IRCloggy #git 2007-04-04

Logs Search ←Prev date Next date→ Channels Documentation

Provider of IRC logs since 2005.
WARNING: As Freenode became unjoinable and lost all warnings in topics, we cannot log channels on Freenode anymore.

2007-04-04

→ DrNick joined00:08
rkaway → rambokid00:14
dwmw2_BOS → dwmw2_gone00:24
gitster /away commute00:51
/away z00:51
→ gitster joined00:52
→ Romster joined01:07
→ spearce joined01:08
mugwump my new merge script is unstoppable!01:22
rambokid → rkaway01:29
clee spearce: check the git list :)01:37
spearce mugwump: what does your merge script merge?01:38
clee: i'm waiting for my fetchmail to catchup...01:38
clee (if you have any suggestions for other commands that might be useful to test against one-pack-versus-partitioned-packs, let me know)01:38
spearce: you should use retchmail instead ;)01:38
mugwump spearce: utsl.gen.nz/scripts/smartmerge01:38
sorry 'bout the server mime type01:38
the comment at the top is a bit wrong01:39
spearce whoops, gotta run.01:39
mugwump it's now good for resolving merges left by git-core too01:39
→ sgrimm joined01:50
→ Avochelm joined01:55
→ kanru joined01:55
dwmw2_gone → dwmw202:00
→ kanru joined02:12
→ sgrimm_ joined02:13
← sgrimm_ left02:13
→ tcoppi joined02:15
→ sgrimm joined03:08
→ orospakr joined03:43
→ Tv joined03:51
→ Tv joined03:54
ribas Does anyone has some Presentation to introduce GIT?03:59
i'm making my own [in portuguese], but if someone already made something may help =)03:59
people already sleeping :)04:11
→ spearce joined04:11
→ rkaway joined04:28
dwmw2 → dwmw2_gone04:28
mugwump ribas: http://members.cox.net/junkio/200607-ols.pdf04:28
spearce "the scm that came before git" ah, fond memories.04:29
ribas mugwump: i'll look at those stuff tomorrow!!! =) [wednesday afternoon]04:37
mugwump: thanks for the presentation =)04:37
spearce wha? its wed. already. ;-)04:37
ribas it is 1:37am!!!! i'll see after 12:00 ;)04:38
spearce heh.04:38
i guess you are about in the same timezone as me then.04:38
ribas cool!04:39
whre do you live?04:40
spearce upstate new york04:40
dwmw2_gone → dwmw204:41
→ beu joined04:43
ribas nice04:43
spearce eh, we get snow sometimes.04:43
ribas hehe.. i never saw snow!!!04:43
spearce it finally melted, but sounds like we might get more before the end of the week.04:43
→ gitster joined04:44
spearce my wife insists we move to a warmer climate. she doesn't like the snow. i don't either, actually.04:44
ribas and i'm having some really hot days.. passing 36º CELCIUS04:44
gitster it's not *that* warm here in SoCal...04:51
spearce we were in the mid-50s F today. not a bad day today.04:51
spearce wonders how he used git before bash completion...05:01
aeruder_ that stuff does make things a lot easier05:02
gitster wonders how he did any development before git.05:02
mugwump spearce: you used git-<tab> instead of git␠<tab> ?05:02
spearce hah!05:02
muguwmp: I just completed an ugly remote name (anthony) as an argument to git fetch. it was nice.05:02
aeruder_ mugwump: yea except imagine tab-completing branch names and such ;)05:02
gitster I encountered a few commands recently that I wanted completion but contrib/completion did not know about. I'll bring them up on the list when I see them next.05:03
mugwump hey that sounds good05:03
spearce please do, i'll hack it up for you.05:03
i haven't found anything missing myself, which is why there hasn't been any work on it recently. ;)05:04
→ spearce_ joined05:17
spearce_ → spearce05:18
spearce crap. gitk doesn't like it when you restart it on your laptop after it was on the external display... darn thing opens on the external display, which doesn't exist...05:21
gitster irritating, isn't it? My note's screen is much smaller and I sometimes have trouble after working on desktop.05:22
rm -f ~/.gitk is your friend.05:22
spearce yea, i actually used to do that on cygwin, to that point that i patched my local git.git clone. ;-)05:23
i think its fixable, i just haven't taken the time to do it.05:23
