Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-03-29

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

All times shown according to UTC.

Time Nick Message
00:00 teatime i.e., a single codepoint which would otherwise be part of an EGC / NFG char ?
00:00 idiosyncrat_ We don't do NFG.  I don't know if that answered your question.
00:00 teatime it does, that will work fine.
00:01 teatime ... I think.
00:01 teatime definitely easier to deal w/ than the other extreme :)
00:01 idiosyncrat_ Also, there's a problem with UTF in Marpa *scripts* -- no in the input to Marpa ...
00:01 idiosyncrat_ s/no in the input/not in the input/
00:01 idiosyncrat_ but in the Marpa grammar description itself, there are bugs with UTF8 handling.
00:02 idiosyncrat_ These I don't think affect what you are talking about.
00:02 teatime ah, unfortunate... not a complete deal-breaker, but unfortunate...
00:02 teatime what kind of bugs?
00:03 idiosyncrat_ Basically your Marpa SLIF grammar description should be in ASCII-7.
00:03 teatime alas.
00:03 idiosyncrat_ Anything beyond that is a roll of the dice.
00:03 teatime can I at least have unicode in comments?
00:03 idiosyncrat_ The input to your grammar can be UTF8 and that's been thoroughly tested and is quite robust.
00:04 teatime the thing I want parse involves lots of unicode chars, so I can handle writing them as \Uxxxx or whatever if I have to
00:04 idiosyncrat_ In comments -- you can't rely on it not breaking the script.
00:04 teatime heh, ok
00:05 idiosyncrat_ This is something we are thinking of fixing, but the question is does it justify a new Marpa::R2 release?
00:05 idiosyncrat_ Jean-Damien has identified the problem and the fix itself looks straight-forward.
00:05 teatime I lack context to make an attempt at answering that.
00:06 teatime obviously, I would vote yes, fix it and roll a new release :)
00:06 teatime v.R2-teatime0
00:07 idiosyncrat_ Your own fork! :-)
00:08 teatime ok, I'll take a look at more of the docs and intro materials, and if I have more questions later I will definitely bug you again.  TYVM for the info.
00:08 idiosyncrat_ You're welcome and I hope this helped.
00:08 teatime quite.
00:10 idiosyncrat_ teatime: for context in reading the docs -- the  UTF8 problem I described is in SLIF scripts ...
00:10 idiosyncrat_ everything else works.
00:10 teatime ok.
00:11 idiosyncrat_ dxvd: re http://irclog.perlgeek.de/marpa/2016-03-28#i_12250023
00:11 idiosyncrat_ Peter Stuifzand did some work on LISP parsers in Marpa:
00:12 idiosyncrat_ https://github.com/pstuifzand/Marpa-parser-generator/blob/master/examples/lisp/lisp.mp
00:12 idiosyncrat_ https://gist.github.com/pstuifzand/5930450
00:13 idiosyncrat_ I'm not sure if they even work, but they look like Peter had the right idea -- he was very, very good with Marpa, to put it mildly.
00:14 idiosyncrat_ LISP should be easy for Marpa to do -- it has hardly any syntax at all.
00:16 idiosyncrat_ dxvd: there is also this by amon: http://stackoverflow.com/questions/16431079/lisp-like-grammar-compile-into-javascript-with-perl
00:17 idiosyncrat_ again, I've not tested, but amon is a Marpa expert, and his answer was the one selected as best by the OP
01:12 dvxd idiosyncrat_ thank you, will click the links once I get home.
03:42 ceridwen joined #marpa
03:42 ceridwen joined #marpa
04:15 idiosyncrat_ Good night!
06:42 pczarn joined #marpa
07:01 maybekoo2 joined #marpa
09:15 sadmac joined #marpa
11:37 kaare_ joined #marpa
13:03 maybekoo2 joined #marpa
17:07 ceridwen_ joined #marpa
17:07 ceridwen_ joined #marpa
17:10 sadmac joined #marpa
17:18 iarna joined #marpa
17:56 koo7 joined #marpa
18:07 idiosyncrat_ joined #marpa
18:29 jdurand joined #marpa
18:30 jdurand Re http://irclog.perlgeek.de/marpa/2016-03-29#i_12250513 - if it not working (Peter?) gimme a yacc or whatever grammar and I do it -;
18:30 jdurand Lunch time - AFK
20:01 jdurand joined #marpa
20:31 idiosyncrat_ About to go out on errands, but a thought --
20:32 idiosyncrat_ I'm thinking of starting Marpa::R3, as basically a "cleaned-up" Marpa::R2, not a radical change ...
20:32 idiosyncrat_ Kollos is still the plan, and Marpa::R3 would be a "baby step" toward it.
20:33 idiosyncrat_ Among the clean-ups:
20:33 idiosyncrat_ No interfaces except the SLIF.
20:33 idiosyncrat_ The UTF8 fix in SLIF scripts by default.
20:34 idiosyncrat_ Also, by default --
20:35 idiosyncrat_ :default ::= action => [name,values];   lexeme default = latm => 1
20:36 idiosyncrat_ *Mostly* backward compatible, so not a radical step forward, but ...
20:36 idiosyncrat_ where it is backward compatibility versus eliminating annoying behavior, we eliminate the annoyance.
20:37 idiosyncrat_ I'd lead on this effort, because I think that's what folks what, but I'd hope for major help on specific sub-tasks.
20:38 idiosyncrat_ If we go with Marpa::R3, then immediately Marpa::R2 becomes "major bug fix only".
20:39 idiosyncrat_ jdurand: also, various of your excellent ideas which we haven't been able to do because of backward compatibility, these can now go in.
20:41 idiosyncrat_ I think one benefit of cleaning out the non-SLIF interfaces, may be lower overhead at start-up time.
20:41 idiosyncrat_ I look forward to your comments.
20:41 idiosyncrat_ AFK
23:13 idiosyncrat_ joined #marpa

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