Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2017-09-23

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

All times shown according to UTC.

Time Nick Message
01:20 idiosyncrat joined #marpa
01:21 idiosyncrat Demat!
01:56 ilbot3 joined #marpa
01:56 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
03:55 idiosyncrat I've been working on converting parse evaluation into its own object on the level of grammars and recce's.
03:56 idiosyncrat Coding and testing are now done -- on to documentation.
03:56 idiosyncrat In Marpa::R2, the valuer was built into the recognizer and there could be only one.
03:57 idiosyncrat If you wanted to start over, possibly with new setting, there was a $recce->series_restart() method.
03:59 idiosyncrat Also, ASF (parse forests) used that same valuer, so you couldn't do both.  This was especially inconvenient if you asked for (or automatically were given) a description of a parse's ambiguity, because the ASF's were used to do this, and took over the valuer.  So you couldn't just go on getting values.
04:00 idiosyncrat All that is fixed in Marpa::R3.  You can have as many valuers, and as many ASF's as you want.  To start over, you just create a new valuer, so $recce->series_restart(), along with a whole lot of internal state, is no longer necessary.
04:02 idiosyncrat The fashion, when Marpa::R2 started, was for overuse of objects.  I deliberately went minimal, creating those objects which clearly justified their existence, and no others.
04:03 idiosyncrat But in the case of valuers I went too far.  I built them into the recognizer, but they were always "wanted" to be objects on their own.  Now they are.
04:26 idiosyncrat As one example of the improvement: the change allows me to eliminate my POD document on "parse phases" entirely: https://metacpan.org/pod/distribution/Marpa-R2/pod/Semantics/Phases.pod
04:28 idiosyncrat How Marpa handles the "phases" of reading, multiple valuers, etc., which got quite complex when it invovled coordination of different method calls for the same object, is now obvious from the object model.
04:29 idiosyncrat Mostly that document explained limitations and gotchas which no longer exist.
07:32 kaare__ joined #marpa
09:45 kaare__ joined #marpa

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