i think git-gui might also suffer from the same fate, sadly...05:23
gitster Personally I do not use gui tools that much myself, so I do not understand why many applications want to keep the location/geometry from previous session automatically.05:24
mugwump use an Xresource05:24
spearce when i have 8 copies of git-gui running in 8 different repositories, its nice that they come up at predictable positions on screen. each of the 8 is in a different "area"...05:24
of course this is windows, where Xresource don't exist. and command line arguments would serve just as good of a job.05:25
gitster I do not oppose to "Save current layout" action. I can even live with "Do not autosave geometry changes" option that only weirds like me enable.05:25
spearce heh. yes, that option would be better than destroying your .gitk every time, as maybe you want a different font, but otherwise don't want the ui saved.05:26
gitster Or perhaps saved configuration "per display".05:26
(or at least keyed with geometry obtained from xdpyinfo or something like that)05:27
spearce indeed. or just damned bounds checking of the config file geometry against the current geometry.05:27
i love how Tk just applies whatever it gets, as if that part of the desktop is still there...05:27
gitster It's good that I have only 30 or so commits since -rc3 and most of them are minor fixes and docs.05:32
spearce indeed, 1.5.1 went very smoothly i think.05:32
gitster Although there are five fixes to rather grave ones...05:32
spearce but i'm not the maintainer. ;-)05:32
→ gitster joined05:36
ribas going to sleep05:51
g'night05:51
→ sgrimm joined05:53
gitster night.05:55
gitster changed the topic to: 1.5.0.7/1.5.1-rc3 | Everyone asleep or clueless? Try [email@hidden.address] | Channel log http://colabti.de/irclogger/irclogger_log/git05:56
spearce wow. gitster updated the topic! ;-)05:56
gitster so 1.5.0.7 is out, not that it matters.05:59
mugwump rushes to upgrade05:59
gitster nah, tomorrow we will have 1.5.105:59
mugwump I'll merge that in too!!!06:00
spearce indeed. and users are sadly still running 1.4.4.406:00
gitster what's a british slang for naval admiral?06:00
mugwump that's very specific06:01
as opposed to "skipper" ?06:01
spearce big-wigs: http://www.history.navy.mil/faqs/faq47-1.htm ?06:02
gitster ok I can live with that. people may unofficially call the release "skipper".06:02
spearce that may be my only meanginful contribution to 1.5.1. ;-)06:02
use it wisely. ;-)06:03
→ Romster joined06:03
gitster You have 76 commits (many of them git-gui) since 1.5.006:03
spearce heh. i've felt like i haven't paid any attention to git since 1.5.0 shipped. been busy the past two weeks on non-git projects. ;-)06:04
gitster Maybe that was why 1.5.1 went so smoothly ;-)06:05
spearce hah! yea, very possible. ;-)06:05
gitster 1.5.0 really started as Carl Worth release and ended up as Shawn Pearce release.06:05
spearce hey now, a lot of other people put a huge amount of effort into 1.5.0.06:06
i just snowplowed in two huge projects i had been working on for like 6 months. :)06:06
gitster shortlog -n -s v1.4.0..v1.5.0 ^v1.4.4.4 shows 529 to you, nobody comes even close.06:07
spearce i think over 300 are git-gui, i was pretty fine-grained in that. and another 100 or so is git-fast-import, and maybe another 40 for sliding mmap...06:07
dwmw2 → dwmw2_gone06:08
dkagedal I tried git-gui yesterday, and one thing that immediately annoyed me was that the fonts were practically unreadable. What is the cure for that?06:11
I'm running gnome on ubuntu edgy, btw06:12
spearce dkagedal: edit your options, to pick a more suitable font. but i think i missed a few -font options on some widgets. i have a patch here to fix it, i just haven't had time to review it, apply it, and test the rest of that series on my systems.06:12
→ Dodji joined06:13
dkagedal thanks, I wonder why I didn't try to change it the obvious way...06:15
