| 2009-10-26 |
| → j416 joined | 00:01 |
| ← neunon left | 00:02 |
| → neunon joined | 00:02 |
| → fcuk112 joined | 00:03 |
| → bremner_ joined | 00:03 |
| → bremner__ joined | 00:07 |
| → drubies_ joined | 00:09 |
| ← Garen left | 00:09 |
| ← icwiener left | 00:10 |
| ← eno left | 00:10 |
| ← bremner__ left | 00:10 |
| ← javatexan left | 00:10 |
| ← flazz left | 00:11 |
| ← sgh left | 00:12 |
| → eno joined | 00:14 |
| ← fualo_ left | 00:16 |
| → ChenKaie joined | 00:20 |
| ← bremner_ left | 00:22 |
| → fualo_ joined | 00:23 |
| ← drubies left | 00:24 |
| ← babilen left | 00:28 |
| ← JasonWoof left | 00:28 |
| → JasonWoof1 joined | 00:28 |
| JasonWoof1 → JasonWoof | 00:29 |
| → hyperair joined | 00:31 |
| ← tbuehlmann left | 00:31 |
| ← Yuuhi left | 00:32 |
| ← mithro left | 00:33 |
| ← dennda left | 00:34 |
| ← ChenKaie_ left | 00:34 |
| ← jlilly left | 00:38 |
| → jlilly joined | 00:40 |
| → muthu joined | 00:40 |
| ← corni left | 00:42 |
| → Adlai` joined | 00:42 |
| ← onigiri left | 00:43 |
| ← krynnotaur left | 00:44 |
| ← Adlai left | 00:45 |
| Adlai` → Adlai | 00:45 |
| ← martianlobster left | 00:46 |
| → Garen joined | 00:48 |
| → javatexan joined | 00:49 |
| → rchavik joined | 00:50 |
| → davegb3 joined | 00:52 |
| ← drubies_ left | 00:53 |
| ← orafu left | 00:55 |
| → OuLouFu joined | 00:55 |
| OuLouFu → orafu | 00:55 |
| davegb3 → davegb3| | 00:58 |
| ← NET||abuse left | 00:59 |
| → NET||abuse joined | 00:59 |
| → drubies joined | 01:00 |
| → chrisr joined | 01:00 |
| ← kentnl_ left | 01:00 |
|
chrisr
| What is the consequence of not committing your changes before switching to another branch? Ive only just started using git today and i was wondering what would happpen | 01:00 |
| → kardan| joined | 01:00 |
|
chrisr
| does anyone know | 01:00 |
|
| ? | 01:01 |
|
FauxFaux
| Try it and see. | 01:01 |
|
sitaram
| the uncommitted changes get carried along | 01:01 |
| → kent\n joined | 01:01 |
|
sitaram
| as long as, for the files you changed, there are no differences between the branch you came from and the one you are going to | 01:01 |
|
chrisr
| So the file stays the same? What if the files are massively different? | 01:02 |
| → jinzhu joined | 01:02 |
|
sitaram
| massively different from what? | 01:03 |
|
chrisr
| If the file i change and dont commit is massively different than the same file on the branch I have changed to | 01:03 |
|
| What happens then? | 01:03 |
|
| If I then commit on the branch i have changed to, will I override the file in that branch? | 01:04 |
|
| This is hard to describe! | 01:04 |
| ← ereslibre left | 01:04 |
| → cytrinox_ joined | 01:05 |
|
sitaram
| chrisr: read what I said above carefully. if the from-branch and to-branch have the same copy of *this* file, it doesn't matter | 01:05 |
|
chrisr
| yeah, but what if the file isn't the same between branches, what happens then? | 01:06 |
|
sitaram
| try it and see | 01:06 |
|
| chrisr: also http://permalink.gmane.org/gmane.comp.version-control.git/104457 might be useful, though the question was asked in a different context | 01:06 |
| ← kpreid_ left | 01:06 |
|
sitaram
| chrisr: hint: git will try very, very, hard no to lose any "in progress" work, and you have to explicitly tell it to do so in most cases | 01:07 |
|
chrisr
| ah ok, thats good to know. | 01:07 |
| ← kardan_ left | 01:07 |
| ← shruggar left | 01:08 |
| → scarabx joined | 01:10 |
|
chrisr
| it still doesn't explain what happens to the changed file in the branch i am changing to. | 01:10 |
|
Ilari
| chrisr: It if the committed file is the same in both, the changes get carried over. If they aren't, the checkout will abort. | 01:11 |
| → DrTrevorkian_ joined | 01:12 |
|
Ilari
| chrisr: Working tree changes are not part of any branch. Nor are staged changes. | 01:12 |
| → vinse joined | 01:13 |
|
chrisr
| llari: I see, so that allows you to decide which branch to make the change on after you have made a change | 01:13 |
| → davey_ joined | 01:13 |
|
chrisr
| makes sense | 01:13 |
| ← DrTrevorkian left | 01:14 |
|
Ilari
| chrisr: There's also two switches: -f makes it discard your changes and -m makes it attempt to merge them (may produce conflicts, and is as such bit of russian roulette). | 01:14 |
|
chrisr
| ok cool | 01:15 |
|
| what does "git reset" do? and "git reset --hard"? | 01:15 |
|
sitaram
| http://sitaramc.github.com/concepts/soft-hard-mixed.html | 01:15 |
|
| might help; once you've read the man page | 01:16 |
|
shd
| chrisr: "git reset" undoes any 'git add' commands you might have done before the commit | 01:16 |
|
| chrisr: 'git reset --hard' undoes your git add commands and resets files to their original state before modifications | 01:16 |
| ← DrTrevorkian_ left | 01:16 |
|
shd
| chrisr: git reset doesn't touch any files, only undoes the staging (that is, git add) | 01:16 |
| ← name left | 01:16 |
| → DrTrevorkian_ joined | 01:17 |
|
Ilari
| Also any 'git rm --cached's | 01:17 |
|
shd
| chrisr: in addition, git reset and git reset --hard may place the branch pointer to arbitrary revision | 01:17 |
|
| chrisr: often one does: git reset HEAD^ after a bad commit, then git add && commit again | 01:17 |
| ← cytrinox left | 01:18 |
| cytrinox_ → cytrinox | 01:18 |
| ← relix left | 01:18 |
| → donpdonp joined | 01:18 |
|
Ilari
| git commit --amend | 01:18 |
| ← n3m left | 01:18 |
|
chrisr
| is there an easy way to remove all the .svn folders from all my brances? I still want them to exist, but not in git | 01:18 |
|
donpdonp
| ive got all these untracked files i want to delete. git reset --hard doesnt remove them. | 01:18 |
|
shd
| donpdonp: git clean --help | 01:18 |
|
donpdonp
| shd: thx | 01:19 |
|
Ilari
| chrisr: Import using git svn? | 01:19 |
|
chrisr
| llari: too late for that | 01:19 |
| → a389742 joined | 01:19 |
| ← simosx left | 01:19 |
|
chrisr
| I only found out about git-svn after i had been using git a couple of hours | 01:19 |
| DrTrevorkian_ → DrTrevorkian | 01:20 |
| → Mezner joined | 01:21 |
| ← donpdonp left | 01:21 |
|
Ilari
| chrisr: Export your changes as patches, import using git svn and apply the patches? | 01:21 |
|
chrisr
| ok ill do that? | 01:22 |
|
Ilari
| chrisr: Useful commands: git format-patch and git am | 01:23 |
| ← abw_ left | 01:29 |
| → paltman94 joined | 01:30 |
| ← xyz left | 01:30 |
| ← coolcat left | 01:31 |
| ← smuf left | 01:32 |
| → krynnotaur joined | 01:34 |
| ← meanburrito920_ left | 01:36 |
| → jinzhu_ joined | 01:37 |
| ← jinzhu left | 01:37 |
| ← Chillance left | 01:37 |
| → meanburrito920_ joined | 01:38 |
| → abw_ joined | 01:38 |
| ← krynnotaur left | 01:39 |
| ← KIAaze left | 01:41 |
| ← hyperair left | 01:42 |
| → schmukkk joined | 01:46 |
|
clever
| how do i undo changes to a local file, to sync it back to the git repo? | 01:51 |
|
| git revert? | 01:51 |
|
Ilari
| clever: 'git checkout -- <file>'. | 01:51 |
| ← gartenstuhl left | 01:53 |
| → hyperair joined | 01:59 |
| ← johnmclane left | 02:00 |
|
clever
| Ilari: thanks | 02:00 |
| ← j416 left | 02:00 |
| ← aziz left | 02:01 |
| → warthog9 joined | 02:01 |
| ← jinzhu_ left | 02:02 |
| ← intripoon left | 02:03 |
| ← kingfishr left | 02:07 |
| ← sagsousuke left | 02:10 |
| → murph__ joined | 02:11 |
| → kpreid_ joined | 02:11 |
| ← patcito left | 02:13 |
| → kingfishr joined | 02:18 |
| ← davey_ left | 02:18 |
| ← kingfishr left | 02:19 |
| → kingfishr joined | 02:19 |
| ← jschoolcraft left | 02:19 |
| → davey_ joined | 02:20 |
| ← davey_ left | 02:22 |
| ← murph_ left | 02:27 |
| ← lulika left | 02:33 |
| → sergiodj_ joined | 02:34 |
| ← sergiodj left | 02:34 |
| ← cenuij left | 02:34 |
| ← squentin left | 02:36 |
| → upul joined | 02:39 |
| ← hyperair left | 02:43 |
| → hyperair joined | 02:43 |
| ← jepler left | 02:45 |
| → jepler joined | 02:49 |
|
rektide
| i'm trying to find out which tip of a remote is the most recent | 02:51 |
|
| any suggestions ? | 02:51 |
| → hagabaka joined | 02:52 |
|
hagabaka
| I had a git tracked project, but for some reason created a new directory to "rewrite" it, but I still ended up copying a lot of the old code. how can I let git track the new directory and detect some of the old code if possible? | 02:54 |
| ← DrTrevorkian left | 02:57 |
| ← paltman94 left | 02:57 |
| ← warthog9 left | 02:58 |
| → warthog9 joined | 02:58 |
|
Ilari
| rektide: Fetch them and use merge-base to compare? | 03:01 |
| ← kpreid_ left | 03:04 |
| → kpreid_ joined | 03:04 |
| → zheng joined | 03:04 |
| → saintdev joined | 03:08 |
| ← muthu left | 03:09 |
| → sheriff joined | 03:09 |
|
sheriff
| I want to run git status on a directory to which I don't have write permissions | 03:09 |
|
| But I'm getting an error about not creating a lock file | 03:09 |
| → Pewpewarrows joined | 03:09 |
|
sheriff
| Any suggestsions? I want to get the key of the last commit in that dir | 03:10 |
| → daddius joined | 03:12 |
|
impulze
| hm where do i find parsecvs to import cvs into git? | 03:15 |
|
| seems like all mirrors are down | 03:15 |
|
sitaram
| sheriff: "key" of the last commit? do you mean the hash/SHA1? you can try 'git log'; git status is more for looking at uncommitted stuff | 03:15 |
|
sheriff
| Thanks - I found it in FETCH_HEAD in the end, but git log looks like it works too | 03:17 |
|
Ilari
| impulze: git://repo.or.cz/parsecvs.git maybe? | 03:20 |
|
saintdev
| i have a branch that introduces some new files. i just merged in changes from upstream, and this broke something, how do i bisect this? | 03:20 |
|
| ie: normally bisect checks out a commit in upstream, but i can't test it because it doesn't contain the new files. | 03:20 |
| → cytrinox_ joined | 03:20 |
|
Ilari
| saintdev: If both parents of merge work but the merge doesn't then that's pretty much mismerge. | 03:21 |
|
impulze
| Ilari: hm last commit in march, is that still the way to go? | 03:21 |
| → mumutou joined | 03:21 |
| → rchavik1 joined | 03:23 |
|
Ilari
| impulze: Might very well be... | 03:23 |
| ← rchavik left | 03:26 |
|
impulze
| hm also that clone doesn't seem to be compilable | 03:26 |
| rchavik1 → rchavik | 03:26 |
| → bcardarella joined | 03:26 |
|
saintdev
| Ilari: so it isn't possible (other than manual search) to figure what commit in another branch broke something? | 03:27 |
| → rchavik1 joined | 03:28 |
| ← rchavik left | 03:30 |
| rchavik1 → rchavik | 03:30 |
| → sparta joined | 03:32 |
|
sparta
| is there a git diff command to get the recursive difference between 2 local repositories | 03:33 |
|
| ? | 03:33 |
|
mumutou
| yes | 03:34 |
| ← epoxy left | 03:35 |
| → cedricv joined | 03:35 |
|
Ilari
| saintdev: The problem is that if the commits themselves are not testable for the bug, bisect can't locate the offending commit... | 03:36 |
| ← cytrinox left | 03:36 |
| cytrinox_ → cytrinox | 03:36 |
|
mumutou
| sparta:you can use the git command as "git diff branch1..branch2" | 03:36 |
| ← zheng left | 03:37 |
|
Ilari
| Just take into account that diff can only diff two commits that both exist locally. | 03:37 |
|
| (if diffing commits). | 03:37 |
|
sitaram
| sheriff: FETCH_HEAD doesnt necessarily show you what I think is meant by "last commit in that dir" (even assuming "dir" meant "branch"). It was a coincidence that your last *operation* was probably a pull which resulted in a fast forward merge. Ask if you need me to explain this better | 03:37 |
|
sparta
| what if i want to find diff between 2 codes in 2 different local repositories using git? | 03:38 |
| → flazz joined | 03:38 |
|
Ilari
| sparta: Fetch the branch from one to another first. | 03:38 |
|
sitaram
| sparta: you'll have to fetch one into the other and then diff | 03:39 |
|
saintdev
| Ilari: thanks, didn't think there was a way, but wanted to ask first ;) | 03:39 |
| ← Theravadan left | 03:40 |
|
sparta
| whats the exact command for fetch? | 03:41 |
|
Ilari
| saintdev: Well, bisect is binary bisection search. Perhaps use similar algo for finding the offending commit? | 03:41 |
| ← joeytwiddle left | 03:41 |
|
Ilari
| sparta: 'git fetch <remote> <branch>'. That fetches the branch as FETCH_HEAD. | 03:41 |
|
saintdev
| Ilari: yeah, that was the plan ;) | 03:41 |
|
sparta
| ok thanks | 03:44 |
| ← m0 left | 03:48 |
| ← jonmasters left | 03:48 |
| ← daemoen left | 03:48 |
| ← galador left | 03:49 |
| ← meanburrito920_ left | 03:49 |
| → galador joined | 03:49 |
| ← LRN left | 03:50 |
| → daemoen joined | 03:50 |
| → m0 joined | 03:50 |
| → jonmasters joined | 03:50 |
| → _jason553839 joined | 03:51 |
| → tjafk joined | 03:52 |
| ← m0 left | 03:53 |
| ← daemoen left | 03:53 |
| ← _|Nix|_ left | 03:53 |
| ← jonmasters left | 03:53 |
| → _|Nix|_ joined | 03:53 |
| → m0 joined | 03:53 |
| → daemoen joined | 03:54 |
| → epoxy joined | 03:54 |
| ← timj left | 03:54 |
| → jonmasters joined | 03:55 |
| ← cedricv left | 03:56 |
| ← kpreid_ left | 03:57 |
| → kpreid_ joined | 03:57 |
| ← bcardarella left | 03:58 |
| ← saintdev left | 04:02 |
| → bcardarella joined | 04:04 |
| ← eno left | 04:08 |
| ← _jason553839 left | 04:09 |
| → eno joined | 04:11 |
| ← javatexan left | 04:15 |
| ← engla left | 04:17 |
| → Creap_ joined | 04:20 |
| → nadim_ joined | 04:21 |
| → JackWinter2 joined | 04:21 |
| ← jonmasters left | 04:21 |
| ← m0 left | 04:21 |
| ← daemoen left | 04:21 |
| ← ChenKaie left | 04:21 |
| ← fcuk112 left | 04:21 |
| ← lresende left | 04:21 |
| ← Aikawa_ left | 04:21 |
| ← ryoma left | 04:21 |
| ← ia left | 04:21 |
| ← JackWinter left | 04:21 |
| ← pgalbraith left | 04:21 |
| ← kmamykin left | 04:21 |
| ← asksol left | 04:21 |
| ← jaalto left | 04:21 |
| ← Swimming_Bird left | 04:21 |
| ← corecode left | 04:21 |
| ← cantoma left | 04:21 |
| ← BTK left | 04:21 |
| ← Helios- left | 04:21 |
| ← Aestas left | 04:21 |
| ← \\steve left | 04:21 |
| ← cxreg left | 04:21 |
| ← jujule2_ left | 04:21 |
| ← tchan left | 04:21 |
| ← henr_k left | 04:21 |
| ← vkroll left | 04:21 |
| ← Creap left | 04:21 |
| ← fr| left | 04:21 |
| ← nadim left | 04:21 |
| ← tazle_ left | 04:21 |
| ← thermal_ left | 04:21 |
| ← Thumper_ left | 04:21 |
| ← a-atwood left | 04:21 |
| ← mugwump left | 04:21 |
| ← Iff left | 04:21 |
| ← agrippa left | 04:21 |
| ← lharc left | 04:21 |
| ← kyron left | 04:21 |
| ← johnjay left | 04:21 |
| ← ricky left | 04:21 |
| ← TDJACR left | 04:21 |
| ← tomoyuki28jp left | 04:21 |
| → fr joined | 04:21 |
| → johnjay joined | 04:22 |
| → ChenKaie joined | 04:22 |
| → BTK joined | 04:22 |
| → kyron joined | 04:22 |
| → corecode joined | 04:22 |
| → vkroll joined | 04:22 |
| → Thumper_ joined | 04:22 |
| → shad joined | 04:22 |
| → fcuk112 joined | 04:22 |
| → jujule2 joined | 04:22 |
| → ricky joined | 04:22 |
| → Aestas joined | 04:22 |
| → Iff joined | 04:22 |
| → m0 joined | 04:22 |
| → cxreg joined | 04:22 |
| ← segher left | 04:22 |
| → \\steve joined | 04:22 |
| → henr_k joined | 04:22 |
| → Aikawa joined | 04:23 |
| → krynnotaur joined | 04:24 |
| → asksol joined | 04:26 |
| → kmamykin joined | 04:26 |
| → pgalbraith joined | 04:26 |
| → tchan joined | 04:26 |
| → Helios- joined | 04:26 |
| → lresende joined | 04:26 |
| → agrippa joined | 04:27 |
| ← scarabx left | 04:27 |
| → ryoma joined | 04:27 |
| → bcardarella_ joined | 04:28 |
| ← bdimcheff left | 04:28 |
| → Swimming_Bird joined | 04:29 |
| ← bcardarella left | 04:30 |
| bcardarella_ → bcardarella | 04:30 |
| → loops joined | 04:30 |
| ← Mezner left | 04:31 |
| ← ahf left | 04:31 |
| ← lresende left | 04:31 |
| ← KingBug left | 04:31 |
| → tiglionabbit joined | 04:31 |
| → ahf joined | 04:32 |
| → KingBug joined | 04:33 |
| → thermal_ joined | 04:34 |
| → lharc joined | 04:34 |
| → ia joined | 04:35 |
| → tomoyuki28jp joined | 04:35 |
| → cantoma joined | 04:37 |
| → scarabx joined | 04:37 |
| → daemoen joined | 04:37 |
| → Jconyers joined | 04:38 |
| → wanders_ joined | 04:40 |
| ← wanders left | 04:41 |
| → rchavik1 joined | 04:41 |
| ← rchavik left | 04:42 |
| rchavik1 → rchavik | 04:42 |
| ← daddius left | 04:42 |
| ← scarabx left | 04:43 |
| ← Rolenun left | 04:44 |
| → reidab joined | 04:46 |
| ← kpreid_ left | 04:50 |
| → kpreid_ joined | 04:50 |
| ← JoeConyers left | 04:52 |
| → cedricv joined | 04:52 |
| → priidu joined | 04:56 |
| → ArdRigh joined | 04:58 |
| → lucs joined | 04:59 |
| → FunkeeMonk joined | 05:00 |
| → _jason553839 joined | 05:00 |
| ← sparta left | 05:01 |
| → abms1116 joined | 05:02 |
| ← flazz left | 05:03 |
| → panpainter joined | 05:08 |
| ← panpainter left | 05:08 |
| → tierney joined | 05:10 |
| ← tierney left | 05:10 |
| → DrTrevorkian joined | 05:11 |
| ← univate left | 05:12 |
| → elliottcable joined | 05:15 |
| ← Garen left | 05:16 |
| → univate joined | 05:18 |
| → mcdebugger joined | 05:21 |
| ← _jason553839 left | 05:24 |
| ← mumutou left | 05:24 |
| → amigojapan joined | 05:26 |
| → Dude-X joined | 05:27 |
| → RobertLaptop_ joined | 05:27 |
| ← RobertLaptop left | 05:27 |
|
amigojapan
| hi, I am trying to clone a repository on a remote linux machine using tortoise.... what should I enter in the url part? | 05:27 |
| ← RobertLaptop_ left | 05:28 |
| → RobertLaptop joined | 05:28 |
| → bsdvodsky joined | 05:31 |
| → chupino joined | 05:31 |
| ← cloudhead left | 05:31 |
| → girishr_ joined | 05:32 |
| ← chupino left | 05:33 |
|
sitaram
| amigojapan: I have no clue about tortoise, but man git-clone will show you a bunch of example URLs somewhere near the bottom (search for "ssh:" on that web page); I imagine tortoise will use the same URLs | 05:35 |
|
Gitbot
| amigojapan: the git-clone manpage can be found at http://git.or.cz/man/git-clone | 05:35 |
|
amigojapan
| thanx | 05:35 |
|
sitaram
| amigojapan: URLS to be precise | 05:36 |
|
| amigojapan: oops; I mean http://www.kernel.org/pub/software/scm/git/docs/git-clone.html#URLS | 05:36 |
|
amigojapan
| ok, anyway I am going to eat, try it after I get back, thanx guys | 05:36 |
|
| thanx sitaram | 05:36 |
|
sitaram
| you're welcome! | 05:36 |
| → yhager joined | 05:41 |
| ← drubies left | 05:41 |
| ← cedricv left | 05:41 |
| → _jason553839 joined | 05:42 |
| ← kpreid_ left | 05:43 |
| → kpreid_ joined | 05:43 |
| → Adlai` joined | 05:46 |
| Thell → Thell|Away | 05:46 |
| ← hyperair left | 05:46 |
| ← Adlai left | 05:47 |
| Adlai` → Adlai | 05:47 |
|
Dude-X
| whats' the default install directory of git in debian linux? and where can i find the config files | 05:52 |
|
Adlai
| Dude-X: try `which git' at the shell | 05:52 |
|
Dude-X
| ah which | 05:53 |
|
Adlai
| `man git-config' to read about the system-wide and personal config files. | 05:53 |
|
Gitbot
| Adlai: the git-config manpage can be found at http://git.or.cz/man/git-config | 05:53 |
|
Dude-X
| that's the command not "where' | 05:53 |
|
Fissure
| Dude-X: if it's a binary, and not essential, and not something that only root should use, it goes in /usr/bin | 05:54 |
| → sprasanna joined | 05:54 |
|
Fissure
| see: filesystem hierarchy standard | 05:55 |
| ← kingfishr left | 05:55 |
|
Dude-X
| basically i want to set the set name for the remote, so that on my home system i can clone the repo | 05:55 |
| ← agenteo left | 05:58 |
| → karmi joined | 05:59 |
| ← bcardarella left | 06:01 |
| → mugwump joined | 06:02 |
| → a-atwood joined | 06:02 |
| → jonmasters joined | 06:02 |
| → roop joined | 06:03 |
| ← abms1116 left | 06:03 |
| → tazle_ joined | 06:05 |
| → JoeConyers joined | 06:07 |
| ← JoeConyers left | 06:08 |
| → LRN joined | 06:09 |
| ← a389742 left | 06:09 |
| ← _jason553839 left | 06:09 |
| ← Adlai left | 06:14 |
| → Adlai joined | 06:14 |
| ← epoxy left | 06:15 |
| → maleknet joined | 06:16 |
| ← Jconyers left | 06:16 |
| ← warthog9 left | 06:16 |
| → gunmuc joined | 06:17 |
| → saurabhverma joined | 06:18 |
| → SunilThaha joined | 06:21 |
| → NemesisD joined | 06:23 |
|
NemesisD
| hi all. ive got changes on a separate branch and i want to merge it into master. i do a git diff and i think it will eliminate some of the changes i made in master. they don't conflict at all but im afraid if i checkout master and merge the other branch, it will just toss them out. is there a way to force a conflict or something so i can make sure my changes to master stay? | 06:24 |
| ← univate left | 06:25 |
| → univate joined | 06:26 |
| ← sheriff left | 06:27 |
| → tasslehoff joined | 06:28 |
| → jceb joined | 06:29 |
| → Cesario joined | 06:31 |
| → Tuomas joined | 06:32 |
|
Adlai
| man git-merge | 06:34 |
|
Gitbot
| Adlai: the git-merge manpage can be found at http://git.or.cz/man/git-merge | 06:34 |
|
SunilThaha
| NemesisD: is the edit is on the same line of a file on both branches, you will get a conflict | 06:34 |
|
NemesisD
| im on the manpage now, im looking at the strats but all of them seem to do things automagically | 06:34 |
|
Adlai
| NemesisD: look at the --no-commit option | 06:34 |
|
NemesisD
