Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-11-18

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

All times shown according to UTC.

Time Nick Message
01:29 jeffreykegler joined #marpa
02:40 jeffreykegler I just uploaded Marpa::R2 2.100000 to CPAN.
02:40 jeffreykegler It contains mainly doc updates & fixes.
02:40 jeffreykegler Testing appreciated!
02:45 jeffreykegler Aria: re http://irclog.perlgeek.de/marpa/2014-11-17#i_9676475 -- Thanks!
03:02 ronsavage joined #marpa
03:15 ronsavage Installing and testing V 2.100000 now....
03:18 ronsavage1 joined #marpa
03:47 ronsavage1 Everything passes.
04:50 ronsavage1 Help required for complex case: https://gist.github.com/ronsavage/7b611e27cefd562d2f11. TIA.
06:08 jeffreykegler joined #marpa
06:09 jeffreykegler ronsavage: The error message "Exception: value() called when recognizer is not in tree mode" occurs at two points.  Are these the issues?
06:10 jeffreykegler Also, is this a regression -- that is, did it work in an earlier version, but now fails?
06:22 jeffreykegler ronsavage: I posted a patch as a comment.
06:38 ronsavage joined #marpa
06:40 ronsavage jeffreykegler: I've added new test cases with multiple subgraphs (named, unnamed) and the failure is immediately after the first if anything (not just subgraphs) in the input follows the subgraph.
06:41 ronsavage The multiple errors are 1 per test case.
06:41 ronsavage The error is always: Parse is ambiguous. Status: Length of symbol "statement_token" at line 8, column 2 is ambiguous
06:43 ronsavage So the problem is where the grammar says at line 559 that a subgraph ends with a graph_statement, but Marpa won't (to my way of thinking) go back to lines 514 .. 520 and accept more tokens.
06:45 ronsavage I just happen to use subgraphs in the gist after the first. I would have been clearer to add another node name there instead in various cases. The error /is/ the same.
06:45 ronsavage 'I' => 'It'.
06:48 ronsavage Thanx for the comment, but I won't use it because I definitely don't want such errors at all. I need to understand what it is about the grammar that let's such cases arise. Meaning: The only solution acceptable to me is to edit the grammar.
06:51 ronsavage Hmm. The same problem arises if I put a node before the subgraph in the 2nd case. It's understandable in that the node name could be a subgraph name. That may be a different problem.
06:52 ronsavage I did think of using priorities, but could not see how to do that.
06:52 ronsavage HTH.
07:00 ronsavage On another note: Anyone here familiar with AutoDia http://aarontrevena.co.uk/opensource/autodia/index.html? I ask because I've received an email from someone interested in interfacing Graphviz (yeah!) with AutoDia, and my new GraphViz2::Marpa may be the key. That is - on top of the range of features AutoDia has already. He wants to output XML to an XML-oriented editor (unnamed).
09:01 lwa joined #marpa
09:03 rns joined #marpa
09:04 rns ronsavage: With subgraph staff rewritten this way:
09:04 rns subgraph_statement::= subgraph_prefix subgraph_id_token graph_statement
09:04 rns subgraph_statement      ::= subgraph_id_token graph_statement
09:04 rns subgraph_statement      ::= subgraph_prefix graph_statement
09:04 rns subgraph_prefix::= subgraph_literal
09:05 rns subgraph_id_token::= node_name
09:06 rns (avoiding the explicit null rules), the 'Parse is ambiguous' errors go away.
09:11 rns The end result becomes 'Counts: in => 6. success => 2.' but the above rewriting looks useful still.
09:18 ronsavage rns: Brilliant. I did look at getting rid of the explicit null rules, since I felt uneasy about them, but I did not think of your re-write. And yes, cases 2 & 3 pass, and that's a bit improvement. It's 8:20 pm here now, so I look forward to tomorrow thanx to you!
09:26 rns ronsavage: Great, glad it worked.
09:29 ronsavage I should add I'm happy for 2 reasons: (1) You've made progress for me, (2) I know how to disambiguate (probably) all failures.
09:44 ilbot3 joined #marpa
09:44 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
10:16 daxim joined #marpa
11:39 flaviu joined #marpa
15:08 Spads joined #marpa
15:35 koo5 joined #marpa
15:36 jluis joined #marpa
15:36 Spads joined #marpa
15:36 Spads joined #marpa
16:03 Spads_ joined #marpa
16:08 Spads joined #marpa
16:18 Spads_ joined #marpa
16:25 Spads_ joined #marpa
16:35 Spads joined #marpa
16:36 Spads joined #marpa
17:18 Spads_ joined #marpa
17:33 jeffreykegler joined #marpa
18:45 Spads_ joined #marpa
18:58 jeffreykegler I've updated my annotated blog index -- http://jeffreykegler.github.io/Ocean-of-Awareness-blog/metapages/annotated.html
19:00 jeffreykegler Several of my recent posts have been about parsing in general, and there is now a separate section for those.
19:00 Spads joined #marpa
20:11 dmitri joined #marpa
20:35 Spads_ joined #marpa
20:42 flaviu joined #marpa
21:19 Spads_ joined #marpa
21:46 Spads_ joined #marpa

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