spearce I think someone else reported a problem with git-gui on Linux (horrible, horrible fonts), and since I don't really use a Linux platform (shame on me!) I didn't get to really debug it. The patch I have queued looks like it might help on those systems. I'm hoping i can get it applied and tested tomorrow.06:16
dkagedal now I have readable fonts, but it looks like a blast from the past. This was how things looked back in the nineties. :-)06:18
spearce its a quick hack. ;-)06:18
and i miss the '80s...06:18
dkagedal maybe you can write something using athena widgets instead06:19
mugwump emacs ftw06:19
spearce oh gawd, i remember trying to compile athena widgets. yes, i should really use those in git-gui too!06:20
→ sgrimm joined06:27
spearce so here's something i've noticed about git and new users: they (the new users) always expect git-pull to work, even when they have modified files that aren't committed yet. when questioned about what the expected outcome of such an operation should be, they have no clue... and then they ask "why does foo.cpp need an update? i already changed it!"06:30
clee haha06:32
fuck athena.06:32
er, sorry. frak athena.06:32
spearce hey, no fraking on this channel while bsg isn't showing new episodes!06:35
mugwump well sure wouldn't want to use a bloated toolkit like ... erm ... GNOME ;)06:36
spearce nah, use XUL. its crossplatform!06:37
clee mugwump: c'mon. GNOME's not a toolkit.06:38
(not that you could really call GTK one, either, but hey.)06:38
spearce is that the kde in you talking? ;-)06:39
mugwump ducks off before the blows start swinging &06:39
clee grins06:43
clee spearce: I make no secret of my allegiance. :)06:43
orospakr clee: still a big KDE guy?06:44
clee orospakr: I dunno if I'm a big guy, but I'm definitely a KDE guy06:44
and I will be until something better comes along06:44
orospakr heh :)06:44
orospakr is certainly going to switch to KDE4, at least for a little while.06:45
orospakr it looks pretty sweet.06:45
→ kanru joined07:20
→ Romster joined07:26
→ nud joined07:28
← dduncan left07:51
→ ferdy joined08:00
→ dkagedal joined08:38
tomprince 08:50
gitster changed the topic to: 1.5.0.7/1.5.1 | Everyone asleep or clueless? Try [email@hidden.address] | Channel log http://colabti.de/irclogger/irclogger_log/git08:59
→ sgrimm joined09:02
→ gcorriga joined09:09
gcorriga hi all09:09
I'm looking for an admin or mentor partecipating in Google Summer of Code. Are there any here?09:10
gitster The admin is spearce who is a regular here, but he is in TZ=US/Eastern so it is like 05:15 in the morning right now. Try later.09:13
gcorriga gitster: thanks09:14
gitster which project are you interested in, may I ask?09:15
gcorriga gitster: I'm the admin for Squeak, and it looks like one student apply to both Git and Squeak.09:16
*applied09:16
gitster ah, "cross project" projects are always fun. I think there is another one we have between us and eclipse.09:18
gcorriga yes they are, but this isn't the case unfortunately.09:19
it's just that one student applied for a project for Squeak, and for a different project for Git09:19
gitster ... just "two separate jobs held by a single person"?09:19
gcorriga yes09:20
and since this can't happen under SoC rules, we have to coordinate with the Git mentors09:20
gitster I see. Well, I am not involved in GSoC so I should stop wasting your time. I'll drop a note to spearce to look at the #irc log and tell him to contact you.09:21
gcorriga gitster: thanks. And it wasn't a time waste ;)09:22
gitster message sent.09:26
→ gcorrig1 joined09:34
gcorrig1 → gcorriga09:38
→ hkBst joined09:48
hkBst I have a dump of a cvs repo that I want to convert to git. I unpacked it and then tried git-cvsimport, but it complains that cvs doesn't know the server command.09:51
→ chris2 joined10:20
→ pigeon joined10:44
corecode try fromcvs11:03
you'll need ruby and rcsparse though11:03
and some rbtree thingy11:03
pigeon hmm, i keep getting "fatal: refs/heads/origin: not a valid SHA1" when i try to git-cvsimport, any ideas?11:11