| SunilThaha: i don't think so, when i do a gif on the files i see '-' before my master lines that i don't want to lose with no + nearby | 06:35 |
| ← Aestas left | 06:35 |
|
NemesisD
| Adlai: ok, will --no-commit leave those lines in the file or will i have to look at an older version of the file or something? | 06:35 |
|
SunilThaha
| you could rebase with master | 06:36 |
|
Adlai
| no, it'll make the merge, just refrain from committing. | 06:36 |
| ← kpreid_ left | 06:36 |
| → kpreid_ joined | 06:36 |
|
Adlai
| I don't entirely understand your situation though. It sounds as though you'd get a merge conflict. | 06:36 |
|
NemesisD
| SunilThaha: my colleague doesn't have git, i checked out his branch on my machine, rebased with master, pasted his files into the dir, haven't even committed his changes to his branch yet | 06:37 |
| → Theravadan joined | 06:37 |
|
SunilThaha
| Adlai: I think the master has a line, which git deleted in the branch | 06:37 |
|
| s/git/got | 06:37 |
|
Adlai
| so that line will get deleted when they merge | 06:37 |
|
SunilThaha
| yes | 06:37 |
|
Adlai
| unless the master also touched that line | 06:37 |
|
NemesisD
| wait this shouldn't happen if my rebase with master was successful | 06:38 |
|
SunilThaha
| yes !! | 06:38 |
|
Adlai
| in which case there'll be a conflict | 06:38 |
|
NemesisD
| because it would add those non-conflicting lines | 06:38 |
|
SunilThaha
| Adlai: no ... | 06:38 |
|
Adlai
| no? | 06:38 |
|
SunilThaha
| since the line in the master did not change .. . | 06:38 |
|
| there won't be a conflict | 06:38 |
|
Adlai
| so it'll just get deleted. | 06:38 |
|
SunilThaha
| yes | 06:38 |
|
| Adlai doesn't really see what the problem is... | 06:39 |
| ← jceb left | 06:39 |
| ← niki left | 06:40 |
|
SunilThaha
| Adlai: master has 10 lines, he branches out, deletes line 5, and adds line 11,12 | 06:40 |
| → warthog9 joined | 06:40 |
|
SunilThaha
| when you merge with master .. line 5 gets deleted | 06:40 |
| → a389742 joined | 06:40 |
|
SunilThaha
| right? | 06:40 |
| ← nighthwk1 left | 06:41 |
| → ixti joined | 06:41 |
|
Adlai
| I think so... | 06:41 |
|
NemesisD
| wow. looking at the diff i suppose i should fix his mistakes before doing anything | 06:41 |
|
SunilThaha
| he doesn't want that to happen | 06:41 |
|
| NemesisD: you could branch out from master | 06:41 |
|
| and to fix the mistake | 06:41 |
| → hyperair joined | 06:42 |
|
SunilThaha
| then merge both the branches with the master | 06:42 |
|
| Adlai: he doesn't want the line 5 to go away | 06:42 |
| ← univate left | 06:43 |
|
Adlai
| so you merge without making a commit, reset half of it, and commit. | 06:43 |
|
| (assuming you have a recent git with 'reset -p') | 06:43 |
| ← karmi left | 06:45 |
| ← krawek left | 06:45 |
| → krawek joined | 06:45 |
|
NemesisD
| i think this will be okay | 06:47 |
|
| every time he sends me code he keeps changing all the curly braces on functions to start on a different line than the param list :| | 06:48 |
|
| but i have git and i control the repo so i put em back :D | 06:48 |
|
Adlai
| NemesisD: if it's your project, let him know what the coding style is (I find it helps to have a style file for clarity and consistency), and explain that contributions to your project should comply with your style preferences. | 06:51 |
| ← loops left | 06:51 |
| ← dunolie left | 06:53 |
| → Axius joined | 06:53 |
|
NemesisD
| Adlai: school project, i really don't have the authority to do that, so we must result to passive aggression for the time being | 06:55 |
| ← BeerSerc left | 06:55 |
|
NemesisD
| also, git ended up removing lines from my master in the merge :| | 06:55 |
|
SunilThaha
| NemesisD: You could use the astlye | 06:56 |
|
NemesisD
| whats that now? | 06:56 |
|
SunilThaha
| astyle --indent=spaces=4 --brackets=linux \ | 06:56 |
|
| --indent-labels --pad=oper --unpad=paren \ | 06:56 |
|
| --one-line=keep-statements --convert-tabs \ | 06:56 |
|
| --indent-preprocessor \ | 06:56 |
|
| `find -type f -name '*.cpp'` `find -type f -name '*.h'` | 06:56 |
|
| NemesisD: its a tool/utility to format the source | 06:57 |
| → kkoehne joined | 06:58 |
|
NemesisD
| that sounds friggn awesome | 06:59 |
| → univate joined | 06:59 |
| → aliceinwire joined | 06:59 |
| ← FunkeeMonk left | 07:00 |
| ← maleknet left | 07:02 |
| → maleknet joined | 07:02 |
| → sde_antrix2 joined | 07:03 |
| → Azeroth-Working joined | 07:03 |
| → rchavik1 joined | 07:04 |
| ← Octalot left | 07:05 |
| ← da5id left | 07:06 |
| → abms1116 joined | 07:06 |
| ← Azeroth-Work left | 07:07 |
| ← hyperair left | 07:08 |
| → ludde joined | 07:11 |
| → hyperair joined | 07:12 |
| ← rchavik1 left | 07:12 |
| → fluxusps joined | 07:12 |
| → alireza_ joined | 07:13 |
| → elTigre joined | 07:15 |
| ← univate left | 07:15 |
| → Pupeno joined | 07:17 |
| → cedricv joined | 07:17 |
| murph__ → murph | 07:18 |
| ← rchavik left | 07:20 |
| ← reidab left | 07:20 |
| → galderz joined | 07:21 |
| → thiago_home joined | 07:21 |
| → rchavik joined | 07:22 |
| ← girishr_ left | 07:23 |
| → roop_ joined | 07:23 |
| ← rchavik left | 07:25 |
| → dunolie joined | 07:26 |
| ← elliottcable left | 07:26 |
| ← krawek left | 07:26 |
| ← Adlai left | 07:26 |
| ← hyperair left | 07:26 |
| → Adlai joined | 07:27 |
| ← Pupeno left | 07:27 |
| → RDev joined | 07:28 |
| → mithro joined | 07:28 |
|
RDev
| hi | 07:29 |
| → j416 joined | 07:29 |
| ← kpreid_ left | 07:29 |
| → kpreid_ joined | 07:29 |
| ← maleknet left | 07:29 |
| → bentob0x joined | 07:30 |
| → univate joined | 07:31 |
|
RDev
| Can somebody suggest me a Project-/Time-/Version/Issue Management and a Wiki? I'm starting a new project with 5 people. We will develop a midsized PHP applikation for the next 6months. | 07:31 |
|
| we work on linux/windows and macintosh machines | 07:32 |
| → relix joined | 07:32 |
| → khmarbaise joined | 07:33 |
| → ph^ joined | 07:33 |
|
thiago_home
| well, Git for version management | 07:34 |
|
| then look at what issue-management tools are available with Git | 07:34 |
|
| the rest, I never use, so I can't help you | 07:35 |
| ← univate left | 07:35 |
|
RDev
| hmm | 07:36 |
|
| I'm trying to find the optimal working-environment and tools | 07:36 |
| → hyperair joined | 07:36 |
| ← ajonat left | 07:37 |
| ← roop left | 07:38 |
| → VVSiz joined | 07:40 |
|
thiago_home
| Git and the rest :-) | 07:41 |
| ← DrTrevorkian left | 07:41 |
|
thiago_home
| for project management, we're simply using a table in the wiki | 07:43 |
|
| gitorious has wiki support if you want, otherwise just use mediawiki | 07:43 |
| → tango_ joined | 07:44 |
| ← tango_ left | 07:45 |
| → tango_ joined | 07:45 |
| ← ph^ left | 07:45 |
|
sitaram
| RDev: take a look at indefero.com -- sort of like googlecode + git integration (they support other VCSs also I believe, though I only care about git) | 07:46 |
|
| Affero GPL I think or maybe plain GPL | 07:46 |
| → mastro joined | 07:46 |
| ← j416 left | 07:46 |
| → ph^ joined | 07:47 |
|
RDev
| thiago: I'm doing this for a highly professional Software-Developement course at the university. Where we shall use all Software-Development Standards available.. unfortunately nobody tells us which tools,IDE, vcs, etc.. | 07:47 |
|
| currently my problem is to find a way to organise all of the TODO's from a UML Diagramm and a Mind-Map to a realistic workflow | 07:48 |
| → reidab__ joined | 07:51 |
| → da5id joined | 07:52 |
|
thiago_home
| in my experience, there is no connection | 07:54 |
|
| UML and such things are to keep professors and clients happy that they are getting something | 07:55 |
| → Fullmoon joined | 07:55 |
|
thiago_home
| I'm sure other people have uses for them, though | 07:55 |
|
sitaram
| the larger diagrams, once printed, can be used for my kids school projects (the reverse side) | 07:55 |
|
| so dont go knocking UML ok? | 07:56 |
| ← reidab__ left | 07:58 |
| → onigiri joined | 07:58 |
| → mtkd_ joined | 07:59 |
|
RDev
| hmm.. | 08:00 |
| ← ryoma left | 08:00 |
|
sitaram
| ok... considering the dead silence, I'd better admit I was half-joking :) | 08:00 |
| → Weasel[DK] joined | 08:03 |
|
tasslehoff
| sitaram: good morning. mind talking a bit about the recipe you gave me friday? I've read some tutorials about git, so I understand more of it now :) | 08:07 |
| ← NemesisD left | 08:09 |
| → robin joined | 08:09 |
| robin → Guest58660 | 08:10 |
| → karmi joined | 08:10 |
| → univate joined | 08:11 |
| → yaotti joined | 08:13 |
| → yaotti_aw joined | 08:13 |
| yaotti_aw → yaotti_ | 08:13 |
| ← da5id left | 08:13 |
| → cilly joined | 08:13 |
| → nitrate joined | 08:16 |
| ← nitrate left | 08:16 |
| ← univate left | 08:16 |
| → reidab__ joined | 08:17 |
| ← dominikh left | 08:18 |
| → mcella joined | 08:20 |
| → reidab___ joined | 08:20 |
| reidab___ → reidab | 08:21 |
| → gunmuc1 joined | 08:21 |
| ← kpreid_ left | 08:22 |
| → kpreid_ joined | 08:22 |
| → reidab___ joined | 08:23 |
| ← robinr left | 08:23 |
| → psankar joined | 08:25 |
| ← RDev left | 08:25 |
| → dpino joined | 08:26 |
| ← thiago_home left | 08:26 |
| → johan-s joined | 08:26 |
| → j416 joined | 08:28 |
| ← Axius left | 08:29 |
| ← reidab left | 08:29 |
| → univate joined | 08:30 |
| → shruggar joined | 08:31 |
| → transplant joined | 08:31 |
| ← mithro left | 08:32 |
| → ntoll joined | 08:32 |
| ← Guest58660 left | 08:33 |
| → ciskje joined | 08:34 |
| ← gunmuc left | 08:35 |
| ← reidab__ left | 08:36 |
| → zoniq joined | 08:38 |
| ← reidab___ left | 08:40 |
| ← gunmuc1 left | 08:40 |
| ← ph^ left | 08:41 |
| ← j416 left | 08:42 |
| → j416 joined | 08:42 |
| → gunmuc joined | 08:43 |
| ← NET||abuse left | 08:48 |
| → rchavik joined | 08:48 |
| ← chrisr left | 08:48 |
| → omenking joined | 08:48 |
| → univate_ joined | 08:48 |
| → gretch joined | 08:48 |
| → ph^ joined | 08:49 |
| ← univate left | 08:50 |
| ← karmi left | 08:50 |
| → DavidKlein joined | 08:50 |
|
sitaram
| tasslehoff: hi; I'm at work so I'll be in and out but sure... | 08:51 |
| ← hyperair left | 08:51 |
|
tasslehoff
| sitaram: good. http://pastebin.com/m31c04746 (in case you don't have it at hand). | 08:52 |
|
sitaram
| yup | 08:52 |
|
tasslehoff
| sitaram: when you update the local repo and the user machines, why do you fetch instead of pull? you have to merge afterwards anyway, right? | 08:53 |
|
sitaram
| tasslehoff: no particular reason except that I usually want to do the merge manually -- I find I often don't use the defaults and specify everything explicitly (like 'git push origin master' instead of 'git push') | 08:55 |
|
| tasslehoff: so I just write it like that | 08:55 |
| → psoo joined | 08:55 |
|
sitaram
| tasslehoff: if I recall, you have an upstream, but you won't be pushing anything back to it -- is that correct? | 08:56 |
| ← xenoterracide left | 08:58 |
| ← Theravadan left | 08:59 |
| ← dunolie left | 08:59 |
| → hyperair joined | 09:01 |
|
tasslehoff
| sitaram: that's correct | 09:03 |
|
sitaram
| tasslehoff: and you want one central location where your developers can pull from / push to | 09:04 |
|
| right? | 09:05 |
|
tasslehoff
| sitaram: I tried making another cookbook without the mirror, using branches. does this look ok? http://pastebin.com/d62d26974 | 09:05 |
|
| sitaram: yes | 09:05 |
|
sitaram
| tasslehoff: I can't recall if your devs also have direct access to the internet (and thus this upstream repo)... | 09:05 |
|
| looknig | 09:05 |
|
| looking | 09:05 |
|
tasslehoff
| thanks. | 09:06 |
| ← DavidKlein left | 09:06 |
| ← univate_ left | 09:07 |
| → univate_ joined | 09:07 |
|
sitaram
| tasslehoff: my first comment is depending on the scale of your development, you will need "topic branches" that the devs will create. Example "bug-1234" might be a topic branch made to fix that bug. These topic branches will be pushed to the central server for testing/QA/validation etc | 09:08 |
|
| tasslehoff: i.e., they shouldn't be doing all the work off of mystable -- things will become a little messy unless they're all small, quick, localised, changes | 09:08 |
|
| tasslehoff: this also bears upon the fact that your flow does not have any "developer pushes his work up to the server" piece | 09:09 |
|
tasslehoff
| sitaram: an important not I have forgotten. we will most likely not change much in this software | 09:09 |
|
| it's like a third-party product that I want an extra level of stability control over. | 09:09 |
|
sitaram
| tasslehoff: that's fine, then you're saying you may not need topic branches -- fair enough | 09:10 |
|
| tasslehoff: do you know what a "bare" repo is? | 09:10 |
|
tasslehoff
| sitaram: barely (pun intended). I checked the manpage after you gave me your suggestion | 09:11 |
| → perez joined | 09:11 |
| → Jurko joined | 09:11 |
|
sitaram
| tasslehoff: any repo that someone will be *pushing to* (as opposed to only pulling from) should ideally be a bare repo | 09:11 |
|
| tasslehoff: http://sitaramc.github.com/concepts/bare.html has a good example | 09:11 |
| → ab3 joined | 09:11 |
|
_Ikke_
| a bare repo is a repository that only has the .git directory, no working directory | 09:11 |
|
sitaram
| _Ikke_: true but that usually does not help to understand why it is needed :) | 09:12 |
| ← alireza_ left | 09:12 |
|
_Ikke_
| sitaram: indeed, thats the next step | 09:12 |
| ← ashH left | 09:12 |
|
sitaram
| _Ikke_: :) | 09:12 |
| → maleknet joined | 09:13 |
| → joeytwiddle joined | 09:13 |
|
_Ikke_
| sitaram: is it btw possible at all to push to a none bare repo? | 09:13 |
|
tasslehoff
| sitaram: I'm following (I think). The devs might have to do some pushes to the server repo. | 09:13 |
|
sitaram
| tasslehoff: your paste line 22/23/24 says "merge changes from stable over to mystable" seems to indicate that you will do this *on* this repo, which in turn means it is not bare. I suggest you do that on some other repo (maybe one senior dev does that on his workstation and updates "mystable" on the server) | 09:13 |
|
| _Ikke_: sure it is... | 09:13 |
| ← yaotti left | 09:14 |
|
_Ikke_
| sitaram: ah ok, I thought it gave an error when I tried that | 09:14 |
|
sitaram
| _Ikke_: from 1.7 or so it will, if you push to the currently checkedout branch | 09:14 |
|
| _Ikke_: right now its a big bad warning | 09:14 |
|
| (I think) | 09:14 |
|
_Ikke_
| sitaram: ah ok. Clear. | 09:14 |
|
tasslehoff
| sitaram: sounds good, since the senior dev (me) will have to compile and test on hw anyway. | 09:15 |
|
sitaram
| tasslehoff: that will make this repo a bare repo -- it gets dev and stable from the outside, and it gets mystable from the people inside (likely only the senior dev) | 09:15 |
| → dunolie joined | 09:15 |
| ← kpreid_ left | 09:15 |
| → kpreid_ joined | 09:15 |
|
tasslehoff
| sitaram: ok | 09:16 |
| → kumbayo joined | 09:16 |
| ← Dude-X left | 09:17 |
|
sitaram
| tasslehoff: you still have to decide how your normal devs should update mystable. They can push to some other branch and the senior dev -- after testing/qa -- can merge those onto the official "mystable" (this is what *I* would do, unless the team is very small and communicates well) | 09:17 |
|
| tasslehoff: the other alternative is for everyone to be able to push to mystable, which is not a bad thing in small, well-knit teams | 09:17 |
|
tasslehoff
| sitaram: I choose the last one. for now we are three guys sitting next to eachother :) | 09:18 |
|
sitaram
| tasslehoff: agreed | 09:18 |
|
| tasslehoff: the latter means that the "senior dev" does not have any special privs as far as the system is concerned -- anyone can do what he does, although in practice -- like the situation you said -- they wont | 09:18 |
| ← GodEater left | 09:18 |
| → GodEater joined | 09:18 |
| → giallu joined | 09:19 |
| → simosx joined | 09:19 |
|
sitaram
| tasslehoff: one suggestion: turn on receive.denyDeletes and core.logAllRefUpdates -- this ought to protect you in case someone makes a mistake | 09:20 |
| ← transplant left | 09:20 |
| → babilen joined | 09:21 |
| → Pupeno joined | 09:22 |
|
tasslehoff
| sitaram: ok | 09:23 |
|
sitaram
| it wont tell you "who" but it'll tell you when, and more important it'll keep a record of any refs that accidentally got blown away | 09:23 |
|
| (for some 30 days or so anyway... there's some time limit in the config somewhere) | 09:23 |
| → xenoterracide joined | 09:24 |
| → aatanasov joined | 09:24 |
| → Axius joined | 09:25 |
| → JackWinter joined | 09:26 |
| → TDJACR joined | 09:27 |
|
tasslehoff
| sitaram: ok. new try at the server workflow : http://pastebin.com/ma469b0a | 09:27 |
| ← tono left | 09:28 |
|
sitaram
| tasslehoff: I dont get it... who and where will you be merging "stable" and "mystable"? | 09:29 |
| → cenuij joined | 09:29 |
| ← Fullmoon left | 09:30 |
|
tasslehoff
| sitaram: it's probably me that's not getting it :) | 09:30 |
| → ramlev joined | 09:30 |
|
sitaram
| tasslehoff: let me write up something; give me about 20 min -- getting into a meeting | 09:30 |
| → Grahack joined | 09:31 |
| → eevar2 joined | 09:32 |
| → jas4711 joined | 09:33 |
| → lucsky joined | 09:33 |
| ← univate_ left | 09:34 |
|
lucsky
| 'morning | 09:34 |
| → univate joined | 09:36 |
| ← JackWinter2 left | 09:36 |
| ← ab3 left | 09:37 |
| ← bsdvodsky left | 09:38 |
|
tasslehoff
| sitaram: here's a note on how I thought our discussion turned out :) http://pastebin.com/d208c1fc0 | 09:40 |
| ← j416 left | 09:41 |
| ← Pupeno left | 09:41 |
| → Stravy joined | 09:41 |
| → ruxpin joined | 09:45 |
| ← psankar left | 09:45 |
|
ruxpin
| hi, I need some advice | 09:45 |
|
| I stashed my master and checkouted a branch | 09:45 |
|
| then did some unsuccessful merge | 09:45 |
|
| and now I want to undo it | 09:45 |
|
| but I want to make sure that the master stash stays put | 09:45 |
|
| is "git checkout --hard HEAD" safe? | 09:45 |
| → tbf joined | 09:46 |
|
| tasslehoff read "reflog" as "re-flog", and thought it was a severe "blame" | 09:48 |
| → da5id joined | 09:48 |
| ← univate left | 09:48 |
| → corni joined | 09:50 |
| → intripoon joined | 09:51 |
| → psankar joined | 09:53 |
|
ruxpin
| ok, I took a backup and did the operation. stash was ok | 09:54 |
|
sitaram
| tasslehoff: http://pastebin.com/m55fec7c8 | 09:54 |
|
| tasslehoff: bonus: http://imagebin.ca/view/Yh8moN.html | 09:54 |
|
| tasslehoff: note the arrow in that picture; which ones are single headed and which ones are double | 09:55 |
| → Yuuhi joined | 09:55 |
|
sitaram
| tasslehoff: comments on your paste http://pastebin.com/d208c1fc0 -- line 3: you cant pull when using a bare repo (pull implies merge, which requires a working tree and human intervention in case of conflict), otherwise it's largely the same as what I posted except I gave commands | 09:56 |
|
tasslehoff
| sitaram: nice. and that image is quite correct :) | 09:56 |
|
sitaram
| graphviz++ :) | 09:57 |
|
| http://pastebin.com/d699f5a1b if you want the source of that little thing! | 09:57 |
|
tasslehoff
| :) | 09:59 |
| → univate joined | 09:59 |
| → froschi joined | 10:00 |
| ← roop_ left | 10:00 |
| ← tbf left | 10:03 |
| ← jspiros left | 10:04 |
| → capndiesel joined | 10:04 |
| → jspiros joined | 10:05 |
| → tbf joined | 10:06 |
| ← kpreid_ left | 10:08 |
| → kpreid_ joined | 10:09 |
| ← elTigre left | 10:11 |
| → __iron joined | 10:16 |
| ← cehteh left | 10:18 |
|
tasslehoff
| sitaram: hm. I'm having a bit of trouble already on the first step :) | 10:18 |
|
sitaram
| ok which one | 10:19 |
| → j416 joined | 10:19 |
| → schierbeck joined | 10:19 |
| ← Axius left | 10:19 |
|
tasslehoff
| I do a "git clone --bare", but then I can't say "origin/dev" afterwards when I create the branch? | 10:19 |
|
sitaram
| I guess a bare clone doesn't have origin/ -- I forgot :-) | 10:20 |
| → engla joined | 10:20 |
|
tasslehoff
| sitaram: ok. just wanted to check :) | 10:20 |
|
sitaram
| actually neither does it have an origin, it seems... hmm! | 10:21 |
|
| actually the way to do it is to create a bare repo, (not a clone), add a remote called origin, and fetch... testing | 10:22 |
| ← hyperair left | 10:23 |
| → hyperair joined | 10:23 |
| → Axius joined | 10:24 |
|
sitaram
| yeah that seems to work fine; changing doc | 10:24 |
| → jschoolcraft joined | 10:25 |
|
sitaram
| tasslehoff: http://pastebin.com/m61b3d468 -- sorry; still not quite tested (am juggling a few different things at the same time right now) so test it out | 10:27 |
| → amigojapan_ joined | 10:28 |
| muguruma1kensei → muguruma-kensei | 10:28 |
|
tasslehoff
| sitaram: I think I can figure out what I need from your previous paste now | 10:29 |
|
| thanks a _lot_ | 10:29 |
|
sitaram
| no problem | 10:29 |
|
| "previous paste"? you mean the last one... http://pastebin.com/m61b3d468 I assume | 10:29 |
|
| (not the one before) | 10:29 |
| ← gretch left | 10:30 |
| → Tama00 joined | 10:34 |
|
Tama00
