
| → cjb joined | 22:34 |
|
cjb
| Hi! Is there any way to squash commits (using git-rebase --interactive) once I've already pushed them to a remote repo? | 22:34 |
|
| Or does modifying the commit history only work locally? | 22:34 |
|
jengelh
| it works everywhere, but people won't be happy if you change the timeline | 22:35 |
|
cjb
| I tend to get an error from the server when I push. Would it work with git push --force? | 22:35 |
|
mugwump
| perhaps. if that doesn't work you can usually push a delete, then push a "new" branch | 22:35 |
|
cjb
| I'm only merging several commits into one, rather than changing the timeline. | 22:35 |
|
mugwump
| that's counted as history rewriting | 22:36 |
|
| people who have already pulled will end up requiring a merge | 22:36 |
|
| but that's their problem ;) | 22:36 |
|
larsl
| Bah, screw the history. I'll just add the appropriate versions of the "submodule" in the heads of all branches. | 22:36 |
|
mugwump
| they should know instead to rebase | 22:36 |
| ← advany left | 22:36 |
|
cjb
| Yeah, I don't think anyone has pulled yet. | 22:36 |
|
mugwump
| larsl: appropriate, especially if the history is there in the submodule | 22:37 |
|
cjb
| But I'm a little worried about getting the repo inconsistent. | 22:37 |
| → schacon joined | 22:37 |
|
mugwump
| people can always come along later and cleanup the history, then use a graft to attach it to their history | 22:37 |
|
larsl
| mugwump: Yeah, and I need this set up tomorrow when GSoC starts. =) | 22:37 |
|
mugwump
| right | 22:38 |
|
felipec
| I can't specify the date in a git commit, right? | 22:38 |
|
vmiklos
| GIT_AUTHOR_DATE? :) | 22:39 |
|
| and GIT_COMMITTER_DATE | 22:40 |
|
felipec
| vmiklos: so I specify that on env? | 22:40 |
|
| cool | 22:41 |
|
mugwump
| see git-commit-tree(1) | 22:41 |
|
larsl
| cvsps doing an entire rlog for each incremental cvsimport is a bit annoying though. There should be a way to see what files have changed and just log those. | 22:41 |
| ← akibahara left | 22:43 |
|
cjb
| ok, tried and: | 22:44 |
|
| error: remote 'refs/heads/master' is not a strict subset of local ref 'refs/heads/master'. maybe you are not up-to-date and need to pull first? | 22:45 |
|
| Is it better to git push --force here, or something different? | 22:45 |
|
| it failed: | 22:47 |
|
cehteh
| guess why it suggests to pull first? | 22:47 |
|
cjb
| error: denying non-fast forward refs/heads/master (you should pull first) | 22:47 |
|
| ng refs/heads/master non-fast forward | 22:47 |
|
| cehteh: not because there are changes waiting for me; there aren't. | 22:47 |
|
cehteh
| no-one else cloned that repository yet? | 22:48 |
| → pygi joined | 22:48 |
|
cehteh
| its not shared? | 22:48 |
|
cjb
| I ran git pull moments before, and I can see from gitweb that there are no new commits. | 22:49 |
|
| but I'll see if pulling again helps, now that I've rewritten the history. | 22:49 |
|
cehteh
| well that wont work then | 22:49 |
| ← _graham_ left | 22:49 |
|
cehteh
| with rewriting history you destroy that old branch, you have to be make sure that no one uses that | 22:50 |
|
vmiklos
| felipec: yes | 22:50 |
|
cjb
| Ouch. Now I have both the original commits, and extra ones for my history rewriting. Suck. | 22:50 |
|
| cehteh: it's a tiny repository with few other committers. | 22:50 |
| → p_bclr joined | 22:51 |
|
cehteh
| well even with a few, it might break the others | 22:51 |
