IRCloggy #git 2007-02-03

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.

2007-02-03

corecode robinr: see README00:09
robinr no README!(?), discovered that I need another RCSPARSE00:18
← Drakeson left00:27
robinr togit.rb:154:in `finish': undefined method `success?' for #<Process::Status: pid=8647,exited(0)> (NoMethodError)00:34
from ./fromcvs.rb:933:in `commit_sets'00:34
from togit.rb:24300:34
..and partial conversion00:39
second conversion looks better, but it can't be that fast...00:41
at least the last commit is ok00:42
I'll see tomorrow what the full conversion gives00:45
it does allow spaces in filenames, does it?00:45
→ GeertB joined01:06
→ kumbayo joined01:11
→ GyrosGeier joined01:14
→ qfyre joined01:22
→ robfitz joined02:00
→ spearce joined02:03
→ DebolazX joined02:09
→ Myen joined02:17
→ DebolazY joined02:28
→ orospakr joined02:35
→ benlau joined02:43
→ segher joined04:08
→ GyrosGeier joined05:06
→ GyrosGeier joined05:08
→ crime_minister joined05:14
→ rambokid joined05:28
→ _crime_minister joined05:54
→ Romster joined06:43
→ Oejet joined07:12
→ devogon joined07:30
→ DebolazX joined07:34
→ DebolazY joined07:52
devogon mutt07:59
Oejet pine08:33
→ Romster joined08:36
corecode yo08:48
Oejet Hi, corecode.08:48
corecode robinr: yes08:48
robinr: you found the README?08:48
robinr: seems your ruby is too old?08:48
robinr: what do you mean with "it can't be that fast"08:49
→ Eludias joined09:00
→ DebolazX joined09:04
→ DebolazY joined09:12
robinr corecode: a joke09:52
the ruby version is the one that's on SLES909:53
corecode well dunno09:53
ruby -v?09:53
robinr 1.8.109:54
the bigger run was 42 minutes09:55
corecode m09:58
1.8.109:58
→ nud joined09:58
corecode could you try to check how i am supposed to check for a process exit status?09:58
instead of finish?09:58
robinr ? I don't know ruby10:00
→ lyakh joined10:01
corecode ri Process::Status10:02
should give you a list of methods10:02
chris2_ → chris210:10
→ chris2 joined10:11
robinr it seems not docs are present10:20
I have another machine with 1.8.5 and a doc package10:22
..and no disk free10:29
→ alley_cat joined10:30
robinr in 1.8.5 there is a success? method10:33
but apparently not 1.8.110:33
do ruby introduce new methods in minor versions?10:33
corecode seems so10:34
1.8.1 is quite old10:34
robinr Still, it's the third digit10:35
corecode yah, not my idea10:39
so if you find out how i can check if there was success or failure, i'd change it10:39
i hope you are happy with the conversion result10:40
robinr I need to to it incrementally10:40
corecode it works incrementally10:40
if you didn't notice10:40
robinr ah :)10:40
corecode i don't handle repo surgery yet, thouhg10:40
robinr not necessary10:41
Rerun with the same arguments?10:41
corecode yep10:42
you shouldn't run while committing, though10:42
and don't commit to the branches10:42
the repo should be read-only10:42
or weird things happen10:42
and i don't convert tags (yet)10:43
(patches welcome)10:43
robinr ok, I do that with the old cvs-imported already10:43
corecode do what?10:43
robinr use a conversion-only-repo10:43
corecode k10:43
→ lhz joined10:53
corecode is there a reason why git-mailinfo doesn't handle nested multipart messages?11:00
i'd like it to apply signed mails as well11:00
hm.11:04
git log -p isn't as powerful as git blame11:04
robinr you mean it doesn't track moves?11:07
corecode: the reason is probably that noone has written the code11:08
corecode k11:08
robinr: yes, doesn't follow them11:08
robinr unless there is a -M-flag, but I don't think so11:08
corecode there is11:11
well, for git-diff-tree11:11
hm.11:11
hack in multipart/signed tolerance11:11
or implement a proper mime parser?11:12
hm11:16
anybody wanna flip a coin for me?11:16
clee heads.11:17
corecode you gotta translate that11:17
clee hack it11:17
corecode k11:17
sure?11:17
clee yep11:17
corecode why?11:17
clee because that's what the coinflip said, dude11:18
heads = hacks11:18
corecode k and your personal opinion?11:18
clee I always do the quick solution first11:18
and then go back later to do the optimizing11:18
corecode writing a proper mime parser isn't an optimization11:21
it is *rewriting811:21
clee it's optimizing design11:22
corecode yah11:23
but when you do that anyways, why bother hacking it first?11:23
→ devogon joined11:26
corecode how would I use git log to pickaxe not only changes, but changed lines?12:09