| Hey | 10:34 |
| ← froschi left | 10:34 |
|
corecode
| can i change author info with rebase or amend? | 10:36 |
| ← scrooloose left | 10:36 |
| ← ArdRigh left | 10:37 |
| → scrooloose joined | 10:38 |
|
sitaram
| corecode: latest commit; git commit --amend --author should do. Others; you can rebase -i, choose "edit" so it stops, do the commit --amend --author, and proceed | 10:38 |
|
corecode
| i see | 10:38 |
|
| i did a format-patch + edit | 10:38 |
|
sitaram
| probably easier; for some silly reason its asking for an existing commit with the new author name I want to use :) | 10:40 |
| ← mtkd_ left | 10:40 |
| ← corni left | 10:40 |
| → corni joined | 10:41 |
|
sitaram
| probably to prevent horrible typos in manually typed stuff | 10:42 |
| → ihoka joined | 10:43 |
|
eevar2
| how do you guys back up your git repos? we have ~a dozen shared repos, some of them with a bit of data, on a central server, but only 1-2 of them are actually used regularly | 10:44 |
| → parasti joined | 10:46 |
| ← aliceinwire left | 10:46 |
| ← amigojapan left | 10:46 |
| → n3m joined | 10:47 |
|
eevar2
| worth taring them up individually and only keeping ones with changes? -- we probably should keep a few old backups around, in case someone decides to do some creative history-rewriting? | 10:47 |
| → aliceinwire joined | 10:47 |
| → patrikf joined | 10:47 |
| → gasbag joined | 10:48 |
| ← gasbag left | 10:49 |
| → yaotti joined | 10:51 |
|
eevar2
| ^^ or is there a failsafe method to prevent commits that change history? | 10:53 |
|
sitaram
| eevar2: last question first: yes: man git-config and look for receive.denyNonFastForwards | 10:54 |
|
Gitbot
| eevar2: the git-config manpage can be found at http://git.or.cz/man/git-config | 10:54 |
|
eevar2
| thanks | 10:54 |
| → name joined | 10:55 |
| → mattayers joined | 10:58 |
|
sitaram
| eevar2: my "backup" is a script that clones any new repos, and git fetch + git including tags in old ones | 10:58 |
| → dominikh joined | 10:58 |
|
sitaram
| I guess I could just rdiff-backup the repo itself too; should work | 10:58 |
| → ashH joined | 10:59 |
| ← onigiri left | 11:00 |
|
eevar2
| okies, just using rsync or similar sounds great | 11:01 |
| → kpreid__ joined | 11:02 |
| ← kpreid_ left | 11:02 |
| ← Axius left | 11:02 |
| ← Stravy left | 11:04 |
| ← muguruma-kensei left | 11:04 |
| → scarabx joined | 11:06 |
| → rraasch joined | 11:06 |
| → harinath joined | 11:08 |
| ← harinath left | 11:08 |
|
tasslehoff
| sitaram: I meant the one from before :) | 11:09 |
| → onigiri joined | 11:09 |
| ← univate left | 11:10 |
|
tasslehoff
| sitaram: but forget that :), now I mean the latest one :) | 11:11 |
| → solofight joined | 11:12 |
| → univate joined | 11:21 |
| → _Cro_Crx joined | 11:22 |
| → froschi joined | 11:23 |
| → altrux joined | 11:23 |
| ← VVSiz left | 11:24 |
| ← amigojapan_ left | 11:24 |
| → tfarina joined | 11:26 |
| ← upul left | 11:31 |
|
sitaram
| :) | 11:32 |
| → hugoxrosa joined | 11:32 |
| → Cro_Cr715 joined | 11:33 |
| ← brizly1 left | 11:33 |
| → brizly joined | 11:35 |
| → sagsousuke joined | 11:35 |
| → Axius joined | 11:36 |
| ← Cro_Crx left | 11:36 |
| → Araneidae joined | 11:39 |
| → lamont`` joined | 11:39 |
|
Araneidae
| I need to rebase from one Linux patch fix to another ... and I don't want to rebase the fixes, of course, which a simple `rebase <new-branch>` forces. What can I do instead? | 11:40 |
| ← lamont` left | 11:40 |
|
sitaram
| I didnt understand; maybe if you elaborated a little...? | 11:40 |
|
Araneidae
| Sure. I currently have some patches based on top of Linux 2.6.30.4 (or something like that) | 11:41 |
|
sitaram
| ok | 11:41 |
|
Araneidae
| I just ran `git rebase v2.6.31.5` and it came up with *lots* of patch conflicts | 11:41 |
|
| But of course, the patch conflicts are nothing to do with my patches, they're just conflicts between the two bug-fix branches | 11:42 |
|
Pieter
| you should probably only rebase your own changes | 11:42 |
|
Araneidae
| yep | 11:42 |
|
Pieter
| se the --onto option in man git-rebase | 11:42 |
|
Gitbot
| Pieter: the git-rebase manpage can be found at http://git.or.cz/man/git-rebase | 11:42 |
|
sitaram
| oh that's like this then: git rebase --onto v26315 v26304 mybranch | 11:43 |
|
| Araneidae: ^ | 11:43 |
|
Araneidae
| Ah, I think I see. | 11:43 |
|
sitaram
| Araneidae: in general, rebase --onto A B C says, take the changes between B and C, and put them on A, and call the new tip "C" | 11:43 |
| ← solofight left | 11:44 |
|
sitaram
| does that sound like what you're trying to do? | 11:44 |
|
Araneidae
| Thanks, that's what I need! | 11:44 |
|
sitaram
| ok | 11:45 |
| ← _Cro_Crx left | 11:45 |
| ← Azeroth-Working left | 11:45 |
|
engla
| Is it asocial to say "Only master won't be rebased, everything else are WIP branches" for my repo on github | 11:45 |
|
sitaram
| engla: I do that myself | 11:45 |
|
Araneidae
| Given that C is my current branch, is it optional? | 11:46 |
|
sitaram
| engla: although I *try* to not rebase the others its not a promide | 11:46 |
|
engla
| I have some trouble with a project I maintain. I have a clean master with some merges but mostly pick and rebase in contributions | 11:46 |
|
sitaram
| promise* | 11:46 |
|
engla
| yeah I want to make my work open but at the same time it is WIP | 11:46 |
|
| that collaborators can comment on or build upon | 11:46 |
| ← jschoolcraft left | 11:46 |
|
sitaram
| Araneidae: yes it is optional; but in general I try to specify everything in general (just a personal habit) | 11:46 |
| ← eevar2 left | 11:47 |
|
engla
| sitaram: great to hear you do the same | 11:47 |
|
sitaram
| yup -- the gitolite project says so on the front page of its home on github :) | 11:47 |
|
jjuran
| engla: git.git has branches that are documented as having forced updates. Also, you might consider keeping them in a namespace, like 'wip/topic'. | 11:47 |
|
engla
| sitaram: I just miss a field for this in the github project page, the description field is too small, I want a field where I can say "Branch policy for this repo::" (like repo.or.cz has a "field" for this) | 11:47 |
|
| jjuran: also a good idea | 11:48 |
|
sitaram
| engla: I just typed all that in there; only the entry field is too small -- just type it | 11:48 |
| → jschoolcraft joined | 11:48 |
|
sitaram
| engla: my current 'desc' reads: standalone, souped-up (started with per-branch permissions as the only extra feature, now lots more), version of gitosis. In perl. [Note: ALL branches except master are subject to rebase/rewind :-)] | 11:49 |
| ← jas4711 left | 11:49 |
|
engla
| it's not pretty. but there are many things about github's interface that could be better. I like the technical/social function still | 11:50 |
|
| the problem is that I manage the project, and I have one main collaborator that contributes a lot | 11:50 |
|
| it is a new collaboration, I mostly take his changes and rebase to my master, fixing them up | 11:50 |
|
| our trouble is finding the right way of branching and merging.. branches have to start in my master to be mergeable, but my collaborator prefers his master which was criss-cross merged with mine | 11:51 |
|
| which I can't merge of course | 11:51 |
| ← Axius left | 11:52 |
| → jas4711 joined | 11:52 |
|
engla
| So, does anyone have a template for an unformal workflow like this? | 11:52 |
| → d0k joined | 11:53 |
|
sitaram
| that's a tough problem, but it's a social one. I'd offer to make a feature freeze or something, fix up a nice master, have him fetch and start from that, and from then on everything he does has to be rebased off the origin master or something | 11:53 |
| → lorandi joined | 11:53 |
| ← __iron left | 11:53 |
|
sitaram
| he doesnt *have to* rebase though | 11:53 |
|
jjuran
| engla: I believe you need a common master. Perhaps your partner could rebase his work onto the common (older) master. | 11:53 |
| ← Adlai left | 11:54 |
|
engla
| yes, perhaps. should he just reset his master to mine, after that? | 11:54 |
|
| that will be a break, but perhaps not so important | 11:54 |
| → Adlai joined | 11:55 |
| ← kpreid__ left | 11:55 |
| ← doener left | 11:55 |
| → kpreid_ joined | 11:55 |
| lamont`` → lamont | 11:56 |
| ← sdboyer left | 11:56 |
| → sdboyer joined | 11:56 |
| ← jepler left | 11:56 |
| sergiodj_ → sergiodj | 11:57 |
| → cannonball joined | 11:58 |
| → doener joined | 11:59 |
| → solofight joined | 11:59 |
| ← jds left | 11:59 |
| → rolfb joined | 11:59 |
| ← abw_ left | 12:01 |
| → jepler joined | 12:01 |
| → qrush_ joined | 12:04 |
| ← tetha left | 12:08 |
| → tetha joined | 12:08 |
| ← Jurko left | 12:08 |
|
tasslehoff
| any guides on how I could/should setup git & git-daemon on Ubuntu? | 12:13 |
|
Adlai
| try `apt-get install git' | 12:15 |
|
ludde
| how do i change branch when i have local changes? | 12:15 |
|
Adlai
| ludde: man git-stash | 12:15 |
|
Gitbot
| ludde: the git-stash manpage can be found at http://git.or.cz/man/git-stash | 12:15 |
|
ludde
| hm | 12:16 |
|
| ok | 12:17 |
|
Adlai
| the examples section is pretty good | 12:17 |
|
| also, if you use gitk, you'll see where you've created stashes | 12:17 |
|
| (sp?) | 12:17 |
| → bdiego joined | 12:17 |
|
ludde
| why can't i just switch branch and it applies all my local changes on the new branch? | 12:17 |
|
Adlai
| oh, I assumed you wanted to keep your changes branch-local | 12:18 |
|
| you should be able to just checkout the new branch, and it'll preserve any changes in your working tree while checking out the new branch. If there are conflicts, though, it won't merge, it'll just refuse to switch. | 12:19 |
|
| so use the stash if there are conflicts and you want to merge them | 12:19 |
|
shruggar
| is there any way to automatically discard repeated trivial merges? | 12:19 |
|
Adlai
| shruggar: I'm not sure what you mean by discard | 12:21 |
| → DavidKlein joined | 12:22 |
| ← scarabx left | 12:22 |
| → elTigre joined | 12:23 |
|
shruggar
| Adlai, basically I've been using "git pull" to pull in upstream changes every-now-and-then, and usually this results in a trivial merge (no differences, just two parents). I'd like to find all the trivial merges and act as if I used "rebase" instead of "merge" on those | 12:23 |
| → roop joined | 12:24 |
| ← psankar left | 12:24 |
| → bcardarella joined | 12:25 |
|
Adlai
| hmm, I'm not sure exactly how to do that, but I can guess... first make a backup branch / lightweight tag, so you don't lose your current commit | 12:25 |
|
| then try just rebasing onto the upstream | 12:25 |
|
sitaram
| shruggar: by trivial merge do you mean "one which resulted in no conflicts"? because otherwise a pull should have become a fast-forward, not a merge at all | 12:25 |
| ← awnstudio left | 12:25 |
|
Adlai
| I -think- that rebase will ignore the merges, and only rebase your commits | 12:25 |
|
shruggar
| sitaram: I mean one which has no diff-output. Is there another term for that? | 12:26 |
|
Adlai
| sitaram: why? | 12:26 |
|
| a merge with no conflicts is still a merge, if the upstream and the local repo both have changes | 12:27 |
|
sitaram
| correct | 12:27 |
| Cro_Cr715 → Cro_Crx | 12:27 |
| ← corni left | 12:28 |
|
doener
| shruggar: note that "no diff output" doesn't necessarily mean "no changes" | 12:28 |
|
shruggar
| doener, okay, in what cases would there be changes but no diff output? | 12:28 |
| → corni joined | 12:29 |
|
doener
| shruggar: the default format for diffs for merge commits is "condensed combined" (--cc), which only shows differences if the contents in the merge commit differ from _all_ parents | 12:29 |
| → Mezner joined | 12:29 |
|
shruggar
| doener, well I guess I mean "empty patch". I'm not actually using diff output to judge | 12:29 |
| → n3m_ joined | 12:29 |
|
doener
| shruggar: so upstream repeatedly did "git commit; git revert HEAD" or what? | 12:30 |
|
| shruggar: if a merge commit was created, there were new commits | 12:30 |
|
| shruggar: or did you basically cherry-pick each upstream commit before you merged? | 12:31 |
|
shruggar
| upstream had a bunch of regular commits. I, down here, used "git pull" to pull in changes. I didn't use "git rebase", because I don't know until after I've done it whether or not the changes touched the same files. I _do_ want to keep anything which touched the same files, but anything else I just want to discard the merge-commits, since as far as I know they're worthless | 12:32 |
| → tstclair joined | 12:34 |
| → nevyn_ joined | 12:34 |
|
| doener is seriously confused | 12:35 |
| ← Cro_Crx left | 12:35 |
|
doener
| I guess "same files" means "the same files that I modified in my local commits" | 12:35 |
| ← Mezner left | 12:36 |
|
shruggar
| yes | 12:36 |
|
doener
| but the distinction between "merge" and "rebase" makes absolutely no sense to me | 12:36 |
| → Mezner joined | 12:36 |
| ← marcoecc left | 12:36 |
|
doener
| especially given the previous "I want to act as if I have used 'rebase' instead of 'merge' on those [trivial merges]" | 12:36 |
|
| when doing "git merge foo; git rebase foo", the "git merge foo" part is pretty much entirely pointless | 12:38 |
| ← n3m left | 12:38 |
| n3m_ → n3m | 12:38 |
| ← nevyn left | 12:38 |
| → Cro_Crx joined | 12:38 |
|
doener
| so any merge commits before a "trivial merge" would of course also be dropped | 12:38 |
|
shruggar
| (typing up a graph) | 12:39 |
| ← onigiri left | 12:39 |
| → Pupeno joined | 12:43 |
|
tasslehoff
| what do I need to setup to let other devs clone git repos from my pc? is it enough to start git-daemon with the correct options? | 12:44 |
| → bdiego_ joined | 12:45 |
| ← Adlai left | 12:46 |
| → Adlai joined | 12:46 |
|
Altreus
| I am in the middle of a rebase, and the conflict markers says on one side HEAD and on the other side is a summary from a commit I made | 12:46 |
|
tasslehoff
| I want them to be able to push & pull/fetch | 12:46 |
|
Altreus
| err | 12:47 |
|
| never mind | 12:47 |
| ChanServ set mode: +o | 12:47 |
| Mikachu changed the topic to: 1.6.5.2 | Homepage: http://git-scm.com | Everyone asleep or clueless? Try [email@hidden.address] | Git User's Survey 2009! http://git.or.cz/gitwiki/GitSurvey2009 | Channel log http://tinyurl.com/gitlog | Mailing list archives: http://tinyurl.com/gitml | Gits on git: http://tinyurl.com/gittalks | Pastebin: http://gist.github.com/ | "Alles wird Git" is over! | 12:47 |
| Mikachu set mode: -o | 12:47 |
|
Altreus
| I was going to ask how I find out what HEAD is but ... I know that | 12:47 |
|
| Altreus thick | 12:47 |
| → Ryback_ joined | 12:47 |
| ← kpreid_ left | 12:48 |
| → kpreid_ joined | 12:48 |
| → aziz joined | 12:49 |
|
Altreus
| hmm | 12:49 |
| → nenne joined | 12:50 |
|
Altreus
| no HEAD seems to have nothing to do with the change I can see | 12:50 |
|
nenne
| hello, i am a bit of a newbie when it comes to git and version control and need some help | 12:50 |
|
Altreus
| how can I find out where the other half of the conflict is coming from? | 12:50 |
|
nenne
| i have created a repo at my webb server and cloned it to my development machine, made a change and pushed it back upp and commited it. after this. how do i make it actually modify my files according to this commit? | 12:51 |
|
shruggar
| okay, I apologize for my horrible explanation, mixing of terms, and not fully thinking through what I was talking about before asking. Perhaps this will help: http://gist.github.com/218625 | 12:52 |
|
charon
| Altreus: during a rebase, HEAD is the last rebased commit and the other side will be the patch that it is currently trying to apply | 12:52 |
|
Altreus
| charon: How do I find out what commit that is | 12:53 |
|
| the "other side" patch | 12:53 |
|
| charon: The problem I'm having is that an entire function has been deleted, and it is a function that I have added to, hence the conflict, but I can't work out in what commit it was deleted. | 12:54 |
|
charon
| Altreus: interactive or normal rebase? | 12:55 |
|
Altreus
| normal | 12:55 |
| ← univate left | 12:55 |
| → univate joined | 12:56 |
|
charon
| Altreus: hm... never needed this, but i think the corresponding patch file should be in .git/rebase-apply/patch | 12:56 |
|
shruggar
| perhaps a way of describing what I want is: "I want to remove every merge-commit which can be removed without my hand-resolving any conflicts. If there were no conflicts, the result would be identical to if I had used rebase from the start. Otherwise it would just be as if I had only run merge when there were conflicts" | 12:56 |
|
Altreus
| thanking | 12:56 |
|
charon
| shruggar: this isn't very constructive because i can see that your request should be possible "usually"; but note that this is not possible in general, as a single merge might conflict where a series of merges did not | 12:57 |
|
Altreus
| ach | 12:57 |
| ← bdiego left | 12:57 |
| → boto joined | 12:57 |
|
Altreus
| This shows me the code I did but does not tell me who made the conflicting code :< | 12:57 |
|
shruggar
| charon, ah right, I forgot about that. Wasn't there a plan to have it fall-back to a more incremental approach when a conflict is detected? (but then I suppose the opposite is true too) | 12:58 |
|
charon
| oh | 12:58 |
|
Mizar-jp
| mm... may have failed to send the patch to [email@hidden.address] (content is http://github.com/mizar/gitk/blob/c20bcd2fdcb0583bbcf5fe4a5262374a84bbaafc/po/ja.po | 12:58 |
|
charon
| Altreus: i see. so you really want to know what HEAD has in that function, don't you? since the patch "on top" is yours | 12:59 |
|
| Altreus: so i'd just try 'git blame HEAD -- $file' | 12:59 |
|
Altreus
| charon: git show HEAD seems to just show a section of the patch that .git/rebase-apply/patch is | 12:59 |
|
| charon will have to actually test his own advice if the blame didn't help | 13:00 |
|
Altreus
| charon: Issue though | 13:01 |
| → _Cro_Crx joined | 13:01 |
|
Altreus
| There is nothing in the blame, because I am trying to blame a deletion | 13:01 |
| ← perez left | 13:01 |
|
Altreus
| So there's no lines in the file that correspond to it | 13:01 |
| ← priidu left | 13:02 |
|
Altreus
| hmm how do I find the FCA of master and $branch, again? | 13:02 |
|
charon
| i see. if you have a function name, you can try 'git log -p -Sfunction_name -- $file' | 13:02 |
|
| or you can play with blame --reverse | 13:02 |
| → johnnyg joined | 13:02 |
|
charon
| $(git merge-base A B), but what does the F stand for? | 13:03 |
| ← johnnyg left | 13:03 |
|
Altreus
| first common ancestor | 13:03 |
|
| I suppose it is redundant | 13:04 |
|
Mizar-jp
| I have this patch to a mailing list is sent to one hour before, the mail is anyone I have not received it? | 13:04 |
|
Altreus
| I will try the -p thing though, that sounds useful | 13:04 |
| → tono joined | 13:04 |
|
kupsch
| I'm trying to clone a svn project. The project has a svn std layout, so this is what I'm doing: | 13:08 |
|
| git svn -s clone https://subversion/repo/foo/foo/trunk foo | 13:08 |
|
| foo is the name of the project | 13:08 |
|
| however, cloning doesn't work. There's one line in the output which is strange: | 13:09 |
| → WebDragon joined | 13:09 |
| → modsognir joined | 13:09 |
|
kupsch
| Using higher level of URL: https://subversion/repo/foo/foo/trunk => https://subversion/repo/foo | 13:09 |
|
| any ideas? | 13:09 |
|
Ilari
| kupsch: Maybe 'git svn clone -s https://subversion/repo/foo'? | 13:10 |
|
kupsch
| .../repo/foo is the location of the repository | 13:10 |
|
| .../repo/foo/foo is the location of the project | 13:10 |
|
| git svn clone -s https://subversion/repo/foo/foo then? | 13:11 |
|
| no /trunk? | 13:11 |
| → roop_ joined | 13:11 |
|
Ilari
| kupsch: Well, try it. | 13:12 |
| ← elTigre left | 13:13 |
|
Altreus
| grr it didn't even find the *creation* of the string | 13:14 |
|
| this is nuts | 13:14 |
|
| charon: how likely would you say that git log -S thing is to find the string? :S | 13:15 |
| ← Cro_Crx left | 13:15 |
| → Axius joined | 13:15 |
|
Ilari
| Altreus: It finds commits where number of times that string is present changes... | 13:15 |
|
Altreus
| hmm so it may miss it | 13:15 |
|
| Time to go with the common ancestor route | 13:16 |
| → veiz joined | 13:16 |
|
Altreus
| According to $(git diff ${FCA}..HEAD $file) the file in question has not changed between the merge base of this branch and master | 13:17 |
|
| Irrespective of how I form that diff (I tried various ways) | 13:18 |
|
| I made sure I was on master when I used HEAD :P | 13:18 |
|
| However, if I rebase $branch onto master, it shows me a conflict in which this function has been deleted completely | 13:18 |
|
| and it does not exist in master | 13:18 |
|
charon
| Altreus: well, the precise definition is: -Sstring finds all commits whose pre- and postimage counts of 'string' differ | 13:18 |
