Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2015-03-01

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

All times shown according to UTC.

Time Nick Message
02:49 ilbot3 joined #marpa
02:49 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
03:03 jeffreykegler joined #marpa
05:36 sivoais joined #marpa
06:10 ronsavage joined #marpa
06:21 rns jeffreykegler: I ack'ed and got an idea how it's organized. Once you start on it, it'll be easier to subtask and I will be able to take individual items, which you can specify.
06:53 rns left #marpa
08:45 pczarn joined #marpa
11:17 sirdancealot joined #marpa
11:56 lwa joined #marpa
13:23 flaviu joined #marpa
17:19 jeffreykegler joined #marpa
17:26 jeffreykegler pczarn: re http://irclog.perlgeek.de/m​arpa/2015-02-28#i_10196407
17:28 jeffreykegler I should mention, as it is often not realized, that "parse success" is very subjective -- that is, in the last analysis it is up to the *user* to decide what success means.
17:28 jeffreykegler This may seem strange, because users get focused on a single application, and in its context success/failure is pretty obvious -- there's nothing really subjective about it.
17:29 jeffreykegler For example a language of matching parentheses succeeds when it is exhausted -- when it cannot go on because the parens have become balanced ...
17:30 jeffreykegler but there is also the question of whether there is additional, unconsumed input.
17:31 jeffreykegler But many programming languages can go on and on -- in typical perl, for example, you can just keep adding statements, so success in is terms of the point where you choose to stop, usually the end of input ...
17:31 jeffreykegler and success means the parse being successful at the end of input.
17:32 jeffreykegler In that kind of language, exhaustion usuallly means failure, and *not* success, because it means the parse has wound up in a situation where it will not be able to finish.
17:33 jeffreykegler This comes up often enough, that I gave it careful attention and wrote a document specificly on it: https://metacpan.org/pod/distribu​tion/Marpa-R2/pod/Exhaustion.pod
17:36 jeffreykegler The most general definition of "parse success" requires you to determine a specific end of input, and test for success there (which is that same thing of trying to take a parse value).
17:38 jeffreykegler But many users ask "Can Marpa tell me when a parse succeeds?", and of course in terms of this "most general definition", it is the APPLICATION which must determine the when, which makes the question pretty meaningless
18:02 pczarn I had a bug where I didn't expect it: I forgot to handle nulling symbols
18:05 pczarn how to go about parsing something like C++'s nested templates with `>>`?
18:06 pczarn LATM will recognize that as "right shift"
20:42 jeffreykegler Perhaps LATM  can tell the difference by context -- it is Longest *Acceptable* Token Matching, meaning that is a right shift does not lead to an acceptable parse at that point, it will not be accepted.
20:42 jeffreykegler I don't know if anyone has tried this in the C++ template vs. comparison case.
21:23 pczarn unfortunately, I'm getting input from another lexer
21:51 Aria Lots of commentary added to my in-progress marpa port: https://aredridel.github.io/lotsawa/
21:52 Aria Funny you were just talking about success. documenting the criteria for success I defined was so arbitrary.
23:56 koo7 wow, nice

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