Camelia, the Perl 6 bug

IRC log for #darcs, 2010-11-22

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:09 Heffalump I found a workaround.
00:10 copumpkin ooh
00:11 intripoon_ left #darcs
00:20 intripoon joined #darcs
00:23 raichoo joined #darcs
00:39 copumpkin left #darcs
01:01 raichoo left #darcs
01:18 Funktorsalat left #darcs
01:24 copumpkin joined #darcs
01:48 twb joined #darcs
03:12 intripoon_ joined #darcs
03:15 intripoon left #darcs
04:14 pumpkin joined #darcs
04:16 copumpkin left #darcs
04:24 peterhil left #darcs
04:54 slyfox joined #darcs
04:54 slyfox left #darcs
04:54 slyfox joined #darcs
05:26 slyfox left #darcs
05:35 lispy_ left #darcs
06:52 sm left #darcs
06:52 lispy joined #darcs
07:05 jutaro joined #darcs
07:24 lispy left #darcs
07:58 balor joined #darcs
08:01 darcscommitbot left #darcs
08:02 darcswikibot left #darcs
08:02 darcscommitbot joined #darcs
08:03 darcswikibot joined #darcs
08:05 darcsohlohcommit left #darcs
08:06 darcsohlohcommit joined #darcs
08:09 mornfall Seems that the response code patch makes darcs segfault on OSX. (Or maybe it's something else, I am just guessing.)
08:09 stepkut left #darcs
08:09 mornfall See buildbot.
08:09 stepkut joined #darcs
08:12 lelit joined #darcs
08:23 ManateeLazyCat joined #darcs
08:24 JaffaCake1 joined #darcs
08:25 JaffaCake left #darcs
08:30 peterhil joined #darcs
08:39 ManateeLazyCat left #darcs
08:41 kowey joined #darcs
08:48 kowey good morning!
08:48 lambdabot kowey: You have 2 new messages. '/msg lambdabot @messages' to read them.
08:51 kowey oh my, lots of patches in
08:51 kowey maybe time to do an optimize --reorder on unstable now that we have that 2.5 tag in
09:04 raichoo joined #darcs
09:05 kowey if issue2001 really was a case of unison eating a file, then yay hashed repositories
09:06 twb left #darcs
09:06 kowey this is the sort of thing that would have resulted in a corrupted old-fashioned repo, ticking time-bomb style
10:11 exlevan joined #darcs
10:11 exlevan morning
10:25 gh_ joined #darcs
10:35 gh_ left #darcs
11:06 JaffaCake1 is now known as JaffaCake
11:33 ManateeLazyCat joined #darcs
11:47 kowey hi exlevan, mornfall (I wonder how patch191 broke the windows slave)
11:47 mornfall Hey.
11:48 kowey I would have guessed a more likely culprit to have been SSH URI support
11:48 mornfall Oh?
11:48 kowey just from the look of the error message, not any fancy thinking mind you
11:48 * mornfall logs into the windows box
11:48 mornfall I simply didn't notice any other patches touching the functionality.
11:48 mornfall So that was my deduction based on that.
11:49 mornfall I hate how ^C doesn't work with haskell programs on windows.
11:52 ManateeLazyCat left #darcs
12:04 intripoon joined #darcs
12:07 intripoon_ left #darcs
12:07 mornfall Building anything with cabal is a total mess now, thanks to mtl 2.
12:08 dcoutts you can pick one or the other using --constraint='mtl >= 2'  or 'mtl < 2'
12:08 mornfall dcoutts: Yes, but you still get both linked in.
12:08 dcoutts um, what?
12:09 dcoutts you mean when doing cabal configure or install?
12:09 mornfall Well, my QC-2 and regex-base is built with mtl 2.
12:09 mornfall But plenty of packages don't built with it.
12:09 mornfall build*
12:09 dcoutts yes, cabal configure has to live with what packages are installed already
12:09 dcoutts cabal install --dry-run --constraint='mtl >= 2'  will show you what it'd rebuild to make things work out consistently
12:09 mornfall What ticked me off somewhat is that darcs picked up mtl 2 when I used -ftest, but mtl 1 otherwise.
