Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-06-15

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

All times shown according to UTC.

Time Nick Message
00:01 rgrinberg joined #marpa
01:19 ronsavage joined #marpa
01:48 ilbot3 joined #marpa
01:48 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
02:52 kaare_ joined #marpa
04:51 idiosyncrat_ Good night!
05:43 ronsavage joined #marpa
11:16 VsyachePuz joined #marpa
11:33 VsyachePuz joined #marpa
12:24 rgrinberg joined #marpa
13:11 rgrinberg joined #marpa
13:25 JPGainsborough joined #marpa
13:37 ceridwen joined #marpa
16:58 idiosyncrat_ joined #marpa
17:56 jdurand joined #marpa
17:57 jdurand Jeffrey, there is one thing unclear to me with marpa_g_sequence_new: its consequence on the parse tree. The doc says it is optimized out, fair enough, but suppose I declare
17:58 jdurand a sequence with marpa_g_sequence_new with a min of 1 of for the symbol X. Then I push and complete the same symbol X three times: will the parse tree call
18:01 jdurand for a value once with {arg_0,arg_n}={0,2}, or call first with {arg_0, arg_n}={0,1}, then for a concatenate with {2}, plus a call with {arg_0,rarg_n}={1,2}, then an append to {0}
19:01 idiosyncrat_ jdurand: No it calls once for the whole sequence.
19:02 idiosyncrat_ That's what I meant by "optimized out" -- all those trivial evaluate, concatenate steps are avoided.
19:02 idiosyncrat_ If done naively, they make the evaluation of a sequence O(n^2) in its length.
19:03 idiosyncrat_ Btw, that's why I'm now doing the Lua semantics.
19:03 idiosyncrat_ Originally, I was going to add features to the grammar, such as more flexibility in defining sequences in the SLIF's DSL.
19:05 idiosyncrat_ But I realized that with my current semantics logic, they would make evaluation O(n^2), and I'd need a much more complex virtual machine to avoid that.
19:05 idiosyncrat_ So, therefore, time to make my semantics Lua-based.
19:06 idiosyncrat_ This will defer my original goal -- more flexible syntax for SLIF sequences -- for some time ...
19:07 idiosyncrat_ but adding Lua will have many nice side effects and is something I intended to do some day, so why not now?
19:07 idiosyncrat_ jdurand: I think that answered your question.  I hope so. :-)
19:31 sivoais joined #marpa
19:44 VsyachePuz joined #marpa
19:49 VsyachePuz joined #marpa
20:59 rgrinberg joined #marpa
21:02 jdurand yep - many thanks
21:58 rgrinberg joined #marpa
22:32 idiosyncrat_ joined #marpa
23:04 ronsavage joined #marpa
23:08 lucs_ joined #marpa
23:08 btyler_ joined #marpa

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