|
| so moving 'string' around does not trigger it, but adding/deleting instances does | 13:18 |
|
Altreus
| hmm so it should have found it *somewhere* in the history | 13:19 |
|
tasslehoff
| "remote HEAD refers to nonexistent ref, unable to checkout" <- is this the way it should be when I clone a bare repository? | 13:20 |
|
Ilari
| Altreus: 'git diff A...B' impiles taking ancestor of A and B and then diffing it to B. | 13:22 |
|
Altreus
| yarr | 13:22 |
|
| I tried A..B, A...B and A B and got no diff at all for the file | 13:22 |
| → kukks joined | 13:22 |
|
Altreus
| ... which is odd because the log shows stuff | 13:22 |
|
Ilari
| tasslehoff: That happens when remote default branch is set to point to something nonexistent. | 13:22 |
|
Altreus
| so that stuff must be older than B. /me checks | 13:22 |
| ← sprasanna left | 13:23 |
| ← corni left | 13:23 |
|
kupsch
| Ilari: now only the trunk is checked out. I can't see the other svn branches in 'git branch'. | 13:23 |
| → corni joined | 13:23 |
|
Ilari
| kupsch: 'git branch -r'. | 13:24 |
| ← Adlai left | 13:24 |
|
kupsch
| aaah. alright. thanks! | 13:24 |
| bdiego_ → bdiego | 13:24 |
| ← cenuij left | 13:24 |
| → Adlai joined | 13:24 |
| ← _Cro_Crx left | 13:24 |
|
tasslehoff
| Ilari: but it's ok? I'm following http://pastebin.com/m61b3d468 that sitaram wrote down for me, and the message come when I clone the repo from my server to a dev-machine | 13:25 |
| ← roop left | 13:26 |
|
Ilari
| tasslehoff: If you don't have branch called 'master' on server, you need to set the default branch to something existent or you get that error. | 13:27 |
| → roop__ joined | 13:27 |
|
Ilari
| tasslehoff: That needs to be done on server (excluding some management interfaces that can set default branch). | 13:27 |
|
tasslehoff
| ah. I'm playing with openembedded, and I think their default is org.openembedded.dev, not master | 13:27 |
| → spearce joined | 13:29 |
| ← veiz left | 13:29 |
|
Ilari
| tasslehoff: The clone suceeds. You just need to create a local branch from some remote branch (as usual). | 13:29 |
|
| spearce: Hi spearce. | 13:29 |
| ← hummingkid left | 13:30 |
| → Blackshark joined | 13:30 |
| → Cro_Crx joined | 13:30 |
| krynnotaur → Tuttle_ | 13:31 |
| Tuttle_ → Krynnotaur | 13:31 |
| ← gunmuc left | 13:32 |
|
tasslehoff
| Ilari: ok. I tried creating a local branch from "git branch --track mystable origin/stable". that worked, but on checkout it said "warning: You appear to be on a branch yet to be born", but that should be okay to, right? | 13:32 |
|
Altreus
| This is nuts. The merge-base of master and $branch is more recent than the last change to $file in master. However, I get conflicts rebasing $branch onto master. | 13:33 |
| ← ihoka left | 13:33 |
|
Altreus
| But if $file has not changed in master since the merge-base, where is the conflict coming from? | 13:33 |
| → bcardarella_ joined | 13:33 |
|
Altreus
| and why does my $file not have the same content? | 13:33 |
|
| And how can I find out? | 13:33 |
|
| being the main question | 13:33 |
| ← bcardarella left | 13:34 |
| bcardarella_ → bcardarella | 13:34 |
| → toscano joined | 13:34 |
|
Ilari
| tasslehoff: Yes. | 13:34 |
| → priidu joined | 13:34 |
|
Ilari
| tasslehoff: Except that branch should be named 'stable' so names match and git push pushes it without any refspecs. | 13:35 |
| ← Adlai left | 13:36 |
| ← sde_antrix2 left | 13:36 |
| → javatexan joined | 13:36 |
| → Adlai joined | 13:36 |
| → paltman94 joined | 13:36 |
| → Jinzhu joined | 13:38 |
| ← jlilly left | 13:40 |
|
spearce
| morning Ilari | 13:40 |
| → bousquet joined | 13:40 |
| → bdimcheff joined | 13:40 |
| ← kpreid_ left | 13:41 |
| → kpreid_ joined | 13:41 |
| ← roop_ left | 13:42 |
| ← Adlai left | 13:42 |
| → keystr0k joined | 13:42 |
| → lohrun joined | 13:42 |
| ← Cro_Crx left | 13:42 |
| → sde_antrix2 joined | 13:43 |
| → Adlai joined | 13:43 |
| ← Swimming_Bird left | 13:43 |
| → Cro_Crx joined | 13:44 |
|
Ilari
| spearce: Oh, and if you want some numbers on remote helpers without any buffering (pushing updates for stuff): TX: 6 reads, 24530 bytes on layer 7 (4088 byte average TU). RX: 2 reads, 397 bytes on layer 7 (198 byte average TU). | 13:47 |
| ← aliceinwire left | 13:47 |
|
phako
| how can I install something that is the output of a custom compiler commanf? | 13:49 |
|
| command | 13:49 |
|
Ilari
| phako: Huh? What binary it is? | 13:50 |
| → _Cro_Crx joined | 13:50 |
|
phako
| er make | 13:51 |
|
| or a custom makefile target | 13:51 |
| ← agile left | 13:51 |
|
| phako needs to build a kernel module | 13:51 |
|
broonie
| This doesn't sound like a git-related question? | 13:51 |
| ← aspotashev left | 13:51 |
|
phako
| oops | 13:51 |
|
| sorry | 13:52 |
| ← bcardarella left | 13:52 |
|
phako
| wrong channel | 13:52 |
|
Ilari
| spearce: The only problematic place is the server initial ref advert. That's flushed at the end, so buffering that for send isn't that hard... | 13:54 |
| ← Adlai left | 13:54 |
| → Adlai joined | 13:54 |
| → aspotashev joined | 13:55 |
| → ctusar joined | 13:56 |
| → Blub\0 joined | 13:56 |
|
Blub\0
| what's the best way to maintain a "stable" branch? | 13:56 |
|
| can I somehow track the commits which have been cherry-picked? | 13:56 |
|
| or compare the trees of two or more branches | 13:56 |
| ← paltman94 left | 13:57 |
|
Blub\0
| so that for example I can see which commits are not yet in the stable tree? | 13:57 |
|
| also, can I somehow pick multiple commits so that they're automatically applied in order? | 13:58 |
|
| as in, start a merging process, pick commits, and then let them be applied in order to avoid conflicts which could emerge by simply cherry-picking the commits in the wrong order? | 13:59 |
|
reto__
| Blub\0: git cherry | 13:59 |
| → GarethAdams|Work joined | 14:00 |
|
Blub\0
| oh... | 14:00 |
|
| hehe | 14:00 |
| ← _Cro_Crx left | 14:00 |
|
reto__
| Blub\0: git cherry -v master stable # <- what commits are applied to stable but not to master | 14:00 |
|
Blub\0
| it's that obvious :D | 14:00 |
|
reto__
| well, but you see the basic problem? the commit ids are not the same | 14:00 |
|
| there are two different commtis | 14:01 |
|
| if you cherry-pick changes | 14:01 |
|
| you should, as long as possible, git merge stuff | 14:01 |
|
Blub\0
| can I merge a certain commit directly? (git merge stable <commit>) ? | 14:01 |
|
reto__
| nah | 14:02 |
|
| you can't merge a commit, you can only merge the commit and its whole history | 14:02 |
| ← tasslehoff left | 14:02 |
| → nighthwk1 joined | 14:02 |
|
reto__
| a given commit contains everything (lets say all features) up to (including) the commit | 14:02 |
|
shruggar
| Blub\0, I'd recommend putting commits into a "fixes" branch, and merging that with both stable and dev | 14:03 |
|
reto__
| shruggar: whats the base of the fixes branch? | 14:03 |
|
Mizar-jp
| Phew, it seems finally succeeded in sending mail. http://permalink.gmane.org/gmane.comp.version-control.git/131250 | 14:03 |
| → bcardarella joined | 14:03 |
|
reto__
| Blub\0: cherry-picking -> just get the changes described in that single commit, git merge, merge everything within this commit | 14:04 |
| ← jas4711 left | 14:05 |
| → vbabiy joined | 14:05 |
|
Blub\0
| reto__: but I'd want to pick commits one by one without merging the full dev tree into stable | 14:05 |
|
reto__
| I know, use the fix branches as shruggar has adviced | 14:06 |
| → jds joined | 14:06 |
| → flazz joined | 14:06 |
|
reto__
| but usually you develop stuff in your master tree | 14:06 |
|
| I know :) | 14:06 |
| ← flazz left | 14:06 |
|
shruggar
| reto__, the base of fixes would be the beginning of "stable" | 14:07 |
| → flazz joined | 14:07 |
|
Blub\0
| shruggar: can you elaborate on that idea? | 14:07 |
| ← alezandro left | 14:07 |
|
reto__
| shruggar: where master and stable diverts? | 14:08 |
|
shruggar
| reto__, right | 14:08 |
|
reto__
| shruggar: hmm, but therefore you should/woluld/could never develop on stable itself? | 14:08 |
|
Altreus
| Is there a simple way of saying "What's the first commit before $(git merge-base $A $B) that changes $file"? | 14:08 |
|
shruggar
| reto__, Blub\0: basically, as if you were applying fixes directly to "stable" and merging them back into "master", but personally I think that looks weird in the history | 14:08 |
|
reto__
| shruggar: alright... :) | 14:09 |
|
shruggar
| reto__, personally I don't like developing on stable itself, but that's just personal preference | 14:09 |
|
Altreus
| I mean I can do it in more than one command but I wondered if git had a feature | 14:09 |
| ← vbabiy left | 14:09 |
|
reto__
| shruggar: got your point :) thx | 14:09 |
|
Blub\0
| oh, so, work on dev, cherry-pick into fixes and merge it with stable? | 14:09 |
|
| what advantage would that give me? | 14:09 |
|
reto__
| not cherry-pick | 14:09 |
| → JNR___ joined | 14:10 |
| → dimsuz joined | 14:10 |
|
reto__
| Blub\0: so basically create a branch called 'fixes' on stable, when you fix something 'in stable' you develop it in fixes, then you merge it from master and you merge it from stable | 14:10 |
|
| but in stable it will be only a fast forward merge | 14:10 |
|
dimsuz
| hi! is there a command to help me find which remote branch is tracked by my local branch? | 14:10 |
|
shruggar
| Blub\0, when you cherry-pick, git doesn't keep track of it having been put from one into the other | 14:10 |
|
Blub\0
| hmmmm | 14:10 |
|
reto__
| dimsuz: less .git/config | 14:10 |
|
Blub\0
| that sounds weird | 14:10 |
|
shruggar
| Blub\0, basically, it bypasses what you might considered to be git's merge-tracking | 14:10 |
| ← ramlev left | 14:11 |
|
Blub\0
| I'd like to "pick" commits and still keep track of them :S | 14:11 |
|
| not necessarily the full tree | 14:11 |
|
reto__
| Blub\0: well if your commit is already in master you can't do anything but cherry-pci kthem | 14:11 |
|
Blub\0
| well, full history | 14:11 |
|
| hmm | 14:11 |
|
reto__
| Bucciarati: but if you switch to the 'stable-fixes' branch before working on these 'stables fxies' | 14:12 |
| ← Adlai left | 14:12 |
| → johnny_ joined | 14:12 |
|
johnny_
| t | 14:12 |
|
reto__
| Blub\0 I mean | 14:12 |
| → Swimming_Bird joined | 14:12 |
|
reto__
| Blub\0: you can use git's much more flexible system | 14:12 |
|
| Blub\0: basically you can do a git checkout master; git merge stable-fixes; and be sure that everything is there | 14:12 |
| → Adlai joined | 14:12 |
|
reto__
| at any time | 14:12 |
|
dimsuz
| reto__: i suspected that :) just wondered if there's a command option perhaps (didn't find any on man git-branch). thanks | 14:13 |
|
Gitbot
| reto__: the git-branch manpage can be found at http://git.or.cz/man/git-branch | 14:13 |
| → perez joined | 14:14 |
|
Blub\0
| okay | 14:14 |
|
| thanks for the help | 14:14 |
|
| I think we found our solution | 14:14 |
|
| :) | 14:14 |
|
reto__
| Blub\0: read man git-cherry if you have to cherry-pick | 14:14 |
|
Blub\0
| (one step closer to moving from svn to git :P hrhr) | 14:14 |
|
Gitbot
| reto__: the git-cherry manpage can be found at http://git.or.cz/man/git-cherry | 14:14 |
| ← johnny_ left | 14:14 |
| ← schierbeck left | 14:15 |
|
reto__
| dimsuz: perhaps there's even a command, but I'd open the config file :) | 14:15 |
| ← Cro_Crx left | 14:15 |
|
Altreus
| can I omit <since> from <since>..<until> ? | 14:15 |
| → jmcantrell joined | 14:16 |
|
dimsuz
| reto__: yep. no probs, i just did it ;) | 14:16 |
|
Blub\0
| another question - related though: | 14:17 |
|
| say, I have a history: A B C D E F G, and the stable branch is now: A B C' D' G' (so, C, D and G have been cherry-picked) | 14:17 |
| ← corni left | 14:18 |
|
Blub\0
| now I'd want to merge master to stable | 14:18 |
| → corni joined | 14:18 |
|
reto__
| wait a sec.. | 14:18 |
|
| you mean merge stable into master? | 14:18 |
|
Blub\0
| eh | 14:19 |
|
| dev to stable | 14:19 |
|
| master would be the dev branch | 14:19 |
|
reto__
| alright | 14:19 |
|
| k! | 14:19 |
|
Blub\0
| so, master/dev = A .. G, and stable with G' at the end, now I merge them | 14:19 |
| → jlilly joined | 14:19 |
|
Blub\0
| so that E and F are merged into it | 14:19 |
|
| could that result in conflicts? | 14:19 |
|
| since I basically apply E and F after having applied G already? | 14:19 |
|
reto__
| you mean because of the cherry-picks? | 14:19 |
|
Blub\0
| yes | 14:19 |
|
reto__
| nana.. git is pretty intelligengt | 14:20 |
|
Blub\0
| because, this currently can easily cause conflicts in SVN | 14:20 |
|
reto__
| should handle that pretty well | 14:20 |
|
Blub\0
| with some strange tool we use | 14:20 |
|
reto__
| :) | 14:20 |
|
Altreus
| how do I get a log for a file *before* $commit, rather than since/ | 14:20 |
|
| I tried git log ..$commit -- $file | 14:20 |
|
reto__
| Altreus: gimme a seoncd | 14:20 |
|
Altreus
| it just blanked me :( | 14:20 |
|
sitaram
| Altreus: ..$commit means HEAD..$commit, which is probably empty | 14:21 |
|
charon
| Altreus: git log $commit -- $file | 14:21 |
|
sitaram
| right | 14:21 |
|
Altreus
| sitaram: oh | 14:21 |
|
| that was silly of me | 14:21 |
|
| well maybe | 14:21 |
|
| I assumed that <since>..<until> meant that ..<until> would DWIM | 14:21 |
| → s0ber joined | 14:22 |
|
reto__
| Blub\0: merging cherry-picked (duplicated) commits shouldn't be a big problem with git ,but perhaps somebody with more experience can comment on that | 14:22 |
| → jas4711 joined | 14:23 |
| → hgb21 joined | 14:23 |
|
sitaram
| Altreus: as charon said, for "everything upto commit" you dont need a special syntax; it's what you get by default anyway. You need something special to make it stop, and that's what the left side of the .. is | 14:24 |
|
Blub\0
| okay, thanks, I hope someone who knows more will comment on it :) | 14:24 |
|
| I'll be around :) | 14:24 |
| → patcito joined | 14:24 |
|
Blub\0
| usually 24/7 - I just forgot to rejoin when the server got rebooted :P | 14:24 |
|
charon
| Altreus: it does DWIM, leaving out one side of .. and ... means HEAD | 14:25 |
| ← JNR___ left | 14:25 |
|
charon
| Altreus: the <commit-ish> arguments mean: all commits reachable from the positive refs, but not reachable from the negative ones | 14:25 |
|
| Altreus: so 'git log A B ^C' gives you all commits contained in A or B (or both), but not contained in C | 14:25 |
|
| Altreus: A..B means B ^A, see man git-rev-parse for the other equivalences | 14:26 |
|
Gitbot
| Altreus: the git-rev-parse manpage can be found at http://git.or.cz/man/git-rev-parse | 14:26 |
|
Altreus
| charon: so I M the wrong thing :) | 14:26 |
| ← priidu left | 14:26 |
| → designs703 joined | 14:28 |
|
designs703
| is there an equivalent to the -i flag for ssh in git commands? | 14:28 |
|
| for passing a different key file | 14:28 |
| ← babilen left | 14:28 |
| → bdiego_ joined | 14:28 |
| → Lemon|work joined | 14:29 |
|
reto__
| designs703: i'm sure there is ,but why dont you specify it in your .ssh/config | 14:29 |
|
| _ | 14:29 |
|
Lemon|work
| Do i loose my subversion commit history if i migrate to git? | 14:29 |
|
| or is there a way to convert to git? | 14:30 |
| → agile joined | 14:30 |
|
reto__
| Lemon|work: there's even a way to use svn as a backend for a git local repo | 14:30 |
|
| Lemon|work: check out git svn | 14:30 |
|
| Lemon|work: there are some limitations of course | 14:30 |
| ← s0ber_ left | 14:30 |
|
designs703
| reto__, I'm working on a deployment app that ideally ships with the ssh key | 14:30 |
|
Lemon|work
| i guess i can just start from "version 1" | 14:30 |
|
| in git | 14:30 |
|
| and archive the subversion repository | 14:31 |
| → loincloth joined | 14:31 |
|
Lemon|work
| got another question | 14:31 |
|
| more concept like | 14:31 |
|
reto__
| Lemon|work: its really easy, if you use the trunk/branches/tags structurein svn | 14:31 |
|
| Lemon|work: its git svn clone -s <path to svn> | 14:31 |
|
Lemon|work
| ah ok | 14:31 |
|
reto__
| Lemon|work: and it transforms all your history.. | 14:31 |
|
Lemon|work
| cool i'll try that | 14:31 |
|
hgb21
| I am trying to make my first commit to my remote serever where git is but I am getting the following error anyone know why ? | 14:31 |
|
| http://gist.github.com/218695 | 14:31 |
|
reto__
| Lemon|work: invest the 20 minutes.. or so | 14:31 |
| → jrmuizel joined | 14:31 |
|
reto__
| Lemon|work: its probably worth it | 14:32 |
|
nenne
| i have created a repo at my webb server and cloned it to my development machine, made a change and pushed it back upp and commited it. after this. how do i make it actually modify my files according to this commit? | 14:32 |
|
Lemon|work
| btw what is the best way to use stable, beta, development branches? | 14:32 |
| ← johan-s left | 14:32 |
|
Lemon|work
| just merge to the beta first and if the beta is stable merge that to the master ? | 14:32 |
|
reto__
| nenne: you shouldn't use a repoistory with a working dir as a central repo | 14:32 |
|
| nenne: I mean: your central repo shouldnt have its data checked out | 14:33 |
|
| nenne: ideally you have three repos, the central one, one checkout at your desk, one where you want | 14:33 |
|
designs703
| it appears that this is not an option (based at least on the help for clone and ls-remote) | 14:33 |
|
reto__
| nenne: btw. github is really ncie for the central one | 14:33 |
|
hgb21
| anyone know why I have this error . . . http://gist.github.com/218695 | 14:34 |
| ← kpreid_ left | 14:34 |
|
nenne
| reto__: aah ok, pretty new to version control and git :) . what i trying to do is using git as a way of cloing my webapp and modding it and then using git to set the changes "live". this might not be the way its ment to be used? | 14:34 |
| ← bdiego left | 14:34 |
|
designs703
| maybe there's a way to include the path in the SSH string? | 14:34 |
| → kpreid_ joined | 14:34 |
|
reto__
| nenne: thats no problem, just use an independet '--bare' central repository | 14:35 |
|
| nenne: imho thats the best way to deploy an application | 14:35 |
|
| nenne: just use a dedicated 'bare' repo as a central hub | 14:35 |
|
| :) | 14:35 |
|
hgb21
| ? | 14:35 |
| ← Aides left | 14:35 |
|
designs703
| :P http://osdir.com/ml/git/2009-02/msg00355.html | 14:36 |
|
reto__
| hgb21: fatal: remote origin already exists. read the error | 14:36 |
|
| hgb21: you alread<y have this remote, check out your .git/config and man git-remte | 14:36 |
|
hgb21
| ? | 14:36 |
|
nenne
| reto__: so i push my changes to this bare repo and use it as a stage, then i retreive it from there to the actual live server? so i have dev - stage and live in a sense | 14:36 |
|
reto__
| hgb21: the first command already spawns a error | 14:36 |
|
hgb21
| i am folowing this http://articles.slicehost.com/2009/5/13/capistrano-series-setting-up-git | 14:37 |
|
reto__
| nenne: you have a checkout 'dev', a checkout 'live' and a central repo which is --bare.. | 14:37 |
|
| of course, thanks to git, all your repos can work independent | 14:37 |
| → temoto joined | 14:37 |
|
temoto
| Is it normal, that git clone URL creates local master and git clone /path/bundle does not? | 14:38 |
|
| What the reasoning behind this difference? | 14:38 |
|
nenne
| reto__: reto__ | 14:38 |
| → LiamH joined | 14:39 |
|
nenne
| reto__: oops mt . i think i understand now. will try it out. thanks alot | 14:39 |
|
reto__
| so i'm afk | 14:39 |
| → bushwakko joined | 14:39 |
|
bushwakko
| How can I diff my undcommited patches against the server | 14:40 |
|
| using svn as a central repo | 14:40 |
| → Turias joined | 14:40 |
| ← Adlai left | 14:41 |
| → Adlai joined | 14:41 |
|
designs703
| mercurial allows this ... | 14:41 |
|
bushwakko
| what is that? | 14:42 |
| → suman joined | 14:42 |
|
designs703
| providing an identity file for an ssh connection manully | 14:42 |
|
| *manually | 14:42 |
| → FunkeeMonk joined | 14:42 |
| dwmw2_gone → dwmw2 | 14:44 |
|
designs703
| I think I will be able to do it this way: http://git-scm.org/gitwiki/GitTips#Howtopasssshoptionsingit.3F | 14:44 |
| → segher joined | 14:44 |
| → squentin joined | 14:44 |
| ← Turias left | 14:45 |
| → fsalgo joined | 14:45 |
| → Turias joined | 14:45 |
| → JoeConyers joined | 14:46 |
|
mattayers
| Hi guys, I've got a branch with 4 commits I want but not the last one. Can I just merge in the SHA IDs of all the other commits? | 14:46 |
| → JohnFlux joined | 14:47 |
| → bdiego joined | 14:47 |
|
JohnFlux
| Hey all | 14:47 |
|
fsalgo
| hey guys! I want to commit a default database config file but want to ignore it in future commits when I have keyed in the actual credentials | 14:47 |
|
JohnFlux
| how do I do: git stash filename ? | 14:47 |
| ← bdiego_ left | 14:48 |
|
teuf
| mattayers: git merge branchname^ maybe ? | 14:48 |
|
fsalgo
