Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-10-19

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

All times shown according to UTC.

Time Nick Message
00:29 ronsavage koo7: Q (1) What's KNF? I visited http://www.allacronyms.com/KNF. Is 'kernel...' the one relevant? I guess so. And are you giving us a Perl wrapper for Lua's regexp engine :-)?
01:48 ilbot3 joined #marpa
01:48 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
03:22 Idiosyncrat I recall djns asking questions about context-sensitive parsing.
03:23 Idiosyncrat Aho&Ullman 1972 has some good pages on that topic.
03:23 Idiosyncrat In the course of doing my paper, I'm looking back at the old texts.
03:25 Idiosyncrat I discover from A&U that derivation trees do not work for context-sensitive grammars, ...
03:26 Idiosyncrat which is pretty obvious, but I've done so little with CS grammars that it hadn't occurred to me.
04:23 koo7 ronsavage: i just googled out of curiosity and pasted an excerpt from the first hit or so
04:39 djns I was confused about the BNF notations for these kinds of parsers/lexers. I had difficulty finding explanations that were unambiguous
04:40 djns that is, clearly and adequately described and explained
05:05 Idiosyncrat Good night!
05:13 shadowpaste joined #marpa
06:03 ronsavage joined #marpa
06:20 ronsavage joined #marpa
06:21 ronsavage jk: Did you see http://irclog.perlgeek.de/marpa/2015-10-18#i_11394295?
09:14 Cheery joined #marpa
15:09 Idiosyncrat joined #marpa
15:10 Idiosyncrat ronsavage: No, I missed http://irclog.perlgeek.de/marpa/2015-10-18#i_11394295 in my backlogging.
15:10 Idiosyncrat See https://metacpan.org/pod/Marpa::R2::Scanless::R#ambiguous
15:10 Idiosyncrat particularly the last graf in the writeup of ambiguous()
15:11 Idiosyncrat This describes the usual cause [ calling ambiguous() ] ..
15:11 Idiosyncrat and the solution -- calling series_restart().
15:13 Idiosyncrat Essentially, I created a completely different evaluation structure for traversing forests ...
15:13 Idiosyncrat (something by the way I think more folks should be interested in doing) ...
15:14 Idiosyncrat and it's not compatible with the ordinary evaluation structure in "tree mode" ...
15:14 Idiosyncrat which is usually no issue because most folks care only about one or the other, except ...
15:15 Idiosyncrat that I found that forest mode can be used to add a really cool feature to Marpa::R2 which is ...
15:15 Idiosyncrat not only detect ambiguity, but describe in a sensible way where it is ...
15:15 Idiosyncrat (find some other parser that does that!)
15:16 Idiosyncrat But if you call the actual ambiguity-describing logic, it sticks you in "forest mode" ...
15:16 Idiosyncrat and since the call is something people who don't know what "forest mode" is might want to use ...
15:16 Idiosyncrat it does cause confusion ...
15:17 Idiosyncrat and would be a good candidate for the FAQ!
15:25 Idiosyncrat \me switches topics
15:25 Idiosyncrat For those looking at Lua syntax: http://permalink.gmane.org/gmane.comp.lang.lua.general/119458
16:30 lucs joined #marpa
21:35 ronsavage joined #marpa
21:36 ronsavage JK: Thanx for the link. I'm already calling both $self -> recce -> ambiguous and $self -> recce -> terminals_expected, so things are becoming clear. And yes, the FAQ will get a boost from this.
21:49 ronsavage joined #marpa
21:49 ronsavage And I'd forgotten about series_restart().
22:33 Idiosyncrat joined #marpa
22:36 Idiosyncrat ronsavage: my idea was that in general folks would call ambiguity_metric() first
22:36 Idiosyncrat https://metacpan.org/pod/Marpa::R2::Scanless::R#ambiguity_metric
22:37 Idiosyncrat This call does *not* put the recognizer into forest mode -- it just tells you if the parse is ambiguous or not.
22:38 Idiosyncrat So -- as I thought -- you'd call ambiguity_metric() and if you thught of ambiguity as an error ---
22:39 Idiosyncrat which I believe is how you typically see it, Ron ---
22:39 Idiosyncrat then you'd call ambiguous() to find out exactly where and what the ambiguity is.
22:40 Idiosyncrat This would leave the parse in forest mode, which is not right for what you originally intended, ...
22:41 Idiosyncrat but you're about to abend/destroy that particular recognizer, in which case which mode you're in makes no difference.
22:42 Idiosyncrat On the other hand, if you plan on going into Marpa's ASF parsing --- https://metacpan.org/pod/distribution/Marpa-R2/pod/ASF.pod
22:42 Idiosyncrat then forest mode is *exactly* what you wanted.

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