Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-08-28

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

All times shown according to UTC.

Time Nick Message
00:29 ronsavage1 I've added some tutorials to Marpa's homepage.
00:32 ronsavage joined #marpa
01:57 ronsavage1 joined #marpa
03:20 ronsavage joined #marpa
03:49 jeffreykegler joined #marpa
04:00 jeffreykegler I'm thinking of blogging re the new web page -- perhaps around Monday.
04:00 jeffreykegler A new public web page, it's appropriate that it be announced.
04:04 ronsavage1 joined #marpa
04:11 ronsavage joined #marpa
04:12 ronsavage Sounds good. Re the new color scheme. I switch every few years, and had decided it was time to do it again before I agreed to curate the new Marpa page, so just coincidence.
04:12 ronsavage AFK
05:12 jeffreykegler I've been working up a new JSON parser: https://github.com/jeffreykegler/perl-JSON-Decode-Marpa/blob/master/lib/JSON/Decode/Marpa.pm
05:12 jeffreykegler (It started with me fooling around with a benchmark.)
05:13 ronsavage joined #marpa
05:13 jeffreykegler It now takes advantage of LATM to do some tricks that were not possible before.
05:14 jeffreykegler In particular, you can now use the G1 level Marpa parser to deal with JSON strings, indentifying escaped and hex characters.
05:14 ronsavage BTW: I'm thinking of a Troubleshooting Guide for Marpa. Email contributions to me. I have some examples planned, explaining why certain grammatical constructs generate the errors they do.
05:14 jeffreykegler The grammar is quite clear.
05:14 jeffreykegler ronsavage: Sounds great.
05:15 ronsavage Of course, any resemblance between defective grammars and my grammars is purely coincidental!
05:15 jeffreykegler Of course, that's why you need contributions. :-)
05:16 ronsavage Ahh, yes. The old self-referential issue again.
05:16 jeffreykegler Re JSON and LATM, the new grammar is a bit faster and much, much more elegant.
05:17 jeffreykegler For evaluation, I create an AST and post-process it.
05:18 ronsavage Sounds like a re-write of MarpaX::Demo::JSONParser will be necessary.
05:19 ronsavage Ahhh. That's Marpa :: Demo :: JSONParser.
05:20 jeffreykegler ronsavage: I tested the new JSON grammar against basic.tests.t from your MarpaX::Demo::JSONParser
05:20 jeffreykegler and everything tests 100% OK.
05:20 jeffreykegler ronsavage: I tested the new JSON grammar against basic.tests.t from your MarpaX::Demo::JSONParser
05:20 jeffreykegler and everything tests 100% OK.
05:20 * jeffreykegler is having Internet troubles again
05:21 ronsavage Thanx. Good testing news is always good news.
05:21 jeffreykegler I tried adding the new grammar to MarpaX::Demo::JSONParser, but ...
05:21 jeffreykegler it uses post processing for evaluation, and it was hard to factor that into your test code.
05:22 jeffreykegler MX::D::JP seems to assume you can just grab the BNF and use the same logic.
05:23 jeffreykegler Also, I started with code using Steven Haryanto's MarpaX::Simple, so it uses those calls, although it's not inextricably tied to MarpaX::Simple
05:25 ronsavage I think they can be unified. I'll look at it, but not immediately.
05:25 jeffreykegler Anyway, what with it getting late California time, and my Internet getting finicky, I had better go AFK
05:26 jeffreykegler ronsavage: Thanks!
05:38 ronsavage joined #marpa
06:53 ronsavage1 joined #marpa
16:27 jeffreykegler joined #marpa
16:28 shadowpaste "jeffreykegler" at 162.232.214.245 pasted "JSON benchmark script" (32 lines) at http://scsys.co.uk:8002/419509
16:31 lucs jeffreykegler: Do you happen to have some results to show?
16:35 jeffreykegler Jean-Damien beat me, so I'm hiding the numbers. :-)
16:36 lucs :-D
16:37 jeffreykegler Actually, I've just emailed them to the G+ group.  But JDD *did* come in faster.
19:55 jdurand_ joined #marpa
21:25 jeffreykegler joined #marpa
21:48 ronsavage joined #marpa
23:42 ronsavage1 joined #marpa
23:43 ronsavage1 I've patched Jeffrey's new JSON code, but it fails 1 test. See http://savage.net.au/Ron/MarpaX-Demo-JSONParser.tgz. Unpack and try prove -Ilib -v t. I've commented out the 2 utf8 tests near the end of t/basic.tests.t.
23:43 ronsavage1 Also you can edit line 240 of t/basic.tests.t to see the code still runs all the old tests successfully.

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