| so the problem is even after putting the file in .gitignore it still shows up at modified | 14:48 |
| → johan-s joined | 14:49 |
| → smuf joined | 14:49 |
|
Thumper_
| j31 | 14:50 |
|
JohnFlux
| fsalgo: I don't think you can do what you want | 14:50 |
|
| fsalgo: it's a pretty common question in here | 14:50 |
|
fsalgo
| JohnFlux: how do you guys go about it? | 14:51 |
| → SjB joined | 14:51 |
|
fsalgo
| JohnFlux: I mean whats the git approach for it? | 14:51 |
|
JohnFlux
| fsalgo: put the credentials in a different, untracked, file? | 14:52 |
|
| and include that file? | 14:52 |
|
hgb21
| reto__: so what is the problem ? | 14:52 |
| → alezandro joined | 14:52 |
| ← Mezner left | 14:53 |
|
fsalgo
| JohnFlux: but then in other file which I'll commit will have reference to the difference file! | 14:53 |
|
designs703
| the technique described here is not working for me, has anyone here had success with it? http://git-scm.org/gitwiki/GitTips#Howtopasssshoptionsingit.3F | 14:54 |
| → sergiodj_ joined | 14:54 |
| ← x-demon left | 14:54 |
|
mattayers
| Is there anyway to merge a commit in without doing it's parents? | 14:54 |
|
fsalgo
| JohnFlux: cant one just start ignoring a file midway? | 14:55 |
| ← sergiodj left | 14:55 |
|
designs703
| I am able to connect when providing the key to ssh with the -i flag and I've written this script with that key provided via an absolute path. I get prompted for a password when connecting to the server after setting GIT_SSH to this env var | 14:55 |
|
fsalgo
| as we have migrated from svn to git and now for future commits we'd like to ignore few files | 14:55 |
|
jjuran
| mattayers: Not really. You'd have to create a new commit without the unwanted parents. | 14:55 |
|
mattayers
| jjuran: Ok thank you | 14:56 |
|
jjuran
| fsalgo: Git doesn't ignore tracked files. If you remove the files, then they'll be ignored if they match .gitignore. | 14:57 |
| ← bgerber left | 14:57 |
| ← sergiodj_ left | 14:58 |
| → sergiodj_ joined | 14:58 |
|
fsalgo
| jjuran: yes I read that git doesnt ignore tracked files. so do you recommend that I make a sample config file commit it, delete it and then create it with actual credentials and start ignoring it? | 14:59 |
| → Haarg joined | 15:00 |
| → nothingHappens joined | 15:00 |
|
fsalgo
| jjuran: as the file was deleted, ignored and recreated will git still track it? | 15:00 |
| ← spearce left | 15:00 |
| → x-demon joined | 15:02 |
|
fsalgo
| jjuran: yes that works :D | 15:02 |
|
| jjuran: JohnFlux: thanks guys | 15:02 |
| → malumalu joined | 15:03 |
| ← abms1116 left | 15:03 |
| → babilen joined | 15:03 |
| → abms1116 joined | 15:03 |
| ← sde_antrix2 left | 15:03 |
| ← abms1116 left | 15:04 |
| ← dpino left | 15:05 |
| ← Axius left | 15:07 |
| → priidu joined | 15:08 |
| → drubies joined | 15:08 |
| → DrTrevorkian joined | 15:08 |
| ← johan-s left | 15:09 |
| gionn_ → gionn | 15:10 |
| → sde_antrix2 joined | 15:12 |
| → ihoka joined | 15:12 |
| → macmaN6789 joined | 15:13 |
| → _jason553839 joined | 15:14 |
| ← javatexan left | 15:14 |
|
macmaN6789
| does anyone use export_auth_hook in their gitweb.conf here? | 15:15 |
| ← yaotti left | 15:15 |
| → javatexan joined | 15:16 |
| → Mezner joined | 15:16 |
|
macmaN6789
| searching through gitlog it would appear export_auth_hook is not a popular topic | 15:16 |
| → evl joined | 15:17 |
| → bgerber joined | 15:18 |
| ← ph^ left | 15:18 |
| → wolog joined | 15:18 |
| → spearce joined | 15:18 |
| → cehteh joined | 15:18 |
| → mrlemao joined | 15:18 |
|
sitaram
| never heard of it, but a quick read suggests it should be quite straight forward, assuming you got gitweb itself running ok | 15:18 |
| ← designs703 left | 15:19 |
|
macmaN6789
| conceptually yes | 15:20 |
| ← mrlemao left | 15:20 |
|
macmaN6789
| just something wack happens when you put simply return 0; for the hook | 15:20 |
| → USA_NO_1_ joined | 15:21 |
| → wwalker_ joined | 15:21 |
|
macmaN6789
| something somewhere spits out some extra garbage after the normal "no projects found" gitweb page | 15:21 |
| ← Grahack left | 15:21 |
|
macmaN6789
| is there a git issue tracker somewhere even? | 15:21 |
|
hyperair
| i believe it's the mailing list. | 15:21 |
| ← evl left | 15:21 |
|
macmaN6789
| thought so | 15:22 |
| → paltman94 joined | 15:22 |
| → Cro_Crx joined | 15:23 |
| → mithro joined | 15:23 |
|
hgb21
| is there a way of changing the password of my remote git server | 15:23 |
|
| I forgotten | 15:24 |
|
hyperair
| ssh? | 15:24 |
| ← a389742 left | 15:24 |
|
hyperair
| ssh <user>@<host> passwd | 15:24 |
| → johan-s joined | 15:24 |
| ← sergiodj_ left | 15:24 |
| → sergiodj_ joined | 15:25 |
|
hgb21
| hyperair: well I have forgotten my old password | 15:26 |
| sergiodj_ → sergiodj | 15:26 |
|
hyperair
| well ask the administrator of that server to reset it | 15:26 |
| → gnukid joined | 15:26 |
| ← ia left | 15:27 |
| ← kpreid_ left | 15:27 |
| → kpreid_ joined | 15:27 |
|
hgb21
| I am the admin . . . | 15:27 |
| ← solofight left | 15:27 |
|
wwalker_
| http://gist.github.com/218730 - I fixed my first git conflict, the command says run add, I did, it still says run add. What do I actually need to do to clear the conflict? <repeating as I think my client went crazy, my nick changed when I posted> | 15:27 |
| ← wolog left | 15:28 |
| → wolog joined | 15:28 |
| → hobodave joined | 15:29 |
| ← Cro_Crx left | 15:29 |
| → Cro_Crx joined | 15:29 |
|
hgb21
| hyperair: ? | 15:29 |
|
hyperair
| hgb21: ...can you access the server at all? | 15:30 |
|
hgb21
| yes | 15:30 |
| → fhr joined | 15:30 |
|
hgb21
| I am loged in | 15:30 |
|
hyperair
| right | 15:30 |
| wwalker_ → wwalker | 15:30 |
|
hyperair
| get root access | 15:30 |
|
| can you get root access? | 15:30 |
| → slaydragon joined | 15:31 |
| ← roop__ left | 15:31 |
|
hgb21
| yes | 15:32 |
| ← qrush_ left | 15:32 |
|
hgb21
| and I did -- su git | 15:32 |
| → qrush_ joined | 15:32 |
|
hgb21
| it asked me for password but I dont remmeber it. | 15:32 |
|
lucs
| hunter2 | 15:32 |
| ← USA_NO_1 left | 15:33 |
|
sitaram
| sck-thebs | 15:33 |
|
| oops! | 15:33 |
|
Ilari
| hgb21: Do 'su git' from root and it won't prompt for password. | 15:33 |
| ← Cro_Crx left | 15:34 |
| → selckin joined | 15:35 |
| → Cro_Crx joined | 15:35 |
|
Ilari
| hgb21: Also, 'passwd git' as root should work to change the password. | 15:35 |
| → ia joined | 15:37 |
|
hgb21
| Ilari: I got this now Cannot execute /usr/bin/git-shell : No such file or directory | 15:37 |
| ← hgb21 left | 15:39 |
|
Ilari
| hgb21: Well, 'passwd git' as root should work. | 15:39 |
| ← _jason553839 left | 15:39 |
| ← Adlai left | 15:39 |
| ← maleknet left | 15:40 |
| ← n3m left | 15:44 |
| → hgb21 joined | 15:44 |
|
hgb21
| and i have it insalled | 15:45 |
| ← asksol left | 15:45 |
| ← Cro_Crx left | 15:45 |
| → abms1116 joined | 15:46 |
| → _Mica_ joined | 15:47 |
| → evl joined | 15:47 |
| ← qrush_ left | 15:47 |
| → __iron joined | 15:48 |
| → Adlai joined | 15:49 |
| ← madewokherd left | 15:49 |
| ← aatanasov left | 15:50 |
| → omenking_ joined | 15:51 |
| ← Adlai left | 15:51 |
|
hgb21
| Ilari: from root ...when I do "su git" I get this | 15:51 |
|
| annot execute /usr/bin/git-shell : No such file or directory | 15:51 |
| → Adlai joined | 15:52 |
|
doener
| wwalker: you resolved the conflict to "no changes", so there's nothing to be committed. Use "git rebase --skip" to skip the commit | 15:52 |
|
Ilari
| hgb21: Also, 'sudo -H -u git /bin/bash' might work to get shell as git user. | 15:53 |
| ← EricInBNE left | 15:53 |
| ← FunkeeMonk left | 15:53 |
|
hgb21
| Ilari: yest that works | 15:53 |
| → Cro_Crx joined | 15:54 |
|
hgb21
| It still ask me to provide my old password when I attempt to change it | 15:55 |
| ← Cro_Crx left | 15:55 |
|
Ilari
| hgb21: 'passwd git' as root. It does not ask for old password. | 15:56 |
| → Cro_Crx joined | 15:56 |
| ← fsalgo left | 15:56 |
| ← bushwakko left | 15:58 |
| → _jason553839 joined | 16:00 |
| ← kpreid_ left | 16:00 |
| ← sde_antrix2 left | 16:00 |
| ← saurabhverma left | 16:00 |
| ← omenking left | 16:01 |
| ← marc_kdab left | 16:01 |
| → Theravadan joined | 16:02 |
| → marc_kdab joined | 16:02 |
| → helo joined | 16:02 |
| ← univate left | 16:02 |
|
| sitaram idly wonders how many people are being inducted into the magic of the Unix command line because of git | 16:02 |
| ← gambler left | 16:03 |
|
Adlai
| sitaram: git was the first command-line app that I really used on the command-line | 16:03 |
|
sitaram
| bingo! score 1 for my hypothesis :) | 16:03 |
|
| Adlai started out half a year ago on the Jaunty Jackalope | 16:03 |
|
| helo aggressively wonders how many more people are avoiding git because they are afraid of the Unix command line voodoo | 16:03 |
|
| Adlai now tiles stumpwm on Archlinux :D | 16:03 |
|
| sitaram idly googles stumpwm | 16:04 |
|
sitaram
| (( ok )) | 16:04 |
|
Adlai
| hehe | 16:04 |
|
sitaram
| (( got (it ))) | 16:04 |
|
hgb21
| Ilari: I still have problem when I attempt to push | 16:04 |
|
| http://gist.github.com/218695 | 16:04 |
|
Adlai
| sitaram: It's definitely a niche WM... | 16:05 |
|
hgb21
| the new password is not been recognised . | 16:05 |
|
Ilari
| hgb21: Maybe ssh settings don't allow remote login as git? | 16:05 |
| ← nevyn_ left | 16:05 |
|
hgb21
| what do you mean ? | 16:06 |
|
Ilari
| hgb21: If you forbid ssh logins to some account, ssh behaves like all authentications fail. Thus you get password prompt which doesn't accept any password. | 16:07 |
| ← warthog9 left | 16:07 |
| → alley_cat joined | 16:07 |
| ← wolog left | 16:07 |
|
hgb21
| yes I have restricted the ssh to the remote git | 16:07 |
|
Ilari
| hgb21: Anyway, this is ssh problem, not git problem. | 16:07 |
|
hgb21
| Ilari: But how can I enable that / | 16:08 |
|
Ilari
| hgb21: sshd configuration file controls that. | 16:08 |
| → _Cro_Crx joined | 16:09 |
| → asksol joined | 16:09 |
| → qhoxie joined | 16:09 |
| → tbuehlmann joined | 16:10 |
| → omenking joined | 16:10 |
| ← johan-s left | 16:11 |
| ← hakunin left | 16:11 |
| → justin-george joined | 16:11 |
| ← justin-george left | 16:11 |
|
sitaram
| /etc/ssh/sshd_config; look for AllowUsers (case insensitive) | 16:12 |
|
Ilari
| Isn't There's also allowGroups ? | 16:12 |
|
sitaram
| I think so; yes. Although I only ever use AllowUsers | 16:13 |
| ← LRN left | 16:13 |
| ← abms1116 left | 16:14 |
| → hakunin joined | 16:16 |
| ← _Cro_Crx left | 16:17 |
| ← omenking_ left | 16:18 |
| → dogmatic69 joined | 16:18 |
|
wwalker
| doener: thank you | 16:18 |
|
dogmatic69
| hi all... im trying to remove a file from a repo, could someone help me | 16:18 |
| → _Cro_Crx joined | 16:18 |
|
dogmatic69
| i tried rm filename.ext and commit --amend | 16:19 |
| → nikki93 joined | 16:19 |
|
Ilari
| dogmatic69: 'git rm filename.extension' and then commit. | 16:20 |
|
nikki93
| Hey guys! Say there's this little experimental change I want to do. It's a small change, and it wont conflict with many things. I make the change, then I continue working on the project. Say in the future I want to undo just that change, how should I do this? | 16:20 |
| ← j416 left | 16:20 |
|
nikki93
| Make a new branch for that change? But, the things I do after the change are normal, just that particular change is experimental. | 16:20 |
|
dogmatic69
| Ilari: that is what i did.. is it the --amend | 16:21 |
|
nikki93
| I will stick to it most probably anyway, but I would like to be able to choose whether I can 'turn it off'. | 16:21 |
| → kennethlove joined | 16:21 |
|
selckin
| branch sounds reasonable then | 16:21 |
|
Ilari
| dogmatic69: --amend does handle file removals too. | 16:22 |
| ← paltman94 left | 16:22 |
|
charon
| dogmatic69: the 'git' in 'git rm' is important. | 16:22 |
|
nikki93
| selckin: So I make a new branch for it, and then I should work on future things on that branch, or on master? | 16:22 |
| → spaceonline joined | 16:22 |
| ← WALoeIII left | 16:22 |
|
dogmatic69
| charon: sorry i had the git also ... all the commands start with git afaik | 16:22 |
| ← Cro_Crx left | 16:23 |
| _Cro_Crx → Cro_Crx | 16:23 |
| → godsflaw joined | 16:23 |
| → qrush_ joined | 16:23 |
| → krawek joined | 16:23 |
| ← giallu left | 16:24 |
| ← _jason553839 left | 16:24 |
| → BeerSerc joined | 16:25 |
|
jgoulah
| whats the best way to send an email on push (ideally with what was pushed) | 16:25 |
|
charon
| dogmatic69: in that case, you did everything right, and should probably describe what you expected to happen and how git disagreed with that | 16:25 |
|
| well, i wouldn't amend though | 16:25 |
|
dogmatic69
| charon: http://github.com/dogmatic/cake-currency-converter i want that loose file gone | 16:25 |
|
Tanguy
| Hi. | 16:26 |
| → juan_arandaalvar joined | 16:26 |
| ← froschi left | 16:27 |
|
Tanguy
| Do you know if a merge tool can be invoked with the true original file as middle file, without the conflicts markers? | 16:27 |
| → murph_ joined | 16:27 |
|
charon
| dogmatic69: you didn't push, or the push failed because you amended | 16:27 |
|
dogmatic69
| ok, let me try again | 16:27 |
| ← Jinzhu left | 16:28 |
|
dogmatic69
| now its saying its not there me git rm currency.php | 16:28 |
|
nikki93
| jgoulah: I believe you can add 'hooks' to events. | 16:29 |
|
selckin
| Tanguy: git mergetool with a diff tool configured that does three way merges, or do i misunderstand | 16:29 |
|
jgoulah
| well i saw this, just don't know where to put it http://git.kernel.org/?p=git/git.git;a=blob;f=contrib/hooks/post-receive-email;h=60cbab65d3f8230be3041a13fac2fd9f9b3018d5;hb=HEAD | 16:30 |
|
| i have a gitosis setup | 16:30 |
|
Gitbot
| [git 60cbab65d]: http://tinyurl.com/yl9f3g3 [blob] | 16:30 |
|
charon
| dogmatic69: do the following: (1) run 'git reflog' and verify that the first (topmost) entry says 'commit (amend)'; (2) having verified that, see if you have any uncommitted changes with 'git status'; (3) having checked that you don't: run 'git reset --hard HEAD@{1}' to undo the amend; (4) run 'git rm currency.php' again; (5) run 'git commit' without(!) --amend this time | 16:30 |
|
| dogmatic69: if all that went well, do 'git push' | 16:31 |
|
Tanguy
| selckin: I use a mergetool that does three way merges, but that expects a middle file intact, without conflict markers. | 16:31 |
|
charon
| dogmatic69: the reason for all this is: you should never rewrite commits that were already pushed elsewhere, so steps (1-3) undo the --amend. | 16:31 |
| ← gnukid left | 16:31 |
|
Tanguy
| selckin: And git gives it the conflict-marked file instead of the original one. | 16:31 |
| → SRabbelier joined | 16:32 |
|
dogmatic69
| charon: amend is at the top \o/ | 16:32 |
|
selckin
| Tanguy: not sure, never seen conflict marks in mine | 16:32 |
| ← corni left | 16:32 |
|
bentob0x
| eeh: Already up-to-date. Yeeah! | 16:32 |
|
| ? | 16:32 |
| ← evl left | 16:32 |
| sdboyer → agentrickards_lu | 16:33 |
|
dogmatic69
| charon: and it says no changes added to commit | 16:33 |
| agentrickards_lu → sdboyer | 16:33 |
|
SRabbelier
| Ilari, madduck: could you s/"Alles wird Git" is over!/GitTogether 2009 in progress/ the channel topic? | 16:33 |
|
charon
| bentob0x: iirc that's the message for an octopus merge of branches that are already contained | 16:33 |
| → Axius joined | 16:34 |
|
bentob0x
| charon: the 'Yeeah' ? | 16:34 |
| ← murph left | 16:34 |
|
charon
| indeed. the up-to-date message is different for ordinary and octopus merges | 16:34 |
|
bentob0x
| is the 'Yeeah' random? | 16:35 |
|
charon
| i just told you it's deterministic :) | 16:35 |
| Ilari changed the topic to: 1.6.5.2 | Homepage: http://git-scm.com | Everyone asleep or clueless? Try [email@hidden.address] | Git User's Survey 2009! http://git.or.cz/gitwiki/GitSurvey2009 | Channel log http://tinyurl.com/gitlog | Mailing list archives: http://tinyurl.com/gitml | Gits on git: http://tinyurl.com/gittalks | Pastebin: http://gist.github.com/ | GitTogether 2009 in progress | 16:35 |
|
Tanguy
| selckin: I think the “meld” definition is just wrong, I will try an modify git-mergetool to see what it does. | 16:35 |
|
dogmatic69
| charon: omg thanks :) | 16:36 |
|
| got it to disapear | 16:36 |
|
nikki93
| So I'd like to make the experimental change, stick with it, but if I have problems, just undo only that. Is a branch the way? | 16:36 |
|
Ilari
| dogmatic69: Oh yeah, there's IIRC delay in github web interface... | 16:37 |
| → Qwindoo joined | 16:37 |
|
dogmatic69
| Ilari: i tried yesterday | 16:37 |
| → ferric joined | 16:37 |
|
charon
| dogmatic69: ok. please take away the following lesson: don't ever use --amend, rebase or filter-branch on commits that were already pushed. | 16:37 |
|
dogmatic69
| charon: i think the lesson is dont copy past crap from the net :) | 16:37 |
| ← Theravadan left | 16:37 |
|
charon
| that too ;) | 16:38 |
|
dogmatic69
| its the first day im using this :) | 16:38 |
|
| properly | 16:38 |
|
ferric
| question, I have a local branch that references a submodule in a commit. the submodule was subsequently pulled into the repo on an upstream branch. how do i get my local branch to understand this? rebase complains. | 16:38 |
|
| s/an/the/ | 16:38 |
| → d3z joined | 16:38 |
| → _jason553839 joined | 16:38 |
| → ab3 joined | 16:39 |
| → univate joined | 16:39 |
| → gambler joined | 16:40 |
| ← Qwindoo left | 16:40 |
| → peff joined | 16:40 |
| → WALoeIII joined | 16:40 |
|
bentob0x
| thx charon :) | 16:41 |
|
Ilari
| ferric: Deleteting the submodule locally would probably at least allow the rebase to start... | 16:41 |
| → icwiener joined | 16:41 |
| → FunkeeMonk joined | 16:41 |
|
JohnFlux
| how do I do a git merge --continue, so to speak? | 16:41 |
|
charon
| nikki93: depends how confident you are; if you feel it has little chances of surviving, make a separate topic branch and use throw-away merges to test the interaction with other branches. if you think it has good chances of staying, just make the commit and if it turns out to be a bad idea, revert it | 16:41 |
|
ferric
| Ilari: cool, I'll try that -- and then probably redo my last commit since I just noticed that it is the only commit that references that submodule | 16:41 |
| ← Pupeno left | 16:42 |
|
charon
| JohnFlux: add the resolved files and then commit. git-merge told you when it complained about conflicts! | 16:42 |
|
Tanguy
| selckin: In fact, it seems that meld is just unable to save into a file different than the original one, thus it is incompatible with git-mergetool's behaviour (four file names: original, local, remote, merged). | 16:43 |
| ← ntoll left | 16:43 |
| ← bentob0x left | 16:43 |
|
nikki93
| charon: Well it ain't that bad, but the thing is, I might want to just 'try it without it' even a few commits later. | 16:44 |
| ← dogmatic69 left | 16:44 |
|
charon
| nikki93: no harm there; you can always say 'git checkout HEAD^0; git revert $change' to do a "test revert" in the second case. remember to 'git checkout $branch' afterwards to re-attach your HEAD | 16:45 |
| ← jchris left | 16:45 |
| ← univate left | 16:46 |
|
nikki93
| charon: I see. | 16:46 |
|
selckin
| Tanguy: maybe you can write a wrapper script to move things around | 16:46 |
| → univate joined | 16:46 |
| ← Mezner left | 16:47 |
|
Tanguy
| selckin: Yes, I shall do that. | 16:47 |
| ← rolfb left | 16:49 |
| ← rraasch left | 16:51 |
| → jaysern joined | 16:51 |
| → LRN joined | 16:53 |
| → adamm joined | 16:53 |
| → kumbayo_ joined | 16:54 |
| ← cantoma left | 16:54 |
|
nikki93
| charon: Is it possible to 'unapply' an older commit, on a current one? | 16:55 |
|
| charon: Without changing the history. | 16:55 |
|
| (by making a new commit that does the reverse) | 16:55 |
|
Mikachu
| nikki93: git revert | 16:56 |
|
charon
| nikki93: that's precisely what 'git revert $commit' does | 16:56 |
| ← mastro left | 16:56 |
|
nikki93
| Yay. :D | 16:56 |
|
| I thought git revert was only for the latest commit. | 16:56 |
| ← crankharder left | 16:57 |
|
nikki93
| So... Any git developers here? | 16:57 |
| ← Cro_Crx left | 16:57 |
|
sitaram
| lots, I'm sure... | 16:57 |
| ← khmarbaise left | 16:58 |
|
d3z
| Although the gittogether is happening right now. | 16:58 |
|
nikki93
| charon: And if I want to reapply that older commit after reverting it? :P | 16:58 |
| ← hgb21 left | 16:58 |
|
Ilari
| nikki93: Revert the revert? | 16:58 |
|
nikki93
| Cool. | 16:58 |
| ← bdiego left | 16:59 |
| Thell|Away → Thell | 16:59 |
|
charon
| though you may want to make up your mind at some point :) | 17:00 |
|
nikki93
| Well most probably this change will stay. | 17:00 |
| ← kumbayo left | 17:02 |
| → lhz joined | 17:04 |
| ← crowbar left | 17:06 |
| → bushwakko joined | 17:06 |
| → eletuchy joined | 17:07 |
| ← slaydragon left | 17:07 |
|
jgoulah