|
| cehteh suggest to switch to a new branch and inform the others to use that too | 22:52 |
|
cehteh
| when all acknowledged then you can roll over branches | 22:52 |
|
cjb
| I see. I guess I'll leave it alone, then, thanks. | 22:52 |
|
cehteh
| (delete the old, rename the new on to the old name) | 22:52 |
| → ckknight joined | 23:00 |
|
ckknight
| hey all, what would cause git log to hang? | 23:00 |
|
Ilari
| ckknight: Errorneous (cyclic) grafts data? | 23:01 |
|
felipec
| and the signers? | 23:01 |
|
ckknight
| I believe the data is correct, though | 23:01 |
|
| I'm calling git log in the update hook | 23:01 |
|
Ilari
| ckknight: Does the remote repo pass 'git fsck --full'? | 23:02 |
|
ckknight
| I'm not sure | 23:02 |
|
| how can I tell? | 23:02 |
| ← kristoffer_ left | 23:03 |
|
mugwump
| pasky: repo.or.cz[0: 2a01:b0:0:2::]: errno=No route to host | 23:04 |
|
| what's your v6 routing like? | 23:04 |
|
Ilari
| ckknight: Can you figure out what parameters the hook is trying to pass to log? Can you reproduce the hang? | 23:08 |
| ← p_bclr left | 23:08 |
|
ckknight
| Ilari: git log --name-only --pretty=format: | 23:09 |
|
| and it seems to reproduce each time a push is tried | 23:09 |
|
| basically, what I'm trying to do is use git log to get the list of all files created by the resultant push. it's hackish but it works. I'm open to another way to do it | 23:10 |
| ← cjb left | 23:11 |
|
jengelh
| for i in `git rev-list a..b`; do git log -p "$i"^! | lsdiff; done | sort -u; | 23:11 |
|
| where a..b is told you by the push operation | 23:12 |
| → d2dchat joined | 23:14 |
|
Ilari
| ckknight: For the net added files (so files added and then deleted won't show up): 'git diff-tree -r --name-only --diff-filter=A <old> <new>' (or something like that)... | 23:15 |
| → mithro joined | 23:15 |
|
d2dchat
| How can I import an existing git repo into my repo? | 23:16 |
|
| There will be no concflicts as far as files go | 23:16 |
| ← thiago_home left | 23:16 |
|
d2dchat
| or maybe it's a better idea to make a branch ? | 23:17 |
|
Ilari
| d2dchat: fetch it as new branch(es) (git fetch), or even merge the branches (git pull)? | 23:17 |
| → igorgue joined | 23:17 |
|
d2dchat
| Ilari: I can git pull within a repo? | 23:18 |
|
| from a remote repo? | 23:18 |
|
Ilari
| d2dchat: 'git pull <repo> <branch>' fetches and merges branch <branch> from repository <repo>. | 23:19 |
|
ckknight
| Ilari: what's <old> and <new>? | 23:19 |
|
Ilari
| ckknight: Revision numbers. IIRC, update hook gets them as parameters... | 23:19 |
|
ckknight
| Ilari: what if old is 40 0's? | 23:20 |
|
Mikachu
| then it's a new branch usually | 23:21 |
|
Ilari
| ckknight: For old version and new version. If old is 40 0s, it is branch creation ('git ls-tree -r --name-only -z <new>'). Also, maybe add -z to that diff-tree command... | 23:22 |
|
| ckknight: Also, new may be 40 0s (branch delete). | 23:22 |
|
Mikachu
| don't you just want git diff --name-only old new? | 23:25 |
|
| maybe i should read everything instead of just the first and last few lines ;) nwm | 23:26 |
|
d2dchat
| Ilari: I did a git fetch but i don't see the files | 23:27 |
|
| Ilari: what happened? | 23:27 |
| ← fujin left | 23:27 |
|
Ilari
| d2dchat: 'git fetch' won't merge. If you did 'git fetch <repo> <branch>', that branch wound up being called FETCH_HEAD. | 23:27 |
|
d2dchat
