Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-03-28

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

All times shown according to UTC.

Time Nick Message
04:17 sirdancealot joined #marpa
04:48 jeffreykegler joined #marpa
04:48 jeffreykegler file:///home/jeffreykegler/Downloads/how-to-read-code.pdf
04:49 jeffreykegler From a talk Aria gave on how to read code -- it includes a bit of Lotsawa, her Libmarpa clone.
04:49 jeffreykegler Nothing about how to read CWeb, I notice. :-)
06:08 rns joined #marpa
06:11 rns re http://irclog.perlgeek.de/marpa/2015-03-28#i_10353138 -- good read -- a lot of diverse information in very terse text.
06:15 rns jeffreykegler: re CWeb -- That's probably because CWeb is so easy to read. :) Seriously, I *started* reading marpa.w without reading anything on CWeb grammar and was able to find my way around by looking at named chunks and how they are used in the code.
06:16 rns And all of a sudden that became self-explanatory enough to make small changes.
06:19 rns But I have admit that Cweb can be daunting at a first look. Oh, and the link to the source of the talk, just in case -- https://github.com/aredridel/how-to-read-code
06:21 jeffreykegler rns: I notice a lot of work on your Regex module
06:24 rns Yes, the compilation from BNFish to regex looks feature-complete -- it supports symbol expansion and recursion.
06:24 jeffreykegler Are you planning to put it on CPAN?
06:28 rns Eventually, yes.
06:28 rns I'm hesitating -- I succesfully converted most famous regexes, but haven't yet really tried to build a regex in BNFish syntax from scratch, except perhaps https://github.com/rns/MarpaX-Regex/blob/master/t/msvc_warnings.t which, along with your idea, really started it all.
06:29 rns And I need to write some doc/tutorial too.
06:30 jeffreykegler It'll run simple regexes via Marpa?
06:32 rns Nope, it compiles everything to Perl 5.010 regexes now. Marpa is used as a parser.
06:32 rns But now that it's feature complete, I'm planning to start compiling to SLIF to run regex via via Marpa.
06:32 jeffreykegler Ah, OK
06:33 jeffreykegler So at some point the user will have their choice?  SLIF or Perl 5.10 regex as target?
06:37 rns Yes, I plan to support target => (SLIF|regex) option with regex being the default and SLIF and marpa left for hard cases.
06:37 rns One hard case is Randal Swartz's json regex -- which I can't really compile to regex, but which can be compiled to SLIF, I think.
06:38 rns Except for that json regex, other notable regexes (s-expression, html) compile ok.
06:39 jeffreykegler Nice.  Time for me to sign off.
06:39 rns Ok, good bye.
07:08 rns left #marpa
10:45 lwa joined #marpa
17:51 sirdancealot joined #marpa
18:02 koo5 joined #marpa
20:11 jeffreykegler joined #marpa
20:28 jdurand joined #marpa
20:30 jdurand rns: about your regex stuff, do you plan to support posix collation and equivalence class? The difficulty is that perl has reserved this syntax and croaks systematically if it is seeing it
20:33 jdurand ... at least one the syntax http://en.wikipedia.org/wiki/Special:Search?go=Go&amp;search=.., maybe not with http://en.wikipedia.org/wiki/Special:Search?go=Go&amp;search=<>
20:33 jdurand oups! My client garbled what I typed!!!
20:33 jdurand let"s put it in pastebin then
20:34 shadowpaste "jdurand" at 217.168.150.38 pasted "[[.test.]]" (3 lines) at http://fpaste.scsys.co.uk/470488
20:53 rns joined #marpa
20:56 rns jdurand: re http://irclog.perlgeek.de/marpa/2015-03-28#i_10355406 -- for now, whatever the application puts in between the square brackets, is passed through to the regex -- https://github.com/rns/MarpaX-Regex/blob/master/lib/MarpaX/Regex.pm#L51
21:00 rns left #marpa
21:28 Aria Hehe. My talk was aimed particularly at novice programmers, so I didn't dive deep into algorithmic code and the complexities of reading and writing and describing it.
21:31 ceridwen joined #marpa
21:32 flaviu joined #marpa
22:47 ronsavage joined #marpa
23:31 ronsavage jk: I notice in metag.bnf, this: <current lexer statement>. Is it documented?
23:54 jeffreykegler ronsavage: No, and never will be -- my work on multiple lexers was a dead end and I backed it all out of the code, for efficiency's sake.
23:55 jeffreykegler And to make the code simpler.
23:56 jeffreykegler Switching lexers *sounds* like a oool idea, but it means there's a boundary between the two on switch-over
23:57 jeffreykegler In programming it's hard to get the switch-over right.
23:58 jeffreykegler Whereas, if you just change priority of, and activate and deactivate lexemes within the same lexer, you get all the same effects,
23:58 jeffreykegler but things are a lot simpler conceptually.

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