| is there a way to configure git so i don't have to do every command from the root dir? | 17:07 |
| → Sigma joined | 17:07 |
|
selckin
| thats the default? | 17:07 |
| ← name left | 17:08 |
|
jgoulah
| if im in a subdir it doesn't know what i'm talkin bout | 17:08 |
| → dfr|work joined | 17:08 |
|
jgoulah
| says stuff about untracked files | 17:09 |
| ← _jason553839 left | 17:09 |
|
helo
| i'm considering using git to keep track of and push configuration/application updates out to remote servers... anyone have experience with this? | 17:09 |
|
charon
| jgoulah: which command(s) in which usage? | 17:09 |
|
jgoulah
| git status , for example | 17:10 |
|
charon
| ah but that's one of the few that work relative to the current directory... | 17:11 |
| ← mattayers left | 17:11 |
|
jgoulah
| ok | 17:12 |
|
charon
| this is admittedly not very consistent between commands | 17:12 |
| ← Lemon|work left | 17:13 |
|
| sitaram *always* works at the root; too confusing otherwise (probably my fault for not researching enough to know which commands work and which don't) | 17:14 |
|
charon
| what _is_ consistent, however, is that commands that do not work with the worktree at all are always(?) relative to the repo root | 17:14 |
| ← ciskje left | 17:14 |
|
sitaram
| _is_ and (?) dont got together :) | 17:14 |
|
charon
| damn you caught me | 17:14 |
|
sitaram
| go* | 17:14 |
|
| :) | 17:14 |
| ← FunkeeMonk left | 17:14 |
|
charon
| i put the (?) because by the time i got to writing "always", i remembered git-reset isn't | 17:15 |
|
| but then, it works with worktree files too if you tell it to | 17:15 |
|
sitaram
| so it doesnt count | 17:15 |
| → niki joined | 17:15 |
|
jgoulah
| yeah but i am converting people from svk and was hoping to reduce confusion since in there it doesn't much matter where you are, no worries tho | 17:15 |
|
sitaram
| s/do not work with.../never work with.../ maybe? | 17:15 |
| → FunkeeMonk joined | 17:16 |
|
charon
| hrm, well not that either. *sigh* | 17:17 |
|
| 'git log -- $file' also resolves the $file first | 17:17 |
|
| i guess the git UI design is: DWIM, or when in doubt, DWLM | 17:18 |
| → Bass10 joined | 17:19 |
| ← Axius left | 17:20 |
| → froschi joined | 17:20 |
|
sitaram
| :) | 17:20 |
| ← priidu left | 17:20 |
|
sitaram
| or rather, :( | 17:20 |
| → _jason553839 joined | 17:21 |
|
jefferai
| If you have a feature branch that you want to merge into mainline, but that you also want to be able to revert if something turns out to e broken, what is the best way to do it? | 17:22 |
|
| I know you can rebase it into one commit and apply, but that loses history | 17:22 |
|
| but it's also not nice to have to revert a ton of commits | 17:22 |
|
| is there a possibility of simply reverting the merge commit that merged the branches? | 17:22 |
| → evl joined | 17:24 |
| → aliceinwire joined | 17:24 |
| → elTigre joined | 17:24 |
| → joevandyk_ joined | 17:25 |
|
joevandyk_
| git push HEAD~2:origin does not work. what would the syntax for that be? | 17:25 |
|
| i want to push out to the origin branch, but i don't want to include the last two commits in the push | 17:25 |
|
numbah
| git push origin HEAD~2:master | 17:26 |
| → raalkml joined | 17:26 |
|
sitaram
| jefferai: if you push somewhere after the merge, doesn't matter how you do it; a revert is the only way. But you can revert many commits in one revert if you like | 17:26 |
|
| (not by the revert command) | 17:26 |
| ← jmcantrell left | 17:26 |
|
jefferai
| sitaram: err, sorry -- you can revert many commits in one revert, but not by the revert command? | 17:27 |
|
raalkml
| he means: revert takes only one commit in its argument | 17:27 |
| → d2dchat joined | 17:28 |
| → catphish joined | 17:29 |
|
jefferai
| so how do you revert a range of commits? | 17:29 |
| ← godsflaw left | 17:29 |
| ← _jason553839 left | 17:29 |
|
sitaram
| jefferai: it involves an incantation using git reset twice; let me think... | 17:30 |
|
jefferai
| ah | 17:30 |
|
patrikf
| jefferai: you can use git revert to revert commit-by-commit, possibly using "-n" to have all the reverts happen in one commit. | 17:30 |
|
| alternatively, if your history isn't yet public, you can use git reset | 17:31 |
|
jefferai
| sitaram: also, when you merge, the history of the branch merges into the log, right? so what would be a good way to pull out just the various commits that were in that branch? | 17:31 |
|
patrikf
| or git rebase -i | 17:31 |
|
jefferai
| nha, history will be public already | 17:31 |
|
| :-| | 17:31 |
|
sitaram
| patrikf: I'd say using revert assumes you cant use rebase but not sure if jefferai meant it that way | 17:32 |
|
| (meanign someone pulled already) | 17:32 |
|
jefferai
| people will have pulled already | 17:32 |
|
patrikf
| sitaram: consider that part of my sentence that started with "if your history isn't yet public" | 17:32 |
|
jefferai
| (is a safe assumption) | 17:32 |
| → shaggyoaf joined | 17:33 |
|
shaggyoaf
| Hi, all. I've got two SVN repositories that I want to merge into a single git repository | 17:34 |
|
| I figured out git-svn clone and I can bring them both down | 17:34 |
|
| and I want one repo to contain another | 17:34 |
|
| is there a way I can do that and preserve the svn history? | 17:34 |
|
sitaram
| patrikf: I did; sorry it sounded like I didn't... I meant thats not the use case here | 17:34 |
|
patrikf
| yeah, just mentioned it for completeness of the picture :) | 17:35 |
|
sitaram
| jefferai: git reset --hard HEAD~3; git reset --soft HEAD@{1}; now make new commit | 17:35 |
|
| jefferai: assuming top 3 to be reverted | 17:35 |
| ← altrux left | 17:35 |
|
jefferai
| yeah, that's not a good assumption either | 17:35 |
|
| if we have to revert this, it's likely to be weeks from now | 17:35 |
|
| (and many commits on top) | 17:35 |
| → altrux joined | 17:36 |
|
sitaram
| jefferai: then the "-n" method someone mentioned has to be explored | 17:36 |
| → mbroeker joined | 17:36 |
| → godsflaw joined | 17:37 |
|
sitaram
| jefferai: it does work (just tested) but make sure you do them in reverse order. And watch for conflicts | 17:37 |
| → johan-s joined | 17:37 |
|
jefferai
| sitaram: what does work? | 17:38 |
|
| you didn't actually paste your -n foo :-) | 17:38 |
| ← Lisimba left | 17:38 |
|
sitaram
| jefferai: yup; sorry. So you want to revert HEAD~2, HEAD~3, and HEAD~4. You do "git revert -n HEAD~2" then 3 then 4; then commit with some suitable message | 17:39 |
|
| the first 3 reverts are with -n | 17:39 |
| ← raalkml left | 17:39 |
|
sitaram
| s/first// | 17:39 |
|
| s/someone/patrikf/ too I guess -- sorry again! just doing too many things at the same time *and* getting old :( | 17:40 |
| ← shruggar left | 17:40 |
|
jefferai
| sitaram: OK, I see. But, when you merge the histories of the branches are merged together right? | 17:41 |
|
| so is there a good way to find that list of revisions that constituted that branch? | 17:41 |
| ← kmamykin left | 17:42 |
| → Mezner joined | 17:43 |
|
sitaram
| jefferai: eh? where did a branch get in here? | 17:43 |
|
| sitaram checks backlog | 17:43 |
|
sitaram
| aah | 17:43 |
|
| jefferai: when you find that something you merged was bad, you do this in the original branch you merged into mainline, then merge again | 17:44 |
|
jefferai
| wait, you do the revert in the original branch? | 17:45 |
|
| that doesnt' make sense | 17:45 |
|
sitaram
| and now that I read your original question again, yes you can simply revert the merge commit if you wish; play with "-m 1" or "-m 2" | 17:45 |
| ← cedricv left | 17:45 |
|
sitaram
| jefferai: why doesnt it make sense? the problem was caused there; you fix it there and propagate the fix | 17:46 |
|
peff
| sitaram: reverting merges is a bit dangerous, though | 17:46 |
|
sitaram
| indeed | 17:47 |
|
peff
| If you later want to remerge, the history doesn't quite do what you expect. | 17:47 |
|
jefferai
| sitaram: no, the problem wasnt' caused there | 17:47 |
|
| the problem would be merging that branch into mainline | 17:47 |
|
| like, if we want to revert, we will want to revert the entire merge | 17:47 |
|
| not specific revisions in the merged-in branch | 17:47 |
| → ryoma joined | 17:48 |
| → jceb joined | 17:48 |
|
sitaram
| jefferai: try -m 1 or -m 2 on the revert | 17:49 |
|
jefferai
| sitaram: OK -- so, if I wanted to revert merging that branch, I could go to the merge commit, then could do -m X ... | 17:50 |
| ← fhr left | 17:50 |
|
jefferai
| I see in the man-page the references to revert-a-faulty-merge | 17:50 |
|
| I'll look at it | 17:50 |
| → fhr joined | 17:50 |
| → warty9_andy joined | 17:51 |
| → priidu joined | 17:51 |
|
warty9_andy
| anyone seen sop / spearce? | 17:52 |
| → Pupeno joined | 17:52 |
| → kpreid_ joined | 17:52 |
| ← engla left | 17:53 |
|
peff
| warty9: yeah, he was sitting right in front of me a minute ago | 17:53 |
|
bnovc
| how can I create a remote branch without creating a local branch and committing first? | 17:54 |
|
| i have servername/foo-devel and I want servername/bar based on foo-devel | 17:54 |
|
peff
| bnovc: you can't. A branch exists only as a pointer to some commit. If you have no commits, you have no branch. | 17:54 |
| ← kpreid_ left | 17:54 |
|
bnovc
| there are commits, but not in that branch i mean | 17:54 |
|
peff
| Ah. | 17:54 |
|
bnovc
| i just want it to point to HEAD of foo-devel | 17:55 |
| → Theravadan joined | 17:55 |
| → _jason553839 joined | 17:55 |
|
Ilari
| bnovc: 'git push servername servername/foo-devel:refs/heads/bar'. | 17:55 |
|
bnovc
| thanks, ill try that | 17:55 |
|
peff
| Ilari: that's right, though I think you can drop the refs/heads in recent git | 17:55 |
|
bnovc
| thats what i was going to ask about | 17:55 |
|
Ilari
| peff: No, you can't. | 17:55 |
|
peff
| Ilari: I thought we added automagic detection of branch type for remote branches as well as local | 17:56 |
|
bnovc
| why is that necessary/what does it significy? | 17:56 |
|
| *signify | 17:56 |
|
Ilari
| peff: Unless "recent" means more recent than master few days ago. | 17:56 |
| ← zoniq left | 17:56 |
|
peff
| Ilari: recent may mean "in my repo and I haven't submitted to the list yet" ;) | 17:56 |
| ← perez left | 17:56 |
| ← smuf left | 17:56 |
|
bnovc
| heh | 17:56 |
| ← FunkeeMonk left | 17:57 |
| ← Pupeno left | 17:57 |
| → kpreid_ joined | 17:57 |
|
sitaram
| Ilari: I routinely do git push remote foo:foo without using ...;refs/heads/foo -- what am I missing? | 17:57 |
|
| s/...;/...: | 17:58 |
| → manithree joined | 17:58 |
| → Pupeno joined | 17:58 |
|
peff
| sitaram: we detect the type of "foo" and say "oh, it's a branch. The other side is a branch, too" | 17:58 |
| ← manithree left | 17:58 |
| → manithree joined | 17:58 |
|
peff
| sitaram: but in this example, it is actually pushing refs/remotes/server/foo-devel | 17:58 |
| → breitz joined | 17:58 |
| ← warty9_andy left | 17:59 |
|
sitaram
| oh in this case he wants to create a remote branch based on anothe remote branch ? | 17:59 |
|
Ilari
| Yup, not even current pu would handle it right without refs/heads... | 17:59 |
|
| Create remote branch based on remote tracking branch... | 17:59 |
| → PeakerWork joined | 17:59 |
| → vital joined | 18:00 |
|
sitaram
| ok | 18:00 |
|
PeakerWork
| how do I get a THREE way merge in mergetool? git mergetool only shows me the local/remote ver, but not any base | 18:01 |
|
sitaram
| I see the difference | 18:01 |
|
| PeakerWork: depends on which tool you use; kdiff3 shows all 3 at least, so does meld | 18:01 |
|
doener
| sitaram: "fetch ... bla:foo" turns foo into refs/heads/foo, but "push ... bla:foo" failes | 18:01 |
|
sitaram
| actually they all should | 18:01 |
|
doener
| s/les$/ls/ | 18:01 |
|
PeakerWork
| sitaram: I use git mergetool -t meld - it shows me my local and remote on the left/right. in the middle I get the merged result, but not a base | 18:02 |
|
| sitaram: a 3-way merge typically has 4 views. base, local, remote, and result | 18:02 |
|
sitaram
| PeakerWork: meld mutiplexes; try kdiff3 -- it will show you 3 on top, and the result across the bottom | 18:02 |
| ← jceb left | 18:02 |
| ← priidu left | 18:02 |
|
sitaram
| doener: I could have sworn I did git push origin foo:bar once or twice | 18:02 |
|
PeakerWork
| sitaram: you mean the result pane has the base in it? | 18:03 |
|
sitaram
| PeakerWork: yes | 18:03 |
|
| PeakerWork: no; wait the top row has 3 subpanes: base, local, remote | 18:03 |
|
PeakerWork
| sitaram: it seems to have the <<< and === signs from the file -- not the base | 18:03 |
|
sitaram
| PeakerWork: the bottom row is just one file; the result | 18:03 |
| → patmaddox joined | 18:04 |
|
doener
| sitaram: sorry, forgot to add "when 'foo' doesn't exist yet" | 18:04 |
|
PeakerWork
| sitaram: you talking about kdiff3 or meld? I don't think I have kdiff3 installed | 18:04 |
|
sitaram
| PeakerWork: kdiff3 shows <<< and ===? when invoked with "git mergetool -t kdiff3"? this is new to me... I use it all the time and never saw them | 18:04 |
|
PeakerWork
| sitaram: I was talking about meld there | 18:04 |
| → gunmuc joined | 18:04 |
|
sitaram
| doener: even so; but my memory isn't that good | 18:04 |
|
doener
| sitaram: and if "bar" is refs/heads/bar locally, then that's a special case, too | 18:04 |
| ← dominikh left | 18:05 |
|
sitaram
| PeakerWork: rethinking... I dont think meld even shows the base, sorry | 18:05 |
|
peff
| Ilari: ok, you're definitely right about the refs/heads thing. Probably we should add that to the DWIM in remote.c:guess_ref | 18:06 |
|
PeakerWork
| sitaram: I wondered how people praised meld so much when it seemed to me to be very sub-par | 18:06 |
|
evl
| when rebase complains a file isn't up to date, what am I missing? | 18:06 |
|
sitaram
| I moved from meld to kdiff3 some time ago :) | 18:06 |
|
patmaddox
| I ran "git log master.." to see all the commits that existing in my branch but not in master. Is there an easy way to generate a patch from it all? Basically I just want to see the contents of those commits | 18:06 |
| → head joined | 18:07 |
| ← _hp_ left | 18:07 |
|
peff
| patmaddox: do you want one patch per commit, or a patch of the whole thing? | 18:07 |
|
doener
| patmaddox: git log -p master.. | 18:07 |
|
patmaddox
| peff: log -p works nicely for showing one patch per commit (thanks doener). Is there a way to see one giant patch too? | 18:08 |
|
doener
| patmaddox: or "git diff master..." (three dots) to get a diff from the common ancestor of master and HEAD, and HEAD (i.e. changes that would be merged to master) | 18:08 |
|
| patmaddox: OTOH "git diff master.." would show a direct diff between master and HEAD, which is unlikely to be what you want (it would show everything that happened on master as being undone) | 18:09 |
| ← bushwakko left | 18:10 |
| → pantsman joined | 18:11 |
|
sitaram
| doener: http://pastebin.com/m3e7709d4 -- what did I misunderstand this time? :( | 18:12 |
| → rolfb joined | 18:12 |
| ← rolfb left | 18:12 |
| → onigiri joined | 18:13 |
|
doener
| sitaram: foo resolved to "refs/heads/foo", so "bar" is assumed to be "refs/heads/bar" | 18:13 |
| → saurabhverma joined | 18:13 |
|
sitaram
| oh... 'when foo doesn't exist yet' (misread that as ...bar...); how can you do anything like that anyway when foo doesnt exist? | 18:14 |
| → Dashkal joined | 18:14 |
|
doener
| sitaram: in my example, "foo" was the destination, not the source | 18:14 |
|
| sitaram: http://pastebin.com/m55f2e22e | 18:14 |
|
sitaram
| doener: but bar, the destination, doesnt exist yet in mine... | 18:15 |
| → dominikh joined | 18:15 |
| → bugfux joined | 18:15 |
| ← tango_ left | 18:15 |
| → engla joined | 18:15 |
| ← catphish left | 18:15 |
|
doener
| sitaram: but the source was resolved to refs/heads/foo, and then the destination is assumed to live in refs/heads/, too | 18:16 |
|
sitaram
| if you'd tried push ~/a refs/heads/master:master it would have worked? | 18:17 |
| ← patmaddox left | 18:17 |
|
doener
| sitaram: yep, as would have "master:master" or just "master" | 18:17 |
| ← manithree left | 18:17 |
| → warthog9 joined | 18:17 |
|
sitaram
| ok | 18:17 |
|
doener
| sitaram: but IIRC only since 1.5.6 or so. Previously, for refs to be created, you always had to use the full refspec | 18:17 |
|
sitaram
| yeah I'm usually at the latest anyway | 18:18 |
|
bnovc
| I have a servername/foo-devel and I used a mercurial to git importer which created a local "master" branch... to include all these chanegs in servername/foo-devel, do I need to make a new branch that tracks it and cherry-pick everything from the "master" branch? | 18:18 |
| ← godsflaw left | 18:18 |
| → Grahack joined | 18:19 |
|
bnovc
| hopefully there is an easier way than that | 18:19 |
|
Ilari
| bnovc: Check out the new local branch and merge master into it? | 18:19 |
| ← mcella left | 18:21 |
|
bnovc
| I guess my real problem is that I don't know how to push directly without uploading all of these to Gerrit first | 18:21 |
|
| maybe I should just do this on the server | 18:21 |
| → manithree joined | 18:21 |
| → godsflaw_ joined | 18:24 |
| ← _Mica_ left | 18:24 |
|
wwalker
| I've just done a git svn rebase I want to see the changes in some of the files. is there a git equivalent to PREV? | 18:24 |
| ← Weasel[DK] left | 18:24 |
| → mattayers joined | 18:24 |
| → Weasel[DK] joined | 18:24 |
|
Ilari
| wwalker: 'branch@{1}'? | 18:25 |
| ← _jason553839 left | 18:26 |
| ← godsflaw_ left | 18:26 |
|
SRabbelier
| GitTogether talk at #gittogether | 18:26 |
|
wwalker
| Ilari, thank you! | 18:27 |
| → corni joined | 18:30 |
| ← galderz left | 18:31 |
| ← LiamH left | 18:31 |
| → trip0 joined | 18:33 |
| → Markus2 joined | 18:33 |
|
Markus2
| www.europa-irc.de | 18:33 |
| ← Markus2 left | 18:34 |
| → godsflaw_ joined | 18:34 |
| → ph^ joined | 18:35 |
| → Axius joined | 18:36 |
| ← wagle left | 18:37 |
| → bdiego joined | 18:37 |
|
jgoulah
| if you are tracking a remote branch and want to push commits, git push won't do? | 18:37 |
| → broquaint joined | 18:37 |
| → patmaddox joined | 18:38 |
|
jgoulah
| i'm getting this: http://gist.github.com/218910 | 18:38 |
| → broquain1 joined | 18:39 |
| ← broquain1 left | 18:40 |
| ← rchavik left | 18:41 |
|
PerlJam
| jgoulah: did you pull before pushing? Perhaps there have been other changes to that branch while you were making your changes. | 18:42 |
|
jgoulah
| didn't pull, but nobody else is working on it atm, it looks the branch was pushed but that master rejected is my confusion | 18:43 |
| ← tfarina left | 18:43 |
|
PerlJam
| jgoulah: the master branch was rejected because there are differences between the local master and the remote master. | 18:44 |
|
| (usually) | 18:44 |
|
jgoulah
| i didn't think it mattered since i was on the branch but yes that is probably true | 18:44 |
| → owen1 joined | 18:44 |
| → priidu joined | 18:44 |
|
jgoulah
| do i pull to fix ? | 18:44 |
|
PerlJam
| it's a start :) | 18:45 |
|
| jgoulah tries | 18:45 |
|
owen1
| how to see changed i made on a file that i already added with git add? | 18:45 |
|
PerlJam
| pull may fix it, but you might have conflicts that need resolving. | 18:45 |
|
| owen1: git diff --cached | 18:45 |
| → _jason553839 joined | 18:45 |
| ← Sho_ left | 18:46 |
|
owen1
| PerlJam: nice, thanks | 18:46 |
|
jgoulah
| PerlJam: http://gist.github.com/218916 | 18:47 |
|
PerlJam
| jgoulah: then next I would suspect that the lineages of your local master and remote master have diverged somehow. | 18:48 |
|
evl
| Is there a nicer way to push a rebase other than -f? | 18:48 |
|
jgoulah
| hmm, i mean really nobody else is working on it, a file was added and removed from master | 18:48 |
|
| i added an email hook on the gitosis setup | 18:48 |
|
| thats about it | 18:49 |
|
PerlJam
| did you do a git commit --amend or git rebase after you pulled at some point? | 18:49 |
|
evl
| It's only me working on the remote git(hub) repo so I didn't think I had to pull before I rebased since last push, was i wrong? | 18:49 |
|
PerlJam
| evl: no. rebasing is not nice. | 18:49 |
| → jfkw joined | 18:50 |
|
evl
| I was trying to combine two commits to make the history look nicer, since the last commit was really just me forgetting to add a file in the previous commit | 18:50 |
| ← kpreid_ left | 18:50 |
| ← godsflaw_ left | 18:50 |
|
PerlJam
| evl: If you rebased commits that were already pushed, then that would cause problems. | 18:50 |
|
evl
| Is there any better way of doing this? | 18:50 |
|
| Other than being a bit more careful before pushing? :) | 18:51 |
|
PerlJam
| evl: no. Be more careful :) | 18:51 |
|
jgoulah
| PerlJam: sorry, no i wouldn't even know to do that i'm pretty newb | 18:51 |
|
evl
| Alright, good enough for me. It looks nice and all but it messes up the github commit log | 18:51 |
| → drewolson joined | 18:53 |
