Perl 6 - the future is here, just unevenly distributed

IRC log for #darcs, 2014-11-06

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

All times shown according to UTC.

Time Nick Message
01:27 gh_ joined #darcs
02:21 stepkut joined #darcs
02:59 gh_ I suspect the space leak problem of patch index creation comes from the fact that it is not written to disk in a very "pipelinesque" way
03:00 gh_ http://hub.darcs.net/darcs/darcs-screened/browse/src/Darcs/Repository/PatchIndex.hs#501
03:01 gh_ first we write the whole repoStateFile, then touchMapFile, then fidMapFile, then fpMapFile
03:02 gh_ while to avoid the space leak we should write all of them in parallel while consuming pids
03:25 mizu_no_oto joined #darcs
03:45 mizu_no_oto joined #darcs
04:31 lambdabot joined #darcs
06:21 stepkut joined #darcs
06:37 lelit joined #darcs
06:44 masse joined #darcs
09:02 c74d joined #darcs
09:19 c74d joined #darcs
12:07 mizu_no_oto joined #darcs
13:48 mizu_no_oto joined #darcs
15:00 byorgey joined #darcs
16:12 mizu_no_oto joined #darcs
16:43 lelit joined #darcs
17:04 amgarchIn9 joined #darcs
17:38 amgarchIn9 joined #darcs
17:59 gh_ joined #darcs
18:04 amgarchIn9 joined #darcs
18:07 Heffalump hi
18:26 gh_ Heffalump, hi
18:27 gh_ I got caught in trying to understand patch index
18:31 mizu_no_oto joined #darcs
18:31 gh_ anyway I guess I can branch now
18:53 gh_ ok we have http://darcs.net/releases/branch-2.10/ , and http://darcs.net/releases/branch-2.10/branch-2.10.zip
18:53 gh_ the latter being generated again each time we will push into this repo
18:55 dolio joined #darcs
19:11 gh_ Heffalump, did we plan to drop some older ghc version before the release?
19:28 * sm mirrors the 2.10 branch on darcs hub
19:54 gh_ it would be nice to have github mirrors..
20:03 sm joined #darcs
20:17 byorgey joined #darcs
20:29 amgarchIn9 joined #darcs
20:41 mizu_no_oto joined #darcs
21:38 gh_ mkay, patch index is most probably to be disabled by default in 2.10
21:38 gh_ unless someone takes the time to think it through so that it does not consume that much memory on creation
21:56 gh_ joined #darcs
22:42 mizu_no_oto joined #darcs
23:49 gh_ I managed to reduce a little memory consumption by no longer using the String representation of patch hash, but directly its binary representation.
23:50 gh_ also this cuts in half the size of _darcs/patch_index/repo_state

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