Camelia, the Perl 6 bug

IRC log for #darcs-theory, 2008-09-01

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

All times shown according to UTC.

Time Nick Message
07:20 kowey joined #darcs-theory
07:38 Peng_ But those IRC logs only go back to when the bot joined?
08:04 gal_bolle joined #darcs-theory
08:15 kowey Peng_: that's right
08:15 kowey this channel hasn't been around for very long though
08:15 kowey so we haven't "lost" much
08:16 kowey http://tinyurl.com/zfute for old #darcs logs and http://irclog.perlgeek.de/darcs for general #darcs as well
08:32 Peng_ kowey: Still, it would be nice to have complete logs. If the bot supports it, the people who have been here the longest may be able to piece something together.
08:32 Peng_ (But I'm a halfhearted lurker, so I shouldn't bother you. Also, I was gonna brush my teeth. Bye.)
08:34 kowey @tell lispy do you still have logs for #darcs-theory around? if so I might see about getting them integrated into the shiny new ones
08:34 kowey oh :-)
08:37 Peng_ I might be able to provide logs since late on 2008-08-18, but I'm afraid of my irssi client logging a "/msg nickserv identify" or something.
08:37 * Peng_ is really away now
10:02 mornfall --- Log opened Чтв Авг 14 00:40:59 2008
10:02 mornfall With a small gap, I should have everything since then.
10:02 mornfall (CEST zone)
10:03 mornfall Nothing much there though.
10:20 kowey cool... maybe you could put them on the web as a tarball, or email them to me
10:33 mornfall kowey: http://web.mornfall.net/stuff​/%23darcs-theory@freenode.log
10:38 kowey thanks
10:38 kowey I've /msg'd moritz_
14:58 kowey left #darcs-theory
15:26 kowey joined #darcs-theory
15:49 gal_bolle left #darcs-theory
16:37 arjanb joined #darcs-theory
20:32 arjanb Igloo: is definition 7.4 right? the name is the same the previous one
20:33 Igloo arjanb: The definition is right, the name is wrong
20:33 arjanb and what should it mean when the patches in the contexts partially overlap?
20:34 Igloo Contexted patches automagically maintain their invariant, so you have   p:r and p:s then then you make p^ p : s that magically turns into :s
20:34 Igloo Then you try to commute ^r past :s
20:35 arjanb so it's just a merge on sequence of patches?
20:35 Igloo Yes, it's equivalent to "do these sequences cleanly merge"
20:36 arjanb ok
20:36 arjanb only got a little confused by the notation
20:36 Igloo It should have more explanation, sorry
20:37 Igloo e.g. it should actually say that they start from the same context
20:45 arjanb hmm this makes conflictor commute quite costly
20:45 Igloo It's cubic, I think
20:46 Igloo Which is way better than darcs 1 and 2  :-)
20:46 Igloo (I'm pretty sure 2 is still exponential)
20:47 arjanb so conflictors can contain 'catches'?
20:47 Igloo In my world? No
20:47 Igloo Contexted patches are made of a sequence of patches and a patch
20:47 Igloo In darcs they contain catches
20:47 Igloo darcs2, that is
20:48 Igloo (which is why I think darcs2 is exponential)
21:06 kowey joined #darcs-theory
21:07 Heffalump has anyone tried to stress darcs2 out with random records?
21:07 Heffalump and pulls and stuff, obv
21:08 Igloo I don't know, but if they have, please give me the script!
21:10 Igloo Heffalump: With GADTs, I can't ever return Nil from a function that's supposed to conjure up a (Seq from to), right?
21:10 Heffalump sorry, I don't follow the question.
21:11 Igloo e.g. read :: String -> Sequence Patch from to    (where from and to are contexts)
21:11 Igloo That can't return Nil :: Sequence Patch x x, right?
21:11 Igloo I'd have to wrap it up in something to hide one or both of from and to?
21:15 Heffalump I'm still not with you. If from = to then obviously it can return that.
21:16 Heffalump But in that type read is in theory forced to make up a patch of whatever type the caller asks for.
21:16 Heffalump Which is not really possible, and leads to unsafeCoerce nastiness like I was talking about with lispy a few days ago
21:16 Igloo *nod*. I guess I have to wrap it up.

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