|
PerlJam
| jgoulah: Hmm. Did you make another clone and do some work from that other clone? | 18:53 |
|
jgoulah
| PerlJam: actually more probably did happen on master, i convered from svn earlier, and added a couple dirs to master | 18:53 |
|
| and those weren't added on branch | 18:53 |
| → wagle joined | 18:53 |
|
jgoulah
| but yeah those were added from another clone, then work on this clone was done, but only on the branch | 18:54 |
| ← toscano left | 18:54 |
| → godsflaw joined | 18:54 |
| ← psoo left | 18:54 |
|
PerlJam
| jgoulah: Dunno. I'm leaning towards "you screwed something up", but I'm not sure what at the moment. | 18:56 |
|
jgoulah
| oh god, git can't be -that- easy to screw up can it :) | 18:56 |
|
| not much has really happened | 18:56 |
|
| PerlJam: looks like checkout master; pull; checkout branch; push fixes it (don't even need the push) | 18:57 |
|
| but i dont understand why | 18:57 |
| ← godsflaw left | 18:57 |
|
jgoulah
| do i have to pull master just to push a branch? | 18:57 |
| → godsflaw joined | 18:57 |
|
ryoma
| is there a quick way to merge the last 2 commits into one? or do i have to do an interactive rebase? | 18:58 |
|
PerlJam
| jgoulah: oh. you were on a branch other than master when you did a pull before? | 18:59 |
| ← madduck left | 18:59 |
| → thiago_home joined | 18:59 |
|
PerlJam
| jgoulah: you must have been. It fetched the commits from master but only merged the commits for the branch you were on. | 18:59 |
|
jgoulah
| PerlJam: yep | 18:59 |
|
| i see | 18:59 |
|
| so always pull from master? | 18:59 |
|
PerlJam
| jgoulah: you needed to merge the changes on the master branch before pushing. | 18:59 |
| ← corni left | 19:00 |
| → hummingkid joined | 19:00 |
| → corni joined | 19:00 |
| qrush → qrush|away | 19:00 |
|
PerlJam
| jgoulah: you pull from where ever. pull == fetch + merge fetch will get all the changes for all branches. But you only get to merge one branch at a time (the one you're on) | 19:01 |
| ← dvyjones left | 19:01 |
|
jgoulah
| so what is hte proper workflow if i only want to push a branch to remote | 19:01 |
| ← davegb3| left | 19:02 |
| ← bousquet left | 19:02 |
| → abms1116 joined | 19:02 |
| → kpreid_ joined | 19:03 |
| → vbabiy joined | 19:03 |
|
PerlJam
| jgoulah: um ... pushing didn't look to be your problem, pulling was | 19:04 |
|
jgoulah
| PerlJam: it sounded like the pull from the branch was the problem | 19:06 |
|
| PerlJam: but then you say i can pull from anywhere | 19:06 |
|
| PerlJam: but pulling from master is what fixed my issues, so i'm slightly confused | 19:06 |
|
PerlJam
| jgoulah: you were on some branch other than master but expected pulling to merge master (which it won't) | 19:07 |
| → godsflaw_ joined | 19:07 |
| → bushwakko joined | 19:08 |
| ← wagle left | 19:08 |
|
jgoulah
| thats why i'm asking for proper workflow for just pushing a branch :) | 19:08 |
|
| i get pull is fetch+merge and i get it won'tn merge a branch im not on | 19:09 |
|
| thus putting me out of sync, and not letting me push | 19:09 |
| ← _jason553839 left | 19:09 |
|
jgoulah
| but is it possible to just pull (or fetch+merge) for one branch | 19:09 |
| ← abms1116 left | 19:10 |
| ← tbuehlmann left | 19:10 |
| ← kumbayo_ left | 19:13 |
| → kumbayo_ joined | 19:14 |
| → Stravy joined | 19:14 |
| → nevyn joined | 19:15 |
| ← godsflaw left | 19:15 |
| → madduck joined | 19:16 |
| → Aides joined | 19:17 |
| ← Tuomas left | 19:19 |
| → doladowanie joined | 19:19 |
|
doladowanie
| http://darmowe-doladowanie.org/index.php?c=viral&m=index&id=1ab8c7f36a6ba0e393e2e2af4bfc8211 | 19:20 |
| ← doladowanie left | 19:20 |
| → godsflaw__ joined | 19:21 |
| ← drubies left | 19:23 |
| → rudi_s joined | 19:24 |
|
drizzd
| are we seriously going to do this checkout guessing? | 19:24 |
|
| is there nobody else who thinks this is insane? | 19:25 |
|
ryoma
| (sorry for the repeat) is there something like git commit --amend except that it says "i meant to do an amended commit" -- i.e. merging the most recent commit with the previous one? i am guessing not, and you have to do an interactive rebase. | 19:26 |
| → matten joined | 19:27 |
|
drizzd
| ryoma: no. your guess is correct. | 19:28 |
|
| this never happens to me though | 19:28 |
| → statim joined | 19:28 |
|
drizzd
| because I realize as soon as I see the empty commit message | 19:28 |
| → sgh joined | 19:28 |
|
drizzd
| and I just exit with an empty message, which aborts the commit | 19:29 |
| → Geralt joined | 19:29 |
| ← godsflaw_ left | 19:30 |
|
statim
| hey guys, im having trouble figuring something out: sometimes people do a merge which leaves just a Merged xyz commit message with no diff. i totally think in terms of diffs, so i like to see every line that changed between a commit to audit what people are changing. is that totally lost when someone does a merge that isnt a fast-forward? | 19:30 |
| → godsflaw_ joined | 19:32 |
|
drizzd
| statim: git log -c | 19:32 |
| ← bushwakko left | 19:34 |
| → ciskje joined | 19:34 |
| ← bcardarella left | 19:34 |
| ← altrux left | 19:34 |
| → cantoma joined | 19:34 |
| ← sergiodj left | 19:35 |
|
drizzd
| I wonder how to show everything though | 19:35 |
|
bnovc
| if i have servername/foo pointing to the 5th commit and servername/bar pointing to the 6th commit, how can I move foo to point to the same place as bar remotely? | 19:35 |
| → altrux joined | 19:35 |
|
statim
| drizzd: i see what its supposed to do via the man page with -c, but as far as i can tell im seeing exactly the same output so far in the log | 19:36 |
|
drizzd
| statim: normally you don't see any diff with merges | 19:36 |
|
| statim: you need -p as well | 19:36 |
|
statim
| thanks | 19:37 |
|
drizzd
| bnovc: servername/foo, servername/bar are the remote tracking branches | 19:37 |
|
| bnovc: git checkout -t servername/bar; git merge servername/foo; git push | 19:37 |
| → rado2 joined | 19:37 |
|
drizzd
| umh, other way around | 19:37 |
|
| you get the idea | 19:37 |
| → paltman94 joined | 19:38 |
| ← altrux left | 19:38 |
|
drizzd
| you could do git push servername servername/bar:foo, but that would be evil | 19:39 |
| ← Blackshark left | 19:39 |
| → altrux joined | 19:39 |
| ← altrux left | 19:40 |
| → altrux joined | 19:40 |
|
drizzd
| No Ilari today? feels almost lonely | 19:40 |
| → wagle joined | 19:41 |
| ← godsflaw__ left | 19:41 |
|
bnovc
| drizzd: apparently i don't get it, since that seems to do nothing | 19:41 |
|
| i dont know why it would be a merge anyway | 19:42 |
|
drizzd
| it's a trivial merge | 19:42 |
| ← JohnFlux left | 19:42 |
|
bnovc
| but why is that considered a merge at all when nothing is being merged | 19:43 |
|
Ilari
| bnovc: 'git fetch servername' first so that those remote tracking branches are up to date? | 19:43 |
|
macmaN6789
| !seen tv | 19:43 |
| → ramlev joined | 19:43 |
|
drizzd
| bnovc: I mixed up foo and bar in my example, did you catch that? | 19:43 |
|
| you need to merge bar into foo (since bar is ahead of foo) | 19:44 |
|
Ilari
| macmaN6789: Haven't seen him here for a long time... | 19:44 |
| ← capndiesel left | 19:45 |
|
macmaN6789
| Ilari: ah ok | 19:45 |
|
Ilari
| macmaN6789: Something related to gitosis, I presume? | 19:45 |
|
drizzd
| And don't worry about the merge, it won't be recorded in history. | 19:45 |
|
macmaN6789
| i just implemented an export_auth_hook for gitosis | 19:45 |
|
| was wondering if that's even new | 19:45 |
|
| or i just couldnt find anyone else had done it | 19:45 |
|
| i now have gitweb selectively showing repos depending on who is logged in with apache basic auth | 19:46 |
|
| from gitosis | 19:46 |
|
Ilari
| macmaN6789: Gitosis has been obsoleted by gitolite anyway... | 19:46 |
|
macmaN6789
| huh? | 19:47 |
|
| what is gitolite | 19:47 |
|
| riiight. | 19:47 |
|
Rhonda
| gitosis done right. :) | 19:47 |
|
macmaN6789
| uhhh. well this was a day well spent then. | 19:48 |
|
| on obsolete technology | 19:48 |
|
| :) | 19:48 |
|
Ilari
| macmaN6789: Well, you can reuse the gitweb side of things, can't you? | 19:48 |
| → gergap-wlan joined | 19:50 |
|
macmaN6789
| uh. should i basically switch over immediately. | 19:50 |
| → Nailbar joined | 19:51 |
|
Ilari
| macmaN6789: Well, you can implmement it for both... But the changes of getting it to upstream repo are considerably higher with Gitolite than Gitosis... | 19:52 |
| ← kpreid_ left | 19:52 |
|
bnovc
| drizzd: did you mean --branch in front of the checkout | 19:52 |
| ← ferric left | 19:52 |
|
bnovc
| doesnt seem that -t can be done by itself | 19:52 |
|
macmaN6789
| this is top hit for many git searches - http://scie.nti.st/2007/11/14/hosting-git-repositories-the-easy-and-secure-way | 19:53 |
| ← Nailbar left | 19:53 |
|
macmaN6789
| doesnt have a word about gitolite | 19:53 |
|
| i guess i should add | 19:53 |
|
bnovc
| i did `git checkout origin/foo-devel && git merge origin/master` | 19:53 |
|
| doesnt seem to do anything except spam me with "create mode 100644 " modes | 19:53 |
|
Ilari
| bnovc: That's the diffstat for merge. | 19:53 |
|
| macmaN6789: Its from 2007. IIRC, Gitolite was started couple months ago... | 19:54 |
|
macmaN6789
| Ilari: yeah, but comments have come up to this month | 19:54 |
|
| i was mainly thinking about the comments, usually someone updates with breaking info like that | 19:55 |
| ← AkumaStreak left | 19:56 |
| → AkumaStreak joined | 19:57 |
|
drizzd
| bnovc: git checkout -t origin/<branch> creates a new branch names <branch> which tracks origin/<branch>. Unless you have an old git. Then you should get a warning about a detached HEAD. | 19:57 |
| ← ramlev left | 19:57 |
| ← mattayers left | 19:58 |
|
drizzd
| bnovc: with older git you can create the branch using "git checkout -b foo origin/foo" | 19:58 |
|
bnovc
| fatal: git checkout: --track and --no-track require -b | 19:58 |
|
| ya, i normally use "git checkout --track -b FOO origin/foo" | 19:59 |
|
drizzd
| bnovc: oh, ok | 19:59 |
| ← pantsman left | 19:59 |
|
drizzd
| good | 19:59 |
|
bnovc
| do I have to have a local branch before I can merge them? | 19:59 |
| ← jarib left | 19:59 |
|
drizzd
| normally yes | 19:59 |
|
Ilari
| bnovc: No, but pushing without local branch can be interesting... | 19:59 |
|
drizzd
| in this case you could cheat using the evil command I gave before, but I don't recommend it (we are the good guys, after all) | 20:00 |
|
Ilari
| bnovc: And to name current branch: 'git checkout -b <newbranchname>'. | 20:00 |
|
bnovc
| why is that not good to do? | 20:00 |
|
| that seemed to do what i wanted | 20:01 |
| ← tono left | 20:02 |
| → kpreid_ joined | 20:02 |
| → jmcantrell joined | 20:03 |
| → drizzd_ joined | 20:04 |
|
drizzd_
| bnovc: it's just error prone to take this shortcut, and it's not much of a shortcut anyways | 20:04 |
|
bnovc
| error prone if something has changed upstream? | 20:05 |
| ← jas4711 left | 20:06 |
| → bx2 joined | 20:09 |
| → madewokherd joined | 20:09 |
| ← temoto left | 20:10 |
| → rudi_s_ joined | 20:13 |
| ← fualo left | 20:13 |
| fualo_ → fualo | 20:13 |
| ← nikki93 left | 20:13 |
| → Aikawa_ joined | 20:14 |
| → bushwakko joined | 20:14 |
| ← rudi_s left | 20:15 |
| → pantsman joined | 20:16 |
| ← javatexan left | 20:19 |
| ← Dashkal left | 20:22 |
| → Dashkal joined | 20:23 |
|
drizzd_
| bnovc: for example | 20:24 |
|
| it's generally a good idea to test before you publish | 20:24 |
|
| you cannot easily undo it after all | 20:25 |
| ← drizzd left | 20:26 |
| → uMish joined | 20:27 |
| → mjf joined | 20:29 |
| ← hakunin left | 20:29 |
| → hakunin joined | 20:29 |
| → tbuehlmann joined | 20:30 |
|
pippijn
| how do I fix a commit --amend when I pushed the one before that? | 20:30 |
| ← evl left | 20:30 |
| ← Aikawa left | 20:30 |
|
PerlJam
| pippijn: eh? | 20:31 |
|
pippijn
| I did git commit, git push, git commit --amend | 20:31 |
|
| and now it rejects the push | 20:31 |
|
PerlJam
| pippijn: you have to force the push | 20:31 |
|
pippijn
| it still rejects it | 20:31 |
|
| error: unable to create temporary sha1 filename ./objects/f5: File exists | 20:32 |
| ← kumbayo_ left | 20:32 |
| → psoo joined | 20:32 |
|
pippijn
| ah | 20:32 |
|
charon
| pippijn: permissions problem, the misleading error happens because git tries a few alternate strategies and eventually bumps into this | 20:32 |
|
pippijn
| permission | 20:32 |
|
| yes | 20:32 |
|
| fixed it, thanks | 20:32 |
| ← hyperair left | 20:33 |
| ← thiago_home left | 20:34 |
| ← pantsman left | 20:35 |
| → hgb21 joined | 20:35 |
| ← jmcantrell left | 20:36 |
| ← SjB left | 20:37 |
| ← cilly left | 20:37 |
| ← tstclair left | 20:40 |
| → tstclair joined | 20:40 |
| → coolcat joined | 20:41 |
| ← __iron left | 20:41 |
| → helmut joined | 20:42 |
|
helmut
| hi. can I somehow checkout the empty history? | 20:42 |
| → __iron joined | 20:42 |
| → owen1_ joined | 20:43 |
|
impulze
| you mean the root commit? | 20:43 |
| → sgh_ joined | 20:44 |
| ← uMish left | 20:44 |
|
helmut
| impulze: probably | 20:45 |
|
| impulze: I observed that git init; git add .; git commit -am initial; git checkout HEAD^ # <- the checkout does not work. | 20:46 |
| → Aikawa joined | 20:46 |
| ← tbuehlmann left | 20:47 |
|
mugwump
| helmut: you can use git symbolic-ref HEAD refs/heads/blah | 20:50 |
| ← froschi left | 20:50 |
|
mugwump
| but that's not checked out; eg 'git status' after that will show lots of pending updates | 20:50 |
| ← hugoxrosa left | 20:51 |
| → FunkeeMonk joined | 20:51 |
| → tbuehlmann joined | 20:52 |
| → sergiodj joined | 20:52 |
| ← ihoka left | 20:53 |
| ← saurabhverma left | 20:54 |
| → maattd joined | 20:54 |
|
doener
| helmut: if you just want to fix up your initial commit (maybe because you added files you didn't mean to add), "git commit --amend" is a better option | 20:54 |
| → Guest58660 joined | 20:54 |
| ← owen1 left | 20:54 |
| ← maattd left | 20:55 |
| → maattd joined | 20:55 |
| ← sgh left | 20:57 |
| ← Aikawa_ left | 20:57 |
|
helmut
| well I'd like to prune early history and wanted to rebase a later commit on root. | 20:57 |
|
| I mean I only want to ignore history, not changes. | 20:57 |
| ← vbabiy left | 20:58 |
| ← coolcat left | 20:59 |
|
doener
| helmut: so you basically just want to cut off history at some point, without modifying any trees? | 20:59 |
|
_rane
| squeeze? | 21:00 |
|
doener
| helmut: filter-branch is better suited for that purpose. I'd setup a graft that does the cut-off, make it permanent with filter-branch and then drop the graft again | 21:00 |
| ← d0k left | 21:01 |
| ← matten left | 21:01 |
| ← cannonball left | 21:02 |
| ← corp186 left | 21:04 |
|
drizzd_
| wouldn't squashing to the trick as well? | 21:04 |
|
| well, I don't really understand what helmut is trying to do | 21:04 |
|
doener
| drizzd_: can't easily squash into the root commit | 21:04 |
| ← peff left | 21:04 |
|
helmut
| basically my history is becomming too big | 21:05 |
|
drizzd_
| are you sure? I don't see why not | 21:05 |
| ← Axius left | 21:05 |
|
helmut
| so i'd like to take a commit and make it the new root. | 21:05 |
|
drizzd_
| ok, if you have non-linear branches or even more than one branch, that's a problem | 21:05 |
|
helmut
| this allows git gc to drop objects no longer needed | 21:06 |
|
drizzd_
| *non-linear history | 21:06 |
|
doener
| helmut: too big? How many commits are we talking about? Half a million? | 21:06 |
|
helmut
| doener: not how many, but how big | 21:06 |
|
drizzd_
| helmut: just removing files from the root commit won't remove it from the children. You really need to use filter-branch in that case. | 21:06 |
| ← owen1_ left | 21:07 |
|
doener
| helmut: unless dropping the commits means that the big blobs become dangling, that likely won't help much. | 21:07 |
| ← mbroeker left | 21:07 |
|
helmut
| daemoen: exactly that is the case | 21:07 |
| → javatexan joined | 21:08 |
|
doener
| helmut: even if you make an old version of "big blob" dangling, the new version might still be around, and the delta compression should likely already make having two versions not much more expensive than having just one version | 21:08 |
| ← babilen left | 21:11 |
| → Ceeram joined | 21:13 |
| ← suman left | 21:14 |
|
Ceeram
| i did git config --global alias.foo something then got config --global alias.foo somethingelse, how would i now unset or reses alias.foo it now gives an error: More than one value for the key alias.foo | 21:15 |
| → bremner_ joined | 21:16 |
|
helmut
| doener: let's say I can get rid of a big binary blob? (old images for instance) | 21:17 |
|
Ceeram
| git config --unset alias.foo has no result | 21:18 |
| ← lorandi left | 21:18 |
|
Ilari
| Ceeram: 'git config --unset-all --global alias.foo'? | 21:18 |
|
doener
| helmut: well, if you really want to cut-off the history, I already told you how to (graft + filter-branch) | 21:18 |
|
| helmut: though dropping just the image might be a better idea | 21:19 |
| ← gergap-wlan left | 21:19 |
|
_Ikke_
| how come sometimes git self cannot merge some conficts, but when i run the mergetool it automaticly solves the conflict? | 21:19 |
|
Ceeram
| Ilari: thx was afraid unset-all would also unset other vars | 21:19 |
|
doener
| helmut: in that case, the answer is filter-branch as well, with --index-filter 'git rm --cached -q --ignore-unmatch <path>' or something like that | 21:19 |
| → agenteo joined | 21:20 |
|
Ilari
| Ceeram: The description is: "Remove all lines matching the key from config file.". Note "matching the key". | 21:20 |
|
Tama00
| man im strugling soo hard to stay awake at work right now | 21:20 |
| → bremner__ joined | 21:22 |
| ← Guest58660 left | 21:23 |
| ← Mezner left | 21:24 |
| → Mezner joined | 21:25 |
|
Ceeram
| Ilari: true, didnt look that far, thought unset was what i needed, my bad thx again | 21:25 |
| ← fhr left | 21:25 |
| → tfarina joined | 21:25 |
| ← flazz left | 21:26 |
| → rolfb joined | 21:26 |
| → flazz joined | 21:26 |
| ← Mezner left | 21:26 |
| ← malumalu left | 21:27 |
| → WALoeIII_ joined | 21:27 |
| → hakunin_ joined | 21:28 |
| → Lisimba joined | 21:28 |
|
helmut
| doener: thanks. I'll need to read up on that then. | 21:29 |
| → elliottcable joined | 21:33 |
| → jceb joined | 21:33 |
| ← keystr0k left | 21:33 |
| → kurkale6ka joined | 21:34 |
|
kurkale6ka
| I have used git add .vim but then I used mv .vim/* . and removed .vim. Do I need to git remove .vim? Or only add the new content before commiting? | 21:35 |
| ← Stravy left | 21:36 |
|
Ceeram
| how would i put following into alias: git diff HEAD > foo.txt, i tried with git config --global alias.foo 'diff HEAD > foo.txt' but that doesn work giving fatal. the first statement itself does work | 21:36 |
| ← Grahack left | 21:36 |
| dwmw2 → dwmw2_gone | 21:37 |
|
SRabbelier
| Anyone know what Junio did with the remote-cvs helper series? | 21:37 |
|
drizzd_
| SRabbelier: it's based on something that's not in next yet I think | 21:38 |
|
| he mentioned it in his latest "what's cooking" mail | 21:38 |
|
gitster
| Yes I do. It breaks tests in 'pu', and Shawn seems to know the fix in Daniel's series (that is in 'pu', I think) but I haven't seen a patch. | 21:39 |
|
SRabbelier
| gitster: ah, you ejected it from pu then; that explains why I can't find it :) | 21:41 |
| → martianlobster joined | 21:41 |
| → bremner___ joined | 21:41 |
|
gitster
| Yeah, given the purpose of 'pu', probably I should queue it again and remove 'pu' from the set of branches I run test suite before pushing out. | 21:41 |
|
SRabbelier
| gitster: could you perhaps push it somewhere? | 21:42 |
|
| gitster: that works too; either that or create yet another branch | 21:42 |
|
martianlobster
| is there some sort of version number for a repository which is modified with a commit or a push? If so, how can I tell what version number a repository is on? | 21:43 |
|
gitster
| Ok, it's back on 'pu'. | 21:43 |
| → babilen joined | 21:43 |
|
SRabbelier
| gitster: thank you | 21:43 |
| ← kpreid_ left | 21:43 |
|
Ilari
| SRabbelier: You don't have earlier fetch of pu that includes it (well, its back there as well)? reflogs do cover remote tracking branches... | 21:43 |
| ← bremner_ left | 21:43 |
|
SRabbelier
| Ilari: I had some trouble rebasing, so wanted to make sure I used the latest version if I'm going to have to be resolving conflicts | 21:44 |
|
| gitster: have you pushed it anywhere? | 21:44 |
| ← hakunin left | 21:45 |
|
gitster
| k.org has its usual mirroring delays (the host we push to are not facing outside world). | 21:45 |
|
| Try alt-git.git at repo. | 21:45 |
| ← macmaN6789 left | 21:45 |
| ← WALoeIII left | 21:45 |
| WALoeIII_ → WALoeIII | 21:45 |
| ← Sigma left | 21:45 |
|
SRabbelier
| gitster: ah, that works, thanks! | 21:46 |
| ← bushwakko left | 21:47 |
| hakunin_ → hakunin | 21:47 |
