Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-02-20

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

All times shown according to UTC.

Time Nick Message
01:03 jeffreykegler joined #marpa
01:35 jeffreykegler joined #marpa
01:41 ronsavage joined #marpa
02:20 jeffreykegler joined #marpa
02:48 ilbot3 joined #marpa
02:48 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Pastebin: http://scsys.co.uk:8002/marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today
03:44 jdurand joined #marpa
03:44 jdurand Re https://github.com/jeffreykegler/Marpa--R2/issues/246 - oups -; thx.
04:11 jeffreykegler jdurand: https://github.com/jeffreykegler/Marpa--R2/blob/master/cpan/pod/Event.pod#pre-lexeme-events
04:12 jeffreykegler I added some explanation to the description of "pre-lexeme events" based on your issue #235
04:13 jeffreykegler Helping to clarify the difference between predictions and pre-lexeme events, and why the behavior of pre-lexeme events may seem strange.
04:21 ronsavage joined #marpa
04:24 ronsavage left #marpa
04:28 jdurand Yep, many thanks. I admit pre-lexeme events is as powerful as sometimes puzzling.
05:24 ronsavage joined #marpa
05:25 ronsavage Testing after upgrading Perl to 5.20.2.
05:33 rns joined #marpa
05:36 rns jeffreykegler: re http://irclog.perlgeek.de/marpa/2015-02-19#i_10144626 -- built, installed, and runs my modules ok under winxp (msvc 2010, perl 5.20.1) and cygwin (gcc 4.9.2, perl 5.14.2)
06:07 rns left #marpa
06:47 jeffreykegler rns: Thanks!
10:25 pczarn joined #marpa
10:34 lwa joined #marpa
13:55 pczarn The fundamental difference is that G1 looks for full parses and L0 looks for longest partial parses
14:08 pczarn I hope there's a way to simplify this
14:11 pczarn If you think of parsing as a pipeline, it should be possible to unify SLIF (character_classes -> marpa_grammar -> marpa_grammar)
19:12 tjvr joined #marpa
19:59 tb10 joined #marpa
19:59 tb10 left #marpa
21:22 jeffreykegler joined #marpa
21:23 jeffreykegler pczarn: "unified" approaches, that is one grammar which does both lexing and scanning are certainly possible with the SLIF now.
21:24 jeffreykegler You can do character-by-character external scanning, though until Kollos comes along this will be slow.
21:24 jeffreykegler I experimented with this *a lot*, and still think there may be a future in it.
21:25 jeffreykegler Two things to note: until Kollos comes along, external scanning of every character will be slow ...
21:25 jeffreykegler and until strand parsing comes along, having an Earley set for every character will be space-intensive.
21:26 jeffreykegler Also, a 3rd thing to note, much of programmers have trouble with the lexer/parser distinction, they seem to have more trouble in its absence --
21:26 jeffreykegler there seems to be a natural division between the two in human language processing.
21:40 flaviu joined #marpa
21:47 jeffreykegler rns: with the Libmarpa error codes
21:48 jeffreykegler The intent is that they'll be taken from marpa_codes.h or marpa_codes.c or that a similar table will be created by processing error_codes.table
21:48 jeffreykegler These are not compiled in, because I want to keep strings out of Libmarpa
21:49 jeffreykegler But it's safe to separately copy them and use them, even if they get out of sync, because ...
21:49 jeffreykegler Existing error numbers and mnemonics are guaranteed not to change -- more might be added, but none will be changed, even if they become inaccurate or misleading.
22:10 ronsavage joined #marpa
22:17 jeffreykegler My "Parsing: a timeline" has been translated into Russian:  http://tproger.ru/translations/parsing-a-timeline/
22:18 jeffreykegler They added a nice illustration.

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