12:10 mornfall (Without explicit constraint, that is.)
12:10 mornfall Hm. Can it actually pick different mtl for different binaries in the package?
12:10 mornfall (Darcs does not build with mtl 2, but darcs.exe did build, while darcs-test.exe failed.)
12:11 mornfall Ok, it works now (luckily, it doesn't seem to be a problem to have both mtls linked in.)
12:11 dcoutts mornfall: no, all components of the package get consistent dependencies
12:11 mornfall dcoutts: In that case, it failed to rebuild some parts of darcs.
12:11 dcoutts anything else would be really confusing :-)
12:11 mornfall After reconfiguring.
12:12 mornfall Since Darcs.IO is used in both darcs.exe and darcs-test.exe and it only errored out in the latter.
12:12 dcoutts that could be a ghc bug, what ghc version?
12:12 mornfall (Due to overlapping instance with mtl 2.)
12:12 mornfall 6.12.3
12:12 dcoutts mornfall: sounds like it could  be a bug in ghc --make 's recompliation checker
12:13 dcoutts if you can reproduce it then report it with as much detail as possible to help the ghc hackers reproduce it
12:16 sm joined #darcs
12:17 mornfall The general bugginess of Haskell software is worrisome though. I have the feeling that it's often more buggy than more "traditional" software, not less.
12:20 mornfall kowey: You are right, it's not Alexey's patch but the SSH URI one.
12:20 exlevan left #darcs
13:15 gal_bolle joined #darcs
13:27 kowey ah just right person! hey gal_bolle
13:27 kowey there's a chance you might be saying "I hate Windows" soon
13:33 gal_bolle am i not already?
13:35 kowey :-) did you spot the '"ssh": Could not resolve hostname C: no address associated with name' on darcs-devel?
13:36 kowey buildbot++ did its job
13:38 gal_bolle that might be because of ssh:// urls
13:39 gal_bolle i hate windows, but not for that reason
13:40 gal_bolle (line endings)
13:42 kowey oh ugh too (or ugh 2)
13:43 kowey and once upon a time Macs used \r only (I think)
14:05 pumpkin left #darcs
14:12 mornfall Yes, they did.
14:13 mornfall Also, the OSX slaves have darcs segfaulting now.
14:13 mornfall In network tests.
14:13 mornfall (See backlog, I suspect it *might* be related to response error codes. Don't take my word for it, though.)
14:20 dixie hmm, I got idea to install darcsweb on windows for viewing repostories. How hard it is ? :)
14:35 abuiles joined #darcs
14:52 copumpkin joined #darcs
15:23 dixie it works.
15:40 gal_bolle kowey: windows problem solved, i think
15:41 mornfall gal_bolle: You are in the core team nowadays?
15:42 mornfall (If so, grabbing buildbot-0.8 and running buildbot-try from contrib could help you confirm that...)
15:53 gal_bolle yes
15:54 gal_bolle thanks for the tip
15:54 * gal_bolle hopes to be ready for release management before release time comes
15:59 raichoo left #darcs
16:09 mornfall gal_bolle: (Btw. you can expect one failing test, which I so far failed to fix, so the result being red doesn't mean your fix didn't work.)
16:09 mornfall Check the output.
16:10 gal_bolle ok, thanks
16:15 gal_bolle how do I check which try is mine?
16:18 sm mornfall: interesting comment (bugginess of haskell software) - I don't think that's my experience. It's not a magic bullet though
16:19 sm plenty of bugs in darcs, for sure
16:35 raichoo joined #darcs
16:41 lispy joined #darcs
16:41 sm_ joined #darcs
16:42 mornfall And Xmonad, GHC, Cabal, of the tools that I use daily.
16:42 mornfall Although the recent floating layer hack made xmonad much more usable. :)
16:42 mornfall It's still far from perfect.
16:42 sm left #darcs
16:42 sm_ is now known as sm
16:43 dcoutts left #darcs
16:44 mornfall It may be related to the fact that I use these tools a lot, so I run into bugs in them more often.
16:45 mornfall Or maybe I just have higher expectations of them, given how they are Haskell.
16:47 mornfall Also, I suspect that a lot of what you gain on referential transparency, you lose on laziness bugs.
