Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-07-01

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

All times shown according to UTC.

Time Nick Message
00:30 hobbs joined #marpa
00:48 ronsavage joined #marpa
06:22 ronsavage joined #marpa
06:43 jeffreykegler joined #marpa
06:45 jeffreykegler jdurand: the workaround for no "soft" exhaustion is events -- there's enough of a variety of them that it's usually possible, depending on the grammar, to arrange for an event that triggers on exhaustion.
06:45 jeffreykegler In fact, I'd implement "soft" exhaustion by adding an "exhaustion event".
06:47 jeffreykegler One thing to note, if the concern is efficiency.  Much of the SLIF's event logic takes place in Perl code, while Perl's exception mechanism is implemented in C.  That means that, when I do add an exhaustion event, detecting the exhaustion event might very well clock in *slower* than catching an exception.
11:54 LLamaRider joined #marpa
18:12 jdurand_ joined #marpa
18:13 jdurand_ Re http://irclog.perlgeek.de/​marpa/2014-06-30#i_8952180 - c.f. https://github.com/jddurand/MarpaX-Languages-ECMA​Script-AST/blob/master/lib/MarpaX/Languages/ECMAS​cript/AST/Grammar/ECMAScript_262_5/Program.pm#L30
18:13 lucs Il est de retour!
18:14 jdurand_ Re http://irclog.perlgeek.de/​marpa/2014-07-01#i_8954164 - ah yes... you're right... I'll might rewrite the regexp in Program.pm using Marpa then -; I think this is the only place within the full ECMASCript stuff where I do not use only Marpa and bingo -; a bug
18:14 lucs jdurand_: Question: why does the produced AST use hashes? Wouldn't arrays of arrays have sufficed?
18:14 lucs [ $rule_id [@items]] or ['lexeme' [@three_values]]
18:16 jdurand_ lucs it could have been like that, only because 100% of the lexemes are not hardcoded in the grammar as 'something', but SOMETHING ~ 'something'
18:17 jdurand_ Nevertheless, what's wrong with hashes - later I started transpiling ECMAScript into another language, and felt easier from the programming point of view to access ruleId by $hash->{ruleId} instead of $hash->[0]
18:18 jdurand_ Oups, going to dinner
18:18 lucs I guess hashes are okay, but they'll most likely consume more memory, and runtime.
18:18 lucs Bonappétit!
18:18 lucs
18:18 * jdurand_ thx & AFK
18:40 jdurand_ I note that ECMAScript was it seems released before the apparition of 'inaccessible is ok by default' - that's why I imagined at that time the possiblity to 'inject' sub-grammars -;
18:41 jdurand_ Also I guess that was written before the ::lhs and ::name default adverb values
18:59 jdurand_ lucs, can you enter a github issue - tnx
19:13 lucs jdurand_: You mean for yesterday's bug?
19:16 jdurand_ yes
19:17 lucs Okay, sure.
19:17 jdurand_ Jeffrey, I am currently doing an innerr Marpa sub-grammar for this issue. It fits my need because I want to sub grammar to go up to the end. So the output of the subgrammar->read() should be ok in any case: no exception == success, exception == failure
20:32 jeffreykegler joined #marpa
22:02 jdurand_ inner grammar seems to work -;
22:08 jdurand_ oups, no - ok sorry for this live thingy - to be ct'ed

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