| Ilari: so how do I merge it where i want it? | 23:28 |
|
Ilari
| d2dchat: For fetch&merge, use 'pull', not 'fetch'. | 23:28 |
|
d2dchat
| Ilari: can I run a git pull now instead or do I need to do something with the assets git fetch may or may not have pulled down? | 23:29 |
| ← kixx left | 23:29 |
|
Ilari
| d2dchat: 'git merge FETCH_HEAD'? | 23:30 |
|
| d2dchat: That merges in what is stored in FETCH_HEAD. | 23:30 |
|
d2dchat
| Ilari: is there a way to merge it into a specific directory? | 23:31 |
| → fujin joined | 23:31 |
|
Ilari
| d2dchat: There's subtree merge in new enough versions (see git-merge(1)) | 23:32 |
|
d2dchat
| Ilari: perhaps my approach is all wrong | 23:32 |
|
| I'm used to subversion | 23:32 |
|
| where you have trunk, branches, etc. | 23:32 |
|
| is there a git way to just have separate trees? I think there is I just don't know how to do it | 23:33 |
|
Ilari
| d2dchat: Git does not use directories as branches. It has first-class branches. | 23:33 |
|
mugwump
| d2dchat: ok, first thing to do is to take all your SVN experience and forget it, it's worthless | 23:33 |
|
| d2dchat forgets svn experience | 23:33 |
| ← redondos left | 23:34 |
|
mugwump
| good | 23:34 |
|
d2dchat
| mugwump: no what? | 23:34 |
|
| now* | 23:34 |
|
ckknight
| d2dchat: what you're used to trunk being is now a branch called "master" | 23:34 |
|
mugwump
| now read the user manual ;) | 23:34 |
|
Ilari
| d2dchat: 'git branch <name-for-new-branch> FETCH_HEAD'. That would create new branch <name-for-new-branch> from the data you fetched... | 23:34 |
|
d2dchat
| hehe, is there a way to purge the remote stuff that I pulled? I kinda want to re-arrange everything now | 23:35 |
|
mugwump
| you don't need to do | 23:35 |
|
d2dchat
| ah so I don't need to purge that stuff | 23:35 |
|
| I can just put it into the new branch right away | 23:35 |
|
mugwump
| every branch is independent and does not affect the others | 23:35 |
|
Ilari
| d2dchat: And even the concept of 'branch' is different from SVN: One commit can be in multiple branches simultaneously, branches are not necressarily linear, etc... | 23:38 |
|
d2dchat
| noice | 23:39 |
|
| OK I just did a git branch public FETCH_HEAD | 23:40 |
|
| and a git status | 23:40 |
|
| and now all my files on master did something | 23:40 |
|
Ilari
| d2dchat: It should now show up in 'git branch'... | 23:41 |
|
d2dchat
| Ilari: do I have to git commit and push first? | 23:41 |
|
| ah I see it | 23:41 |
|
Ilari
| d2dchat: No need to commit after changing branches. Branches are first-class, not some brain-damaged SVN branches-as-directories stuf... | 23:42 |
|
d2dchat
| :) | 23:42 |
| ← schacon left | 23:43 |
| → ec joined | 23:43 |
|
d2dchat
| Ilari: now how do I switch to a branch? | 23:45 |
|
Ilari
| d2dchat: 'git checkout <branchname>'. | 23:46 |
| → ben_h joined | 23:46 |
|
d2dchat
| this is sweet lol | 23:47 |
|
| what is origin ? | 23:49 |
| ← robin left | 23:49 |
| → Ken69267 joined | 23:50 |
|
Ken69267
| is it possible to clone a branch? | 23:51 |
|
| rather than master | 23:51 |
|
Ilari
| Ken69267: Those should show up in 'git branch -r', but don't get local branches by default. | 23:52 |
|
d2dchat
| git push origin public.. nice | 23:52 |
|
| I love it | 23:52 |
|
| :) | 23:52 |
|
Ken69267
| Ilari: aha, thanks | 23:53 |
|
d2dchat
| Ilari: thx | 23:53 |
| → chjunior joined | 23:54 |
| ← jjlee left | 23:55 |
| ← bdiego left | 23:59 |