so if i changed foo(bar) to foo(spam), i'd like to find this change with git log -Sfoo12:10
is that somehow possible?12:10
robinr I think it is..wait12:22
→ GeertB joined13:41
→ oleg_erm joined14:11
→ spuk joined14:11
→ someone_ joined14:29
→ bonesss joined14:35
← bonesss left14:36
→ drizzd joined14:36
→ someone_ joined14:36
→ someone__ joined14:37
→ someone_ joined14:46
→ Oejet joined15:02
→ benlau joined15:17
→ kanru joined15:49
→ lyakh joined15:52
→ orospakr joined15:59
→ z3ro joined16:07
→ vlajos joined16:10
→ dufflebunk joined16:17
robinr corecode: not happy, it's not the right code :(16:19
corecode what?16:20
robinr: what do you mean16:22
robinr master of the git repo and the CVS repo differs16:23
the last CVS commit is missing, when looking at one particualar file16:27
could you kill keywords?16:31
good news is that it should be possible to track. All diffs come from the same cvs commit16:35
the missing changes were made on a branch and then merged to HEAD (in CVS)16:39
corecode kill keywords?16:40
robinr $Id$ instead of $Id: blahblah$16:40
corecode how is the last commit missing?16:40
just configure CVS not to expand the keywords16:41
robinr hmm16:41
corecode so, how is it missing?16:42
robinr the changes are not in in the master branch of the converted repo16:42
corecode you simply committed them to HEAD?16:43
you're kinda lagging16:45
robinr well, cvs update -j ofcourse. Any better way?16:45
corecode ah16:45
and then cvs commit?16:45
robinr yes16:45
corecode then then you ran the convert tool?16:46
and*16:46
robinr the merge I'm referring to is over a year old16:46
corecode ah!16:46
alright16:46
robinr old history16:46
corecode so head is missing this one particular commit16:46
robinr yes16:46
corecode master16:46
strange16:47
is this a private repo16:47
or can i have a look at it16:47
→ ferdy joined16:48
robinr it is private, but I can probably extract something16:48
corecode it is very strange that it completely omits a change16:49
robinr I'll see if I can cut out a few files16:49
and bring it down to something simpler16:49
corecode try with just one file16:49
robinr yepp16:51
I've done this kind of excercise before :)16:52
→ kampasky joined16:52
corecode you can use fromcvs directly16:53
to get a list of commits16:53
that's basically a print-only target16:53
robinr aha16:53
how?16:54
ah16:55
'twasn't that easy17:05
corecode hm?17:06
robinr I'll descend the conversion down one level at a time and see where it stops breaking17:07
with your script the every run is quite short :)17:09
corecode the commit is just missing?17:22
robinr yes17:25
I diff'ed a cvs checkout with the converted repo17:25
and the last change in CVS on a particular set of files are gone17:26
if I reset the cvs checkout one revision for those file, everything matches17:26
I'm down to nine files17:29
better than the original 2000017:29
but now it gets harder17:30
robinr <3 checkin comments17:30
robinr I know places where they are rare17:30
so, now it gets harder...17:31
eight files17:44
one file!17:46
git is nice. I put a copy of a subset of the repo under CVS before bisecting it17:47
corecode so?17:59
→ nud joined18:02
orospakr can gitk or tcl/tk in general be made to, um, use anti-aliased fonts?18:07
or is that really asking for too much? ;)18:07
→ nud_ joined18:19
robinr corecode: it simplifies adding/removing cvs files18:26
→ Romster joined18:27
→ nud__ joined18:38
robinr hmm, looks like I have the same commit message on a branch AND on HEAD (differenct files though)18:40
that could confuse someone...18:43
so it seems the file that was changed on the branch got commited by the converter to the branch and the other changes were dropped18:45
→ z3ro joined19:01
→ nud joined19:02
→ alley_cat joined19:11
→ jeffpc joined19:23
→ octal joined19:29
→ crime_minister joined19:49
→ nightbreed joined19:51
nightbreed hello19:51
git: fatal error: `chdir' failed: permission denied19:54
whats the deal with this error19:54
→ drizzd joined20:12
→ GeertB joined20:31
→ maks_ joined20:55
→ orospakr joined21:07
Oejet nightbreed: Perhaps a directory in .git/ does not have read/execute permission?21:37
nightbreed ok21:38
→ DrNick joined21:44
→ ilogger2 joined22:34
orospakr Hey, clee.22:45
long time no chat.22:45
clee orospakr: sup, dude?22:45
orospakr clee: oh, hacking on OLPC stuff.22:46
and waiting for a buddy to come over.22:46
→ qfyre joined22:52
→ GeertB joined23:37

Logs Search ←Prev date Next date→ Channels Documentation