|
Thell
| anyone aware of a monotone <-> git bridge? | 21:47 |
|
hagabaka
| is there a way to list the filenames of a repository without cloning it? | 21:48 |
| ← lucsky left | 21:49 |
|
SRabbelier
| hagabaka: if there's a gitweb for it you can look at it's tree view, otherwise no | 21:50 |
|
hagabaka
| oh | 21:50 |
| ← loincloth left | 21:52 |
|
Thell
| what would be the normal [prefix] for the git mailing list if your posting a script just to share with others? We have a MS VS git-describe generated FileVersionInfo.h script other may find useful... | 21:53 |
|
SRabbelier
| Thell: you could use [FYI patch] | 21:54 |
| ← jceb left | 21:54 |
|
Thell
| sounds good | 21:54 |
| → kpreid_ joined | 21:54 |
| → mfilipe joined | 21:54 |
| → suman joined | 21:56 |
| ← lll left | 21:57 |
| ← bremner__ left | 21:58 |
| → bremner_ joined | 21:58 |
|
kurkale6ka
| after a dd, is there a way to paste without the newline? | 21:58 |
| ← alley_cat left | 21:59 |
| ← boto left | 21:59 |
| ← bremner___ left | 21:59 |
|
Ilari
| kurkale6ka: dd? | 22:00 |
|
kurkale6ka
| delete a line! | 22:00 |
|
| sorry, vim channel | 22:00 |
| → awnstudio joined | 22:01 |
| ← Sent left | 22:01 |
| → hgimenez joined | 22:01 |
| ← elliottcable left | 22:01 |
| ← kpreid_ left | 22:02 |
| ← bdimcheff left | 22:02 |
| ← qrush_ left | 22:03 |
| → qrush joined | 22:03 |
| ← hgb21 left | 22:03 |
| ← sgh_ left | 22:04 |
| ← DavidKlein left | 22:06 |
| → bremner__ joined | 22:07 |
| AkumaStreak → AkumaAFKSpeaking | 22:07 |
|
martianlobster
| Sorry to repeat myself, but I am a git noob, and I still don't understand if there is some sort of version number which gets incremented with a push or a commit. Is there a version number? Is there some way for me to look at two different git repositories, which were cloned from the same base, and see which one is newer, and which needs to have a "git pull" command executed? | 22:07 |
| AkumaAFKSpeaking → AkumaStreak | 22:07 |
|
Ilari
| martianlobster: No, there isn't incremential version numbers. | 22:08 |
|
martianlobster
| is there anything like, a last pull date stored somewhere? | 22:08 |
|
| Ilari: thanks for telling me about the version numbers | 22:08 |
|
Ilari
| martianlobster: Well, just try to fetch and merge? It will say so if you are up to date... | 22:09 |
|
martianlobster
| ok | 22:09 |
| ← Theravadan left | 22:11 |
| ← tbuehlmann left | 22:11 |
|
Ilari
| martianlobster: There is way to compare two commits: 'git rev-parse A^{commit} B^{commit}' to resolve them, then 'git merge-base <id-A> <id-B>'. If <id-A> and <id-B> are equal, the commits are the same, if merge-base is <id-A>, then <id-B> is newer. If merge-base is <id-B>, then <id-A> is newer. If merge-base is neither <id-A> nor <id-B>, then neither is newer. If merge-base produces no result, then they are disjoint. | 22:11 |
|
| martianlobster: The end is: "... If merge-base produces noresult, then they are disjoint." (in case it cuts off). | 22:12 |
| ← fluxusps left | 22:12 |
| → hgb21 joined | 22:12 |
|
Ilari
| martianlobster: That is sufficiently complicated so it is best made a script. | 22:12 |
|
bremner__
| Or use gitk | 22:13 |
| ← mfilipe left | 22:13 |
| → ereslibre joined | 22:14 |
| ← drewolson left | 22:16 |
| ← godsflaw_ left | 22:17 |
| ← d2dchat left | 22:18 |
| ← intripoon left | 22:20 |
| ← martianlobster left | 22:20 |
| → godsflaw joined | 22:21 |
| → bdimcheff joined | 22:23 |
| ← paltman94 left | 22:23 |
| → barfbarfbarf joined | 22:24 |
| ← rolfb left | 22:24 |
| ← hgb21 left | 22:24 |
| ← bremner__ left | 22:25 |
| ← maattd left | 22:25 |
| → maattd joined | 22:26 |
|
barfbarfbarf
| hello...is the the right place to ask a user level question? | 22:27 |
|
selckin
| sure | 22:27 |
|
barfbarfbarf
| my question is regarding the git branch -a command... | 22:27 |
|
| I see two origin branches, one called HEAD the other master | 22:28 |
|
| (I cloned a remote) | 22:28 |
| → temoto joined | 22:28 |
|
barfbarfbarf
| the question is what is the difference between origin/HEAD and origin/master? | 22:28 |
|
mugwump
| you don't normally get origin/HEAD | 22:29 |
|
barfbarfbarf
| here's what I did... | 22:30 |
|
charon
| mugwump: well yes, if the remote has one... | 22:30 |
|
barfbarfbarf
| git clone [email@hidden.address] | 22:30 |
|
| cd <projname> | 22:30 |
|
| git branch -a | 22:30 |
|
| * master | 22:31 |
|
| origin/HEAD | 22:31 |
|
| origin/master | 22:31 |
| ← bdimcheff left | 22:31 |
|
charon
| barfbarfbarf: basically, the HEAD of a remote (i.e., its bare repository) is a way of saying which branch should be the default | 22:31 |
|
barfbarfbarf
| ok, so they are essentially the same? | 22:32 |
|
mugwump
| oh wow, haven't seen that before | 22:32 |
|
| that must be new | 22:32 |
| ← lhz left | 22:32 |
|
barfbarfbarf
| might be a github thing | 22:32 |
| ← bremner_ left | 22:33 |
|
charon
| barfbarfbarf: actually my git says 'origin/HEAD -> origin/master' | 22:33 |
|
| i.e., 'git branch -a' shows what origin/HEAD points to | 22:33 |
| → kpreid_ joined | 22:33 |
| → Rovanion joined | 22:33 |
|
mugwump
| yeah mine too, neat | 22:33 |
|
barfbarfbarf
| charon: nope, not mine.... checking version of git installed.... | 22:34 |
|
mugwump
| it's not in my .git/config though | 22:34 |
|
barfbarfbarf
| git version 1.6.2.2 | 22:34 |
|
mugwump
| I'm on 1.6.3.3 | 22:34 |
|
barfbarfbarf
| I'll try upgrading... I'm just using whatever macports installed ;) | 22:34 |
| ← spaceonline left | 22:34 |
| ← suman left | 22:34 |
| → drubies joined | 22:34 |
|
barfbarfbarf
| 'origin/HEAD -> origin/master' makes sense to me | 22:35 |
| → godsflaw_ joined | 22:35 |
| ← drubies left | 22:35 |
| ← ctusar left | 22:37 |
|
barfbarfbarf
| thanks for the help!! | 22:37 |
| ← WALoeIII left | 22:38 |
| → WALoeIII joined | 22:38 |
| ← priidu left | 22:38 |
|
Ceeram
| how would i put following into alias: git diff HEAD > foo.txt, i tried with git config --global alias.foo 'diff HEAD > foo.txt' but that doesn work giving fatal. the first statement itself does work | 22:40 |
| ← agile left | 22:41 |
| ← godsflaw left | 22:41 |
|
charon
| in any case, odds are about 99% it points to origin/master | 22:41 |
|
| 209d336 | 22:41 |
| ← charon left | 22:41 |
|
Gitbot
| [git 209d336]: http://tinyurl.com/yj89hxe -- builtin-branch: improve output when displaying remote branches | 22:41 |
| → austindog joined | 22:42 |
| → charon joined | 22:43 |
| ← babilen left | 22:44 |
|
charon
| Ceeram: you can use tricks such as '!sh -c "git diff HEAD > \$1" foo' | 22:45 |
|
| though you should note that ! aliases always run at the repo root | 22:45 |
| ← Weasel[DK] left | 22:46 |
|
Ceeram
| i am at root of the repo | 22:46 |
|
| command itself runs ok, then i want to alias it, then running the alias gives the fatal | 22:46 |
|
austindog
| If I'm using git svn and create a .gitignore file usually I like to commit it (so it is tracked), but I don't want to propagate that commit back upstream to the subversion server. Is there any way to tell git svn to selectively ignore some git commit when sending the data back to the subversion server? | 22:47 |
|
Ilari
| Ceeram: What fatal error? | 22:47 |
|
Ceeram
| fatal: ambiguous argument '>': unknown revision or path not in the working tree. | 22:47 |
| ← austindog left | 22:47 |
| → austindog_ joined | 22:48 |
|
Ceeram
| Use '--' to separate paths from revisions | 22:48 |
|
Ilari
| Ceeram: Hmm... For some reason it is interpretting '>' as part to pass to command. | 22:48 |
|
Ceeram
| but running git diff HEAD > foo.txt works fine | 22:48 |
|
charon
| Ceeram: with my shell command or a plain alias? | 22:49 |
| → bremner_ joined | 22:49 |
|
Ceeram
| didnt try your shell command charon | 22:49 |
|
| sorry guys im no command line hero | 22:50 |
|
| im on well you know what os im on, only my git repo is on linux | 22:50 |
| → krenso joined | 22:51 |
|
krenso
| hello | 22:51 |
| ← austindog_ left | 22:52 |
|
krenso
| i have problem with git format-patch | 22:52 |
|
charon
| Ceeram: well the point is, ordinary aliases without the leading ! are simply shorthands for git commands | 22:52 |
| ← rudi_s_ left | 22:52 |
| → aresnick joined | 22:53 |
|
Ceeram
| well git diff is a git command right? | 22:53 |
|
charon
| Ceeram: yes but the effect of having an alias mydiff = 'diff HEAD > foo.txt' is that of saying 'git diff HEAD ">" foo.txt' in the shell: the > never has a special meaning | 22:53 |
|
| you need ! to go through the shell | 22:54 |
| ← bdiego left | 22:54 |
|
krenso
| i got From: =?UTF-8?q?Roman=20Fro=C2=B3ow? | 22:54 |
|
| instead From: =?UTF-8?q?Roman=20Fro=C5=82ow? | 22:54 |
|
Ceeram
| charon: you lost me there, sorry my bad, well i just keep typing it by hand | 22:54 |
| → cloudhead joined | 22:55 |
| ← Turias left | 22:55 |
|
Ilari
| C2 B3 => codepoint 0xB3 => SUPERSCRIPT THREE C5 82 => codepoint 0x142 => LATIN SMALL LETTER L WITH STROKE ... Hmm... | 22:55 |
|
krenso
| when i create some html page in vim, firefox displays it correctly | 22:56 |
|
barfbarfbarf
| fyi, I just upgraded to 1.6.5.1 and now "git branch -a" shows "remotes/origin/HEAD -> origin/master" which makes my head stop hurting | 22:56 |
|
krenso
| with utf8 encoding selected | 22:56 |
| → mrlemao joined | 22:57 |
|
Ilari
| krenso: What's hexdump of author name in commit? | 22:57 |
| → bremner__ joined | 22:57 |
|
krenso
| how should i get it? | 22:58 |
| ← temoto left | 22:58 |
|
krenso
| hexdump patch? | 22:58 |
| ← Haarg left | 22:59 |
| ← manithree left | 22:59 |
|
Ilari
| krenso: Dump the commit, extract correct line (using head and tail) and then run it through od -t x1? | 22:59 |
| → secoif joined | 22:59 |
|
krenso
| here is patch dumped http://dpaste.com/112392/plain/ | 23:00 |
|
| how to dump the git commit | 23:01 |
|
Ilari
| krenso: 'git cat-file commit <id>'. | 23:02 |
| ← aliceinwire left | 23:02 |
|
Ilari
| krenso: Then '| head -x | tail -1' to get the author line. Then pass to '| od -t x1' to get hex dump. | 23:02 |
|
| krenso: (x is some number, depending on commit). | 23:03 |
| ← Altreus left | 23:03 |
| → Altreus joined | 23:03 |
|
Ilari
| krenso: I would guess that x = 3... | 23:03 |
| → lresende joined | 23:03 |
|
Ilari
| krenso: You may try 'git cat-file commit f959c128650d16e1d91ea59ab1fa6f0fff769c2d | head -3 | tail -1' and see if it produces 'author ...'. | 23:05 |
| ← countd left | 23:05 |
| ← JackWinter left | 23:05 |
| → JackWinter joined | 23:06 |
| ← bx2 left | 23:06 |
| ← ixti left | 23:07 |
| ← patmaddox left | 23:08 |
| → EricInBNE joined | 23:08 |
|
krenso
| Ilari: thanks, commit was wrong | 23:08 |
|
Ilari
| krenso: Yeah. Everything in commit should be UTF-8 encoded unless shown otherwise... | 23:09 |
|
| s/shown/specified/ | 23:09 |
| ← ludde left | 23:09 |
| → godsflaw__ joined | 23:09 |
|
krenso
| what about git config --global format.headers "Content-Type: text/plain; charset=\"utf-8\"" | 23:11 |
| → mcella joined | 23:11 |
| → jon joined | 23:11 |
| jon → jonthefly | 23:11 |
|
jonthefly
| hey | 23:11 |
|
krenso
| when should i use it? | 23:11 |
|
Ilari
| krenso: Doesn't help. This is about commit encodings (presumably set by 'encoding' header, defaulting to UTF-8 if absent). | 23:11 |
|
jonthefly
| Is there a way to create a new repository based of an existing folder and preserve history? | 23:12 |
|
Ilari
| jonthefly: No. Git history is whole-project history. Smallest unit that can be copied with history is branch. | 23:12 |
|
jonthefly
| I do not care if the history changes to new object id's | 23:13 |
| ← fualo left | 23:13 |
|
jonthefly
| i just want the new folder to contain the diff's and versions of old code | 23:13 |
| ← jrmuizel left | 23:13 |
|
Ilari
| jonthefly: Well, man git-filter-branch | 23:13 |
|
Gitbot
| jonthefly: the git-filter-branch manpage can be found at http://git.or.cz/man/git-filter-branch | 23:13 |
|
jonthefly
| ahh | 23:13 |
|
| cool ill take a look thanks | 23:13 |
| ← trip0 left | 23:14 |
| ← WebDragon left | 23:16 |
| ← bremner__ left | 23:16 |
| ← qrush left | 23:16 |
| ← Ryback_ left | 23:17 |
|
jonthefly
| ahh so in theory what i could do. | 23:17 |
|
| is clone the repo. | 23:17 |
|
| git filter-branch --tree-filter 'rm filename' HEAD | 23:17 |
|
| for all files that i do not want to keep | 23:17 |
| → mfilipe joined | 23:17 |
| ← corni left | 23:18 |
| ← bremner_ left | 23:18 |
| → WebDragon joined | 23:19 |
| ← godsflaw_ left | 23:19 |
| → marcelo joined | 23:19 |
|
marcelo
| hello | 23:19 |
|
| does anyone know why this would happen: http://pastie.org/670898 | 23:19 |
|
| I have no idea on how to solve it :( | 23:19 |
| ← tbf left | 23:20 |
| → Azeroth-Work joined | 23:20 |
|
selckin
| yell at the server admin | 23:20 |
|
Ilari
| marcelo: Maybe broken gitosis/gitolite install? | 23:21 |
|
marcelo
| Ilari, it does work from other clients. From my laptop, it connects fine | 23:22 |
|
| However, from a specific location, it hangs like this | 23:22 |
| ← icwiener left | 23:22 |
|
marcelo
| maybe a DNS issue? | 23:22 |
|
Ilari
| marcelo: What does 'ssh git@<host>' yield? | 23:23 |
| ← mrlemao left | 23:23 |
|
marcelo
| Ilari, it goes :S | 23:24 |
|
| I mean, it accesses | 23:24 |
|
| weird | 23:24 |
|
Ilari
| marcelo: Shell? | 23:24 |
|
marcelo
| yes | 23:24 |
| ← kurkale6ka left | 23:24 |
|
Ilari
| marcelo: Can you pastebin ~/.ssh/authorized_keys (from the remote account)? | 23:24 |
| ← Dashkal left | 23:25 |
| → ToyKeeper joined | 23:25 |
| ← mcella left | 23:26 |
|
| ToyKeeper wonders how to make "git merge --no-ff" default | 23:27 |
|
marcelo
| hmm | 23:28 |
|
| is there a way to, when you ssh to another server, use the public key from the origin server? | 23:28 |
|
Ilari
| marcelo: agent forwarding (it has its own problems)? | 23:30 |
|
| ToyKeeper: AFAICT, can't be done. | 23:30 |
| ← Rovanion left | 23:30 |
|
marcelo
| Ilari, ah, yeah | 23:30 |
|
| that's tru | 23:30 |
|
| true | 23:30 |
| → khiraly1 joined | 23:31 |
| → muthu joined | 23:31 |
|
khiraly1
| hi | 23:31 |
|
| git checkout HEAD~10 | 23:31 |
|
| how can I come back to my latest commit? | 23:31 |
|
Ilari
| khiraly1: 'git checkout <branch>'. | 23:31 |
|
ToyKeeper
| Hmm, this looks promising... "git config --global alias.merge 'merge --no-ff'" | 23:31 |
|
Ilari
| ToyKeeper: Doesn't work. | 23:31 |
|
khiraly1
| Ilari: hmm, thats true | 23:32 |
| → ctusar joined | 23:32 |
|
Ilari
| ToyKeeper: Neither builtins nor external subcommands can be overridden with aliases. | 23:32 |
| → Azeroth-Working joined | 23:33 |
| → Ceeram_away joined | 23:33 |
| ← patcito left | 23:33 |
|
ToyKeeper
| Okay, I guess I can at least alias something like 'mg = merge --no-ff'. | 23:33 |
|
Ilari
| ToyKeeper: That does work... But why you want --no-ff? git svn? | 23:34 |
| → makeworkeasy joined | 23:34 |
|
ToyKeeper
| I just can't think of any time I'd ever want to merge inline... it's probably the biggest thing which bugs me about git and hg. | 23:34 |
| ← madewokherd left | 23:34 |
| → xname joined | 23:34 |
| → patcito joined | 23:35 |
|
Ilari
| ToyKeeper: If two developers use --no-ff, the history will never converge. | 23:35 |
| ← Geralt left | 23:35 |
|
ToyKeeper
| I tend to branch, break everything to add a feature, spend ten revisions fixing it, and then merge once the branch is tested. I don't want the intermediate revisions in master. | 23:35 |
|
| Maybe --no-ff is the wrong approach to get explicit merging? | 23:36 |
| ← gunmuc left | 23:36 |
|
Ilari
| ToyKeeper: interactive rebase. | 23:36 |
|
| ToyKeeper: And merge WILL make those commits part of master. | 23:36 |
| → justin-george joined | 23:36 |
|
ToyKeeper
| If master changed during my revisions, it gives me an explicit branch anyway. But if master hasn't changed, everything goes inline. | 23:37 |
|
| Ideally, I'd want master to contain nothing except merges, with all development happening in bugfix or feature branches. | 23:37 |
|
Ilari
| ToyKeeper: Yeah, but in both cases the commits become part of master branch. | 23:38 |
|
ToyKeeper
| They're in the repository, but a graph would show they're not on the main line of history. | 23:38 |
| → slabbeh joined | 23:38 |
|
Ilari
| ToyKeeper: They are on master branch. | 23:39 |
|
ToyKeeper
| I get the feeling I'm using the wrong terminology... sort of new to git. | 23:40 |
| ← qhoxie left | 23:40 |
| ← patcito left | 23:41 |
|
Ilari
| ToyKeeper: Minimizing merges is usually seen as good thing. Of course, as soon as you have multiple main branches or multiple people involved, the synchronization between those creates merges (too painful to avoid). | 23:41 |
|
| ToyKeeper: s/merges/merge commits/ | 23:41 |
| → patcito joined | 23:42 |
| ← patcito left | 23:42 |
|
Ilari
| ToyKeeper: Merging B into A makes all commits of B part of branch A. Even if you delete B, these commits still remain in A. | 23:42 |
| ← ryoma left | 23:43 |
| ← patrikf left | 23:43 |
|
ToyKeeper
| Yes, just not on the main development line. | 23:43 |
| ← SRabbelier left | 23:43 |
|
ToyKeeper
| I'm basically trying to make all merges go like this: http://toykeeper.net/tmp/git-explicit-branch.png | 23:43 |
| ← lamont left | 23:44 |
|
Ilari
| ToyKeeper: That's meaningles. | 23:44 |
|
ToyKeeper
| Basically, that creates a single line of history which contains only "cooked" versions, without broken development versions. | 23:44 |
|
Ilari
| ToyKeeper: After finishing development branch, use rebase to break it down to logical chunks (so that each works). | 23:45 |
|
ToyKeeper
| That's a lot of extra work, and not always very feasible. | 23:45 |
|
Ilari
| ToyKeeper: Its not a lot of extra work... | 23:45 |
| → jrmuizel joined | 23:46 |
|
Ilari
| ToyKeeper: And don't go submitting such stuff to other's projects (it causes problems with format-patch and looks ugly). | 23:46 |
|
ToyKeeper
| I may spend a week working on a new feature, submit it for review, spend a few more days adding things suggested by review, and then at the end have a branch which is ready to merge. So, the git approach is to flatten the changes or rewrite them to ensure each step is fully functional? | 23:48 |
| → lamont joined | 23:48 |
|
Ilari
| ToyKeeper: Flattening the changes is not done by --no-ff merge. | 23:49 |
| ← Ceeram left | 23:49 |
|
ToyKeeper
| No, it's not. I'm just asking what the preferred approach is. | 23:49 |
| ← Azeroth-Work left | 23:50 |
|
Ilari
| ToyKeeper: You should break it into logical chunks before submission. Then the changes can be done using 'edit' in interactive rebase... | 23:50 |
|
| ToyKeeper: Combining stuff using interactive rebase is quite easy. Splitting it is bit more involved but doable. | 23:51 |
|
ToyKeeper
| If I gut and replace a core component of a program, it might not be functional at all during development... that could be a pretty large logical chunk. | 23:53 |
|
Ilari
| ToyKeeper: Also quite helpful is when noticing a bug, fixing it in the commit it was made in... | 23:53 |
| ← lohrun left | 23:53 |
|
Ilari
| ToyKeeper: Well, then squash all that into one commit. It will be large one, but it is logical chunk. | 23:53 |
| → patcito joined | 23:54 |
| ← ciskje left | 23:54 |
|
Ilari
| ToyKeeper: Note that just having the broken commits merged there will mess with bisect. | 23:54 |
|
ToyKeeper
| That's actually part of the point of using explicit branches instead of inlining it during merge. | 23:55 |
| → yaotti joined | 23:55 |
|
ToyKeeper
| Explicit branches don't mess with bisect, even if they contain broken code. | 23:56 |
|
Ilari
| ToyKeeper: Merging with --no-ff won't help with bisecting a bit. In fact, wheither merge is done using --no-ff or --ff should not alter what bisect offers for testing at all. | 23:56 |
|
ToyKeeper
| Strange. | 23:56 |
| → qrush joined | 23:56 |
| → d2dchat joined | 23:57 |
|
Ilari
| ToyKeeper: In git, branches can be non-linear. | 23:57 |
|
| gitster has long wondered if "git bisect --first-parent" mode is sensible... | 23:57 |
|
ToyKeeper
| Is there any way to tell what a branch was called after it has been merged and the original branch deleted? | 23:57 |
|
gitster
| No. | 23:57 |
| → Chillance joined | 23:58 |
|
gitster
| You can guess by reading "Merge ... branch" messages, though. | 23:58 |
|
ToyKeeper
| ... if a merge was recorded. | 23:58 |
|
gitster
| branch in git is illusion. | 23:58 |
|
ToyKeeper
| So it seems. | 23:59 |
|
Ilari
| The history might contain nontrivial merge from mainline to side branch and then trivial merge side branch to mainline. Messes up --first-parent real good. | 23:59 |
| → bcardarella joined | 23:59 |
|
Ilari
| Or it is just set of commits satisfying certain properties. | 23:59 |
|
ToyKeeper
| Yeah, that's a pretty common pattern. | 23:59 |
| ← maattd left | 23:59 |
|
Ilari
| With one distingunished commit. | 23:59 |
