Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-07-04

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

All times shown according to UTC.

Time Nick Message
00:02 idiosyncrat $valuer = Marpa::R3::Valuer::new($recce); VALUE: while (my $value = $valuer->value()) { print Data::Dumper::Dumper($value); }
00:03 * idiosyncrat notices some of the code got translated into emoticons on Pidgin -- if so please read ::D as : : D
00:05 idiosyncrat As a convenience, I'll keep the $recce->value() method -- it will transparently create a valuer, check for ambiguity and throw a fatal error if
00:05 idiosyncrat 1.) create a valuer
00:05 idiosyncrat 2.) check for ambiguity, throwing a fatal error if it occurs;
00:05 idiosyncrat 3.) return $valuer->value()
00:06 idiosyncrat 4.) The valuer will be local to the $recce->value() method, and will be garbage collected when $recce->value() returns.
00:08 idiosyncrat A side effect of this will be that it will now become possible to create an ASF's and retrieve normal values at the same time for the same recognizer -- there'll be no more "tree mode".
00:10 idiosyncrat ronsavage: This means some of the logic you had to put it to deal with printing a diagnostic which describes an ambiguity (which implicity puts the recce into "ASF mode"), but also continuing the parse, will no longer be necessary.\
00:11 idiosyncrat With the creation of the valuer, I think & hope all the biggest non-backward-compatible IF changes will be behind us.
00:12 idiosyncrat And perhaps by then we'll have a better idea what to do with pre-lexeme events.
00:30 ronsavage That sounds good. And I see no emoticons, also good. I'm surprised you're keeping $recce->value(). Am I right in thinking: (a) Using the new while loop will - effectively - retrieve 1 tree per cycle. (b) $recce->value() will fatal because it's not the new, correct, way to handle ambiguity.
00:31 ronsavage And as for code which will be no longer necessary: I can cope :-)). Getting rid of my validate_events() method was very nice, with the cost moved into the event handlers, which is cleaner. So I'm all in favour of that. And it sounds like the while will also be a much nicer way of doing things.
00:33 idiosyncrat I'm keeping $recce->value() for 2 reasons
00:34 idiosyncrat 1.) The less important: It provides backward compatibility -- probably over 50% of Marpa::R2 logic will work without change;
00:36 idiosyncrat 2.) Going forward, most apps will want only one value, and having to explicitly create the new value object, then call a method, then ensure the object is cleaned up -- all that is a lot of clutter for something that was simple in Marpa::R2.  So I'll keep it simple.
01:11 ronsavage OK!
01:49 ilbot3 joined #marpa
01:49 Topic for #marpa is now Start here: http://savage.net.au/Marpa.html - Code paste/run: https://f.perlbot.pl/#marpa - Jeffrey's Marpa site: http://jeffreykegler.github.io/Marpa-web-site/ - IRC log: http://irclog.perlgeek.de/marpa/today - Youtube channel: https://www.youtube.com/channel/UCYKVfGBtfTqbs1JdYq-dc5g
03:15 ronsavage joined #marpa
03:31 ronsavage I've updated the Kollos FAQ: http://savage.net.au/Perl-modules/html/kollos.faq/faq.html.
06:50 ronsavage joined #marpa
10:55 kook joined #marpa
17:50 jdurand joined #marpa
18:53 idiosyncrat Does anybody know what the desktop is that Larry is using here? --  https://www.youtube.com/watch?v=EwQoSZdEl2s&t=482s
21:28 ernimril linux mint logo
23:03 ronsavage joined #marpa

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