Camelia, the Perl 6 bug

IRC log for #darcs, 2013-04-16

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

All times shown according to UTC.

Time Nick Message
01:14 gh_ joined #darcs
01:20 Igloo joined #darcs
01:46 preflex joined #darcs
01:58 dolio joined #darcs
02:19 gh_ joined #darcs
02:43 preflex_ joined #darcs
03:14 mizu_no_oto joined #darcs
03:45 edwardk joined #darcs
06:44 lelit joined #darcs
08:03 raichoo joined #darcs
08:05 amgarchIn9 joined #darcs
08:37 mornfall Whom can I contact to ask for _ilbot to join another channel? :)
10:02 hualet joined #darcs
10:07 hualet left #darcs
11:29 favonia joined #darcs
12:41 iago joined #darcs
12:43 mizu_no_oto joined #darcs
12:59 mizu_no_oto joined #darcs
13:51 gal_bolle joined #darcs
14:34 dolio joined #darcs
15:10 goldfishtechno joined #darcs
15:19 donri joined #darcs
15:19 mizu_no_oto joined #darcs
16:51 raichoo joined #darcs
17:22 carter joined #darcs
17:26 amgarchIn9 joined #darcs
17:31 mizu_no_oto joined #darcs
17:45 edwardk joined #darcs
18:16 wferi joined #darcs
18:17 wferi Hi! Why does darcs changes show 'Å‘' characters as 'Q'? (version 2.8.1)
18:17 wferi The --xml output contains 'Å‘' characters all right
18:21 kerneis wferi: there is a bunch of environment variables to control which unicode characters are allowed to be displayed
18:21 kerneis (for some security reason I never really understood --- I'm not aware of any software doing this apart from darcs)
18:22 wferi DARCS_DONT_ESCAPE_8BIT=1 is exported from my environment
18:22 wferi but if I understand you right, kerneis, this isn't necessarily enough
18:23 kerneis I remember a variable listing explicitly alowed characters
18:23 kerneis but I don't know if it's still used
18:24 kerneis wferi: http://darcs.net/manual/darcs.pdf section 3.6
18:25 kerneis hmm, it looks like DONT_ESCAPE_8BIT should be enough
18:25 kerneis "As a last resort you can set DARCS DONT ESCAPE ANYTHING to 1"
18:26 wferi I just did, it did not help...
18:26 kerneis no idea, sorry
18:31 wferi hmm, I'm getting something... darcs changes does not create UTF-8 output
18:31 wferi although I'm using an UTF-8 locale
18:31 wferi darcs changes --xml outputs UTF-8, which is correctly displayed by my terminal
18:32 wferi but darcs changes uses some single-byte encoding, which is not!
18:33 wferi actually, my Å‘-s are internally converted to Q-s, probably because they can't be converted to that mysterious single octet encoding
18:34 wferi I really wonder why darcs munges my commit messages at all :(
18:35 sm trying to be compatible across platforms perhaps ?
18:35 sm just a guess, I don't know either
18:39 _ilbot joined #darcs
18:39 Topic for #darcs is now http://darcs.net/ | logs: http://irclog.perlgeek.de/darcs/ | darcs 2.8.4 is out http://darcs.net/Releases/2.8
18:52 Heffalump it turns your commit messages into UTF8 internally so that they render well in other locales
18:53 Heffalump I suspect the behaviour of darcs changes is a bug, but I'm not certain
19:02 wferi Heffalump, I use an UTF-8 locale, so there's not much to convert... Of course it's nice if the messages render well in other locales, but I'd really like if they rendered well in my (original, recording) locale.
19:03 Heffalump so you wrote a commit message and darcs internally corrupted it? Or just that darcs changes messed up when outputting it/
19:04 wferi I think the latter
19:04 benmachine well, if --xml is getting it right, sounds like it's not corrupted
19:04 wferi the xml seems right
19:04 Heffalump ok. The code is a bit of a mess with the way it handles packing/unpacking, so I think it's likely to just be a bug
19:05 wferi try it, I don't use anything special: LANG=en_US.UTF-8
19:05 wferi create a commit message with an Å‘ character
19:05 wferi display it with darcs changes, and find a Q in its place
19:06 wferi but all the other accented characters are converted to one octet
19:06 wferi they aren't written back in UTF-8
19:06 wferi but my urxvt somehow manages to display most of them right...
19:07 benmachine what are they, ooi?
19:07 wferi sorry, ooi?
19:07 benmachine (out of interest)
19:08 wferi the accented characters? in original form or as single octets? anyway: éáí and similar
19:08 wferi Hungarian accented letters
19:10 wferi bbl
19:31 benmachine wferi: éáí are all in ISO-8859-1 which overlaps with unicode
19:31 benmachine so I suggest what is happening here is probably a truncation to eight bits
19:31 mizu_no_oto joined #darcs
19:44 wferi benmachine: can well be.
19:44 wferi anyway, it's surely a bug, I'll submit one if it's not known
20:53 amgarchIn9 joined #darcs
20:54 mizu_no_oto joined #darcs
21:01 mizu_no_oto joined #darcs
21:14 delamonpansie joined #darcs
21:53 favonia joined #darcs
23:14 mizu_no_oto joined #darcs
23:18 dolio joined #darcs
23:53 alexsuraci joined #darcs

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