Camelia, the Perl 6 bug

IRC log for #darcs, 2010-03-08

← Previous day | Index | Channel Index | Today | Next day → | Search | Google Search | Plain-Text | plain, newest first

All times shown according to UTC.

Time Nick Message
01:38 alexsuraci joined #darcs
02:33 abuiles joined #darcs
06:34 azathoth99 joined #darcs
06:35 azathoth99 I read some crap about darcs not scaling?
06:35 how does darcs compare to say git?
07:28 alexsuraci ah, the old in-n'-out
07:46 balor joined #darcs
07:53 gh_ joined #darcs
07:56 lelit joined #darcs
08:02 darcscommitbot joined #darcs
08:03 darcswikibot joined #darcs
09:09 Korusef joined #darcs
09:25 gh_ joined #darcs
09:42 Alpounet joined #darcs
09:54 attila_lendvai joined #darcs
09:58 attila_lendvai is it by design that darcs started to folow symlinks? after upgrading to 2.4, i had to extend my .boring file to ignore some symlinks
10:05 intripoon joined #darcs
10:51 Alpounet joined #darcs
10:52 exlevan joined #darcs
13:42 alp joined #darcs
13:44 Alpounet joined #darcs
13:56 gal_bolle joined #darcs
14:00 stevan joined #darcs
14:04 stevan hi, if i obliterated a patch locally how can i push the change so to say? record says that no change happend. thanks.
14:05 attila_lendvai stevan, obliterate is delete...
14:06 stevan ok?
14:08 gwern joined #darcs
14:08 stevan i pulled a patch which i didn't want, what's the proper way of undoing that? obliterate seemed to work well for my local copy, but the global still has the bad patch. i'd like to undo it globally too.
14:09 rollback?
14:13 gh_ yes
14:13 since that patch was already in the global repo,
14:14 stevan damn it :-/, so i guess the only way out of this mess i created is to remove the file and then readd it?
14:14 gh_ people may have already pulled it
14:15 the global repo you mentioned is a repo other people have access too, right ?
14:16 stevan yes, but it's only a friend and i who use it
14:16 abuiles joined #darcs
14:22 gh_ well then you can obliterate everywhere
14:22 stevan what if i darcs get a new copy and use rollback there instead?
14:25 gh_ you can as well do that rollback in your own repo
14:25 rollback just creates an inverse patch of one or several patches
14:26 stevan is that possible even after i obliterated them?
14:26 gh_ no
14:26 rollback needs the original patches
14:26 it can not just invent them :)
14:27 but for instance if you recorded a big binary by mistake, it may make sense to obliterate everywhere
14:27 if it's just minor, it's better t odo a rollback
14:29 stevan ok, thanks for the help.
14:41 abuiles left #darcs
14:50 stevan left #darcs
15:04 gal_bolle joined #darcs
16:06 gh_ joined #darcs
16:11 sea-gull joined #darcs
16:12 sea-gull left #darcs
16:27 solistic joined #darcs
16:43 gbeshers joined #darcs
16:44 solistic How can I see the formt of a given repository?
16:53 wferi joined #darcs
16:54 wferi mornfall: why do you say in http://bugs.darcs.net/msg10122 that my patch is bogus?
16:55 it's the same as your pastebin link.
17:00 Alpounet joined #darcs
17:07 kowey joined #darcs
17:08 kowey hello, solistic (just saw your question in the online chat log)
17:08 try 'darcs show repo'
17:15 balor joined #darcs
17:22 lisppaste2 joined #darcs
17:25 lispy1 joined #darcs
17:26 lispy1 joined #darcs
17:33 solistic joined #darcs
17:33 beschmi joined #darcs
17:33 mornfall wferi: I say that because your context says ReadWrite where it should say ReadWriteEx.
17:34 wferi: And it's ReadWriteEx after my patch.
17:35 So either you applied wrong patch, or you are working with wrong version of h-s or something like that.
17:36 Your reported symptoms pretty much match the behaviour one would expect from a h-s with ReadWrite in place of ReadWriteEx, using mmap 0.5.
17:36 These 2 things lead me to believe that you indeed have a version of h-s linked against mmap 0.5 but with ReadWrite.
17:45 wferi mornfall: yes, you are right. I applied the patch by hand, and thought that line contained whitespace difference only. I even overlooked the Ex when I specifically looked for a problem... My bad.
17:45 thanks for pointing this out.
