Camelia, the Perl 6 bug

IRC log for #darcs, 2013-04-08

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

All times shown according to UTC.

Time Nick Message
01:26 mizu_no_oto joined #darcs
01:37 carter joined #darcs
01:38 mizu_no_oto joined #darcs
01:38 edwardk joined #darcs
02:11 mizu_no_oto joined #darcs
02:52 preflex_ joined #darcs
02:55 whaletechno joined #darcs
04:52 Heffalump sm: not entirely sure - I killed first and asked questions later
04:52 Heffalump I think it was using ~160MB
05:34 Thaalos joined #darcs
06:22 lelit joined #darcs
06:42 donri joined #darcs
07:04 raichoo joined #darcs
07:11 raichoo joined #darcs
08:02 nomeata joined #darcs
09:14 owst joined #darcs
09:58 alexei joined #darcs
10:33 vikraman- joined #darcs
11:31 vikraman joined #darcs
11:31 vikraman joined #darcs
11:54 lelix joined #darcs
12:20 mizu_no_oto joined #darcs
12:50 iago joined #darcs
12:50 alexsuraci__ joined #darcs
13:25 whaletechno joined #darcs
13:26 mizu_no_oto joined #darcs
13:52 mornfall ProgrammingError: operator does not exist: text = integer
13:52 mornfall lelix: Have you tried using tracdarcs on top of postgres?
13:53 * mornfall tries a resync
14:02 mizu_no_oto joined #darcs
14:02 mornfall Nope. I guess the query is missing quotes somewhere.
14:09 mornfall Hm. Au contraire. Something is text while it should be integer.
14:15 lelix` joined #darcs
14:26 mornfall lelix`: Nevermind, it works with current darcs head of trac-darcs.
14:27 mornfall No idea why though. :-)
14:50 lelix` mornfall: I adjusted it recently
14:51 lelix` too bad it's still not very efficient: there's the main SQL query which I was not able to optimize (yet)
14:51 lelix` but it is very tricky to get right
14:55 mornfall lelix`: I see. It works quite well for us though.
14:56 mornfall Trac was really sick when it was running on a sqlite over NFS, but I moved it to postgres and it's back to good.
14:56 lelix` yes :) the problem manifests itself with medium-to-big repos
14:57 mornfall The resync seems to slow down though as it progresses.
14:57 lelix` yes
14:57 mornfall How big is big? :)
14:58 mornfall GHC big?
14:58 lelix` that very same query is used a) to list the content of a directory and b) at resync time, to determine the current cached version of a changed item
15:00 lelix` for me "big" was the darcs own repo that I used for the tests: at the time I only had a rather slow netbook, and it took two days to complete the resync X-)
15:01 lelix` then I was able to spot the problem, and to go from over a minute-per-execution to under a second... unfortunately introducing a regression in source browse :-\
15:02 lelix` then my brain exploded and I had to find something else to keep it busy :)
15:03 lelix` I intend to spend more time on it though
15:05 dolio joined #darcs
15:09 lelix joined #darcs
15:51 donri joined #darcs
16:37 stepcut joined #darcs
16:52 donri_ joined #darcs
17:01 xymox joined #darcs
17:49 alexei joined #darcs
18:35 edwardk joined #darcs
19:03 carter joined #darcs
19:40 mizu_no_oto joined #darcs
19:54 javier_rooster joined #darcs
20:01 dcoutts joined #darcs
22:02 carter joined #darcs
22:58 saep joined #darcs
23:02 dolio joined #darcs

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