16:47 mornfall Since laziness *is* tricky.
16:55 lispy I've been using xmonad for a year now and only discovered  the following bugs: 1) Java doesn't work well with it due to a java bug, 2) sometimes it puts my windows on the wrong desktop due this is due to problems with the X protocol and not Xmonad's fault, 3) I have to recompile/restart xmonad manually for some reason
16:56 lispy I'd say xmonad is one of the least buggy things on my entire linux install
16:56 lispy Gnome is about 100x more buggy, IMO
16:59 etarasov joined #darcs
17:04 secorp left #darcs
17:20 raichoo1 joined #darcs
17:21 raichoo left #darcs
17:24 gh_ joined #darcs
17:26 mornfall lispy: Interesting. The way how floating windows never worked on top of fullscreen windows always ticked me off horribly. There's a contrib extension with a partial fix now.
17:26 mornfall lispy: Also, there have been memory leaks in xmonad with decorations.
17:27 mornfall (Which is a bug in contrib, arguably, but it still broke xmonad.)
17:28 mornfall I agree that it's still the most usable window manager out there. But I am almost sure that wmii/awesome are more "stable" (albeit less featureful).
17:28 mornfall Similar argument could be made of darcs.
17:35 dcoutts joined #darcs
17:35 dcoutts left #darcs
17:35 dcoutts joined #darcs
17:39 sm buggy: a project whose complexity currently exceeds it's available design clarity/testedness/maturity
17:42 lispy ?tell iago http://hackage.haskell.org/package/Graphalyze  <-- might help you with code metrics
17:43 lispy ?tell iago http://hackage.haskell.org/package/Graphalyze  <-- might help you with code metrics
17:43 lambdabot Consider it noted.
17:43 lispy lambdabot: space sensitive?
17:45 kerneis http://rwmj.wordpress.com/2010/11/19/half-b​aked-ideas-git-commit-dependency-analysis/ << he uses git cherry-pick but what he wants is darcs, actually
17:48 lispy ?tell iago oh, and dons showed me this link too: http://code.haskell.org/~ivanm/Sampl​e_SourceGraph/SourceGraph/SourceGrap​h.html#overall-cyclomatic-complexity
17:48 lambdabot Consider it noted.
17:48 kowey kerneis: the recent thread about spontaneous branches etc may be of interest; didn't have time to read to follow-ups
17:48 kowey (on darcs-users that is)
17:48 kowey left #darcs
17:49 kerneis Richard Jones already posted sth about cherry-picking actually:
17:49 kerneis http://rwmj.wordpress.com/201​0/05/14/git-cherry-pick-wins/
17:49 kerneis (and I already told him to use darcs :-)
18:02 secorp joined #darcs
18:04 abuiles left #darcs
18:28 thorkilnaur mornfall, You guessed right: Following darcs unpull -p 'resolve issue1637: report HTTP response code for HTTP errors', the 6.10.4 OSX test network succeeds.
18:31 mornfall thorkilnaur: Thanks for diagnosing that.
18:54 lelit left #darcs
19:01 zooko joined #darcs
19:11 lispy left #darcs
19:17 iago joined #darcs
19:24 slyfox joined #darcs
19:24 slyfox left #darcs
19:24 slyfox joined #darcs
20:21 exlevan joined #darcs
20:38 arjanb joined #darcs
20:47 kowey joined #darcs
20:48 peterhil left #darcs
21:18 mornfall kowey: I guess we are lucky that we always find a successor (re RMs). : - )
21:18 * kowey was very relieved
21:39 balor left #darcs
21:51 kowey left #darcs
22:22 copumpkin left #darcs
22:22 jutaro left #darcs
22:51 sm_ joined #darcs
22:51 sm1 joined #darcs
22:51 sm left #darcs
22:51 sm1 is now known as sm
23:02 alexsuraci left #darcs
23:10 copumpkin joined #darcs
23:22 slyfox left #darcs
23:23 gh_ left #darcs
23:35 raichoo1 left #darcs
23:56 sm__ joined #darcs

| Channels | #darcs index | Today | | Search | Google Search | Plain-Text | summary