Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-03-20

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

All times shown according to UTC.

Time Nick Message
03:48 lucs joined #marpa
04:09 jdurand IMHO current Marpa::R2 logic is ok, and yes this is a side effect of perl DWIMery - you might want to document in the POD that you are affected by this and do nothing -;
04:23 idiosyncrat_ I am going to look at and think about this some more.
04:23 idiosyncrat_ My current thinking is that the Perl bug laid the trap and I stepped into it ...
04:24 idiosyncrat_ but any fix might break old scripts.
04:27 idiosyncrat_ One approach is to add a new global statement to the SLIF, like "use utf8", to enable the incompatible, but more correct behavior.
04:35 kaare_ joined #marpa
04:40 idiosyncrat_ Good night!
04:49 jdurand yes - though to the same thing - or alternatively a new option to SLIF recce argument
04:49 jdurand Goof night
04:49 jdurand "Good night" -;
07:11 koo7 joined #marpa
07:54 choroba joined #marpa
15:46 idiosyncrat_ joined #marpa
16:29 daxim joined #marpa
16:30 daxim I plugged a computer language grammar into grammophon and it told me the grammar is neither any 5 of LL(1), LR(0), SLR(1), LR(1), LALR(1)
16:30 daxim what is it, then?  and what implications does this have?
16:42 idiosyncrat_ daxim: it's context-free, and parseable by Earley and therefore Marpa.
16:42 idiosyncrat_ Though I'd have to know more to tell if Marpa could parse it in linear time.
16:57 idiosyncrat_ daxim: I say this because everything you can express in BNF is a context-free grammar (CFG).
16:57 idiosyncrat_ So if you could type it into grammophone, it's a CFG.
16:59 daxim what are the criteria for marpa to parse something in linear time?  can I ask marpa its opinion about the class of grammar?
17:04 idiosyncrat_ https://jeffreykegler.github.io/Marpa-web-site/ has the details, under "Qucbbles".
17:05 idiosyncrat_ We first want to know if the grammar is unambiguous.  Unfortunately, in the general case, that is undecidable, but you may be able to determine that for your particular grammar.
17:05 idiosyncrat_ (Marpa also parses many ambiguous grammars in linear time, but that's another topic.)
17:07 daxim it's ambiguous, you demonstrated that to me over a year ago
17:09 daxim the explanation on the website about middle recursions need some examples illustrating a marked or unmarked recursion
17:11 KotH joined #marpa
17:12 daxim same goes for the right recursion, show an example how to rewrite while preserving the semantics
17:19 kaare__ joined #marpa
18:15 pczarn joined #marpa
18:42 pczarn if the site called them "directly ambiguous right recursions", it'd be more specific
18:43 pczarn daxim: Unfortunately, the site doesn't link to this blog post: http://jeffreykegler.github.io/Ocean-of-Aware​ness-blog/individual/2015/08/fast_handy.html
18:49 pczarn Further, you could rename "unmarked middle recursion" to "locally ambiguous middle recursion".
19:19 idiosyncrat_ pczarn: re http://irclog.perlgeek.de/m​arpa/2016-03-20#i_12212870
19:19 idiosyncrat_ Thanks a log for that link!
19:20 idiosyncrat_ I thought I'd given specifics somewhere, but couldn't find where, so thanks for finding that for me.
19:21 idiosyncrat_ Re the terminology, in the sense I understand the word "direct", the problem is *not* limited to "directly ambiguous right recursions", but includes all ambiguous right recursiions, ...
19:21 idiosyncrat_ whether direct or indirect.
19:22 idiosyncrat_ By the way, it's an open question whether an unambiguous grammar *can* have an ambiguous right recursion.
19:23 idiosyncrat_ Neither Joop Leo or I know the answer -- we don't know of a proof that it's impossible, but we also don't know of an example.
19:23 idiosyncrat_ Similarly with "locally ambiguous middle recursion" -- "local" to me suggests that the ambiguity is local to a finite span of tokens in the input, and ...
19:24 pczarn I see
19:25 idiosyncrat_ with an unmarked middle recursion the ambiguity is not necessarily "local" in that sense.
19:27 idiosyncrat_ re http://irclog.perlgeek.de/m​arpa/2016-03-20#i_12212938
19:27 idiosyncrat_ s/Thanks a log/Thanks a lot/
19:28 idiosyncrat_ But anyway thanks a whole bunch -- I was sure I'd written this stuff up, but couldn't find it.
19:30 pczarn I found it with google
20:04 shadowpaste0 joined #marpa
20:05 idiosyncrat_ I tried Google too, but obviously didn't use the right keywords or wasn't as persistent.
20:05 idiosyncrat_ AFK
20:22 shadowpaste joined #marpa
20:23 shadowpaste joined #marpa
20:25 shadowpaste joined #marpa
20:27 shadowpaste joined #marpa
20:30 shadowpaste0 joined #marpa
21:41 ronsavage joined #marpa
23:06 idiosyncrat_ joined #marpa

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