mugwump pigeon: delete the scratch directory or specify a branch11:12
pigeon scratch directory?11:12
i ran cvsimport in an empty dir11:12
mugwump by scratch I mean target11:17
pigeon is git-cvsimport supposed to create branches from cvs's branches?11:36
→ agoode joined11:45
chris2 pigeon: yes11:57
jwb_gone → jwb11:59
pigeon ok, cool, it seems to create branches properly with the right names, but i noticed some weird issues, where i find changes for one branch which should be in another12:05
hkBst corecode: to install rcsparse I need to do "setup.py <what?>"?12:22
corecode build12:22
then as root, install12:23
but12:23
uhm12:23
you need the ruby part12:23
ruby extmod.rb12:23
make && make install12:23
of course you can install the python part, but you don't need it12:23
hkBst corecode: I see no ruby part, so I am still confused12:24
corecode which rcsparse are you using?12:24
hkBst oh, right, extconf.rb12:24
→ yashi joined12:28
→ lcapitulino joined12:28
hkBst corecode: ok, I got it to build rcsfile.so. Now how do I tell ruby about it?12:29
or how do I make that into a ruby module?12:31
corecode make install12:31
hkBst corecode: ok thanks for your help, formcvs is doing something now :)12:52
corecode good12:53
oh12:53
you need togit12:53
not fromcvs12:53
but i guess you figured that out12:53
hkBst corecode: yeah, I was lying I was doing togit :)12:54
corecode k12:55
→ orospakr_ joined13:18
→ Oejet joined13:33
→ benlau joined13:55
hkBst corecode: togit seems to be stuck in a loop now. COnsole output has stopped.14:00
→ spearce joined14:00
→ GyrosGeier joined14:07
→ cworth joined14:17
corecode hm14:31
hkBst: still?14:31
hkBst corecode: quite a long time, but I've since killed it14:34
corecode what was the last output?14:35
some "committing..." stuff?14:35
how big is the repo?14:35
hkBst corecode: no, it was still going through the files. The repo is 1.4GB14:36
corecode hm.14:36
it probably was sorting + starting to commit14:36
oh14:36
no14:36
wait14:36
it is committing in parallel14:36
weird.14:36
which repo?14:36
hkBst gentoo-x8614:36
corecode how much ram do you have?14:37
hkBst 2GB14:37
corecode should be okay14:37
you have the backtrace from killing?14:37
hkBst no, unfortunately14:38
corecode hm.14:38
did it use some cpu?14:38
or did it sleep14:38
hkBst yeah it was 100% on one of my cores14:38
corecode too bad14:39
did it stop in the middle14:39
or was the scanning over14:39
hkBst corecode: in the middle as far as I can tell14:39
corecode weird14:39
maybe you can try to reproduce it with a part of the files14:40
hkBst corecode: how?14:43
dwmw2_gone → dwmw2_BOS14:45
→ kanru joined14:46
hkBst I'll try again, see if it hangs at the same spot, maybe give it more time14:47
spearce is it just me, or does `git apply -p2` not work?14:53
→ nud joined14:54
spearce ahhh.. if it is a git style patch, we throw the user's desires to the wind. cute.15:07
→ riddochc joined15:21
hkBst ok, it's there again15:30
→ hein joined15:30
wereHamster|zZz → wereHamster15:32
→ Astro joined15:32
→ troyt joined15:35
→ krh joined15:39
→ Eludias joined15:39
hkBst corecode: no output for 15 minutes, but plenty of CPU and hdd activity15:43
hein → Sho_15:46
→ alc joined15:47
bowens → tminos15:59
corecode hkBst: gdb it?16:04
hkBst corecode: :) ok16:04
corecode hkBst: just curious what it is doing16:04
hkBst: or strace or such16:05
hkBst corecode: http://rafb.net/p/cK67iu83.html16:08
corecode k16:09
looks good, no?16:09
let it run16:09
you have tons of data to be sorted16:09
hkBst corecode: ok, I'll be patient :)16:09
← Oejet left16:11
← gcorriga left16:16
→ Beg joined16:18
← Beg left16:22
→ orospakr joined16:27
→ tko joined16:31
hkBst still going :)17:00
→ alley_cat joined17:14
→ kumbayo joined17:17
kumbayo hi, does anybody know if .git/objects/info/alternates should work during initial push to a repo?17:18
for me the initial push pushes all the objects i have locally17:18