17:46 Btw. if I link h-s with mmap-0.5, then install darcs, should Cabal notice that h-s was build with mmap-0.5, and refuse to build darcs with mmap-0.4?
17:47 How can I (or cabal) find out which version of mmap h-s was built with, after the fact?
17:50 mornfall Unfortunately, you can't. You can find out what darcs has been built with, using --exact-version, if you are using HEAD.
17:50 Well, ghc-pkg may know, actually.
17:50 wferi Then how is cabal supposed to stop problems like this?
17:51 mornfall Anyway, Cabal will warn you about building with two different versions of a package, but unfortunately it will proceed.
17:51 Ack, ghc-pkg dump will give you enough data to reconstruct the build-deps used.
17:52 wferi Hmm, ghc-pkg describe gives some "depends:" info, maybe that's what I'm looking for.
17:52 david_ joined #darcs
17:52 mornfall You can't find the used builddeps from a binary, after-the-fact, though.
17:52 Yes, describe -> depends is the thing you are looking for.
17:53 wferi Those aren't the builddeps, but the actual chosen versions, aren't they?
17:53 mornfall Yes.
17:53 wferi The builddeps themselves aren't particularly interesting after the fact, I think.
17:53 Ok, thanks.
17:54 So isn't cabal supposed to stop when several different versions of a package are present in the depends chain?
17:54 Or should that really work?
17:55 mornfall It warns about the fact. IMHO, it should stop, but it doesn't, for whatever reason.
17:55 You could take that up with Duncan (dcoutts).
17:55 wferi He'll probably chime in if he's present.
18:01 solistic kowey: thx :)
18:03 solistic joined #darcs
18:41 exlevan joined #darcs
18:55 exlevan how do i save a patch in darcs? e.g. darcs save-patch -p X > x.dpatch
18:59 beschmi do you want something that you can feed to patch(1) or a darcs patch bundle?
19:01 attila_lendvai joined #darcs
19:02 exlevan i want to save darcs pach bundle to file
19:02 beschmi for the later, see darcs send -o foo.dpatch
19:04 exlevan thanks, looks like what i wanted
19:09 beschmi are there redundant patches in the _darcs/patches in case of conflicts?
19:10 e.g., you pull patch p2 which conflicts with p1 already in the repo.
19:11 then you get a conflictor C(p2,p1) which is referenced in the inventory, but p1 is kept in _darcs/patches anyways
19:13 solistic joined #darcs
19:13 beschmi nevermind, I think I get it now: p1 is redundant. If you unpull p2 from p2.C(p2,p1), then p1 can be reconstructed from the conflictor.
19:32 mornfall Hi beschmi! I keep forgetting to reply to your mail...
19:33 My ticket says I should arrive 19.3. at 9:17 am to Zürich HB.
19:38 beschmi hi mornfall
19:41 solistic joined #darcs
19:41 beschmi ok, I haven't decided yet if I want to take the full day off or go to the office in the morning
19:44 edwardk joined #darcs
19:45 beschmi but I can pick you up at the train station and we can put your luggage (if you have anything that you don't want to carry around) in my office
19:45 mornfall Hmm.
19:46 I tend to travel light if at all possible.
19:47 So probably arrange your day however you want. If it suits you to meet up earlier, works for me, or we can meet at ZuriHac at 2 pm.
19:47 I can look around, or maybe Eric organizes some darcs pre-hack?
19:48 beschmi ok, I just saw that tux_rocker is arriving already on the 18th
19:54 tux_rocker joined #darcs
19:55 tux_rocker joined #darcs
19:56 attila_lendvai joined #darcs
20:11 solistic joined #darcs
20:28 tux_rocker_ joined #darcs
21:04 Alpounet joined #darcs
21:13 solistic joined #darcs
21:35 Alpounet joined #darcs
22:13 aavogt joined #darcs
22:13 aavogt is there such a thing that would be called unapply?
22:14 Igloo You can fold definitions
22:15 Oh, ignore that
22:15 You probably want one of obliterate, unrecord and rollback
22:20 aavogt yeah, I was hoping to be able to get rid of all the patches not present in a repo that's pretty slow to get
22:29 kpreid_ joined #darcs
23:16 kpreid_ joined #darcs
23:19 abuiles joined #darcs
23:22 kpreid_ joined #darcs

← Previous day | Index | Channel Index | Today | Next day → | Search | Google Search | Plain-Text | plain, newest first