when i then git-gc in the remote repo the size decreases to some sane size of objects not covered by alternates17:19
until now i tried on push over ssh and on local push17:20
→ Romster joined17:24
tomprince The push detects what the other side has only by looking at the refs on the other side.17:26
So that if you don't have the refs from the alternates, it will push everything.17:26
This used to use a lot of bandwith, but no space, as we exploded the pack on the receiving side.17:27
But now that we keep the pack that was transfered on the wire, it doesn't help for space either.17:27
As you noticed, when we clean the repository we check the actual objects that are available.17:28
kumbayo hmm is it possible to create the ref without writing the objects somehow?17:30
hmm http://repo.or.cz/m/regproj.cgi?name=wine/ advertises, that you only need to push objects you created17:30
maybe it worked different in a older version? maybe HEAD was initialized to the parent repro or something17:31
tomprince I don't know anything about repo.or.cz17:32
If you have shell access on the server, you could just copy or clone from the alternates repository, and then delete the references afterwards.17:32
kumbayo ok, it just provides space to push to, and you can fork a project, and i guess it uses alternates to do that17:32
tomprince You might need to contact pasky to do that on repo.or.cz17:33
kumbayo i wrote a mail to him yesterday, but only tried to analyze furter by trying to push locally today17:34
will wait for his answert, thanks for clarification17:37
tomprince It looks looks like the scripts on repo.or.cz don't setup references to the base project17:40
so you don't get the benifit of alternates until the repository gets repacked.17:41
kumbayo with references you mean alternates?17:41
tomprince No, the protocol negotiates base only on the refrences. Right now, we don't look at the refrences in the alternates to determine what the other side has.17:42
Since for a new project, there are no refrences, we assume that we have to send everything.17:43
kumbayo might be a useful improvement for the future :-)17:44
tomprince This could perhaps be imporved, but we don't assume that the alternates points at an actual git repository, just the object directory from one.17:44
→ sgrimm joined17:51
→ lyakh joined18:11
→ rphillips joined18:43
→ T1 joined18:57
→ fultilt joined19:00
→ GeertB joined19:49
→ dst_ joined19:51
dst_ Hello! I'm trying to fetch from a repository over http from a repository on which git-update-server-info hasn't been run for quite a time. Is there a way to fetch the head anyway?19:54
I even know the sha1 of the current head commit from Gitweb19:54
ferdy http://rafb.net/p/9drVky74.html <-- does anybody know if this is acceptable for non-gnu make ?19:56
→ dduncan joined20:08
GyrosGeier hmm20:12
http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=blob;f=arch/arm/mm/init.c;h=7760193e74cc37fb29f632d673fe01b2dc6dc62c;hb=HEAD20:12
is the HTML broken for you too20:13
?20:13
(extra DIV element after </HTML>)20:13
dst_ doesn't look broken on first sight: the document ends in </html>20:16
GyrosGeier hmm20:16
last three lines for me are20:17
</body>20:17
</html><div class="page_body">20:17
</div>20:17
weird20:17
dst_ </div>20:17
</body>20:17
</html>20:17
GyrosGeier hmm20:18
now it works20:18
weird.20:18
→ Romster joined20:23
→ orospakr joined20:32
dwmw2_BOS → dwmw2_gone21:03
→ T1 joined21:09
→ fultilt joined21:12
→ Kristoffer joined21:20
→ Romster joined21:22
dwmw2_gone → dwmw2_BOS21:22
→ orospakr joined21:24
dwmw2_BOS → dwmw2_gone21:32
→ shevy joined21:45
shevy whoa this channel has grown since last time i was here21:46
jwb → jwb_gone21:49
wereHamster → wereHamster|away22:00
← shevy left22:02
→ dfarnsworth joined22:19
→ nud joined22:21
→ testing joined22:24
testing test22:24
test22:26
→ schemish joined23:10
→ robfitz joined23:16
→ joan joined23:16
→ Romster joined23:47
→ DrNick joined23:51

Logs Search ←Prev date Next date→ Channels Documentation