Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-12-26

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

All times shown according to UTC.

Time Nick Message
00:00 ronsavage joined #marpa
00:00 ronsavage Back! I changed my mind about going shopping.
00:16 ronsavage What I really want is some way of exporting constants, and using my(xxx) doesn't do that. Any ideas?
00:16 ronsavage This is for a set of flags, so a list of subs does not really appeal, but may be what I have to use.
00:17 ronsavage I tried Scalar::Readonly too. Same issue (my).
00:20 koo6 joined #marpa
00:25 jeffreykegler ronsavage: Can't help you.  I gave up on the standard solutions in Marpa::R2 and rolled my own.
00:44 ronsavage Thanx. I've gone with Const::Exporter, which suits me perfectly.
02:45 jeffreykegler I've written a document describing how Marpa can parse those grammars which it parses in linear time, in constant space as well:
02:45 jeffreykegler https://github.com/jeffreykegler/kollos/​blob/master/notes/misc/constant_space.md
02:45 jeffreykegler This technique could be built into Marpa at some point in the future, so that it happens automatically
02:47 jeffreykegler it could also with a bit of effort be used for specific grammars today, using Marpa's current facilities.
03:08 jeffreykegler By the way, one reason I'm aggressive about "tricks" for extending Marpa's capabilities, is that I hope someday Earley/Leo parsing will be implemented in chips, the way graphics processing is now ...
03:09 jeffreykegler if you think about it, parsing is finding structure in sequences of symbols, which is a *very* general problem, not totally unlike floating point and graphics, which special hardware currently target.
03:10 koo5 with parallelization?
03:11 jeffreykegler But yacc was a failure, and recursive descent must be hand-written, so dedicated hardware will not help there.
03:12 jeffreykegler koo5: I'm a software guy.  What happens to an algorithm once the EE's start working on dedicated hardware isn't my area of expertise.
03:12 jeffreykegler I have done firmware, but actually designing chips is outside my area.
03:12 koo5 hmm
03:13 jeffreykegler But I'd assume they'd parallelize, pipeline, etc., etc.
03:14 jeffreykegler And with that Marpa gains a fast constant factor.  Perhaps 10x?
03:18 ronsavage I'd be very grateful if anyone could proof-read the docs for Text::Balanced::Marpa: http://savage.net.au/Perl-module​s/html/Text/Balanced/Marpa.html TIA!
03:20 jeffreykegler Is it on github, or should we just do fixes on IRC, or what?
03:20 ronsavage The module has not been released, but it is on github: https://github.com/ronsavage/Text-Balanced-Marpa
03:21 ronsavage Comments here say, and patches via github.
03:21 ronsavage AFK
03:27 jeffreykegler ronsavage: "The default string length returned is 20 characters."  -- "default" suggests it can be changed.  Can it?
03:32 jeffreykegler ronsavage: I created a Github PR with two changes, one very nit-picky.
04:14 sirdancealot joined #marpa
05:40 sirdancealot joined #marpa
05:47 ronsavage joined #marpa
05:50 ronsavage jeffreykegler: I don't see any PRs on github!?
05:51 ronsavage And, I've been thinking. I'll remove all logging, since most people won't want that, and I'll put real code in the Synopsis (which is broken without it).
06:53 rns joined #marpa
07:07 ronsavage OK. Synopsis fixed (+ scripts/synopsis.pl). Logging removed. Added error_message() and error_number() for both warnings and errors. All documented.
07:10 rns left #marpa
09:21 lwa joined #marpa
09:36 ronsavage Hmmm. This is interesting: http://www.nntp.perl.org/group/perl.​perl5.porters/2014/12/msg223955.html
09:39 ronsavage And now we want a parser for the pack/unpack format: http://www.nntp.perl.org/group/perl.​perl5.porters/2014/12/msg223956.html
09:41 ronsavage And a parser for for printf's format: http://www.nntp.perl.org/group/perl.​perl5.porters/2014/12/msg223958.html
10:05 ronsavage Namespace suggestion: MarpaX::Languages::Perl::PackUnpack and *::Printf.
10:06 ronsavage Grrrr. MarpaX::Languages :: Perl:: PackUnpack and *:: Printf
16:53 jeffreykegler joined #marpa
17:45 sirdancealot joined #marpa
20:23 koo5 joined #marpa
20:39 flaviu Marpa::R2::Advanced::Bibliography has a dead link: http://www.cs.vu.nl/~dick/PTAPG.html
20:39 flaviu I think that http://dickgrune.com/Books/PTAPG_1st_Edition/ is the same page, just moved
20:43 flaviu Leo 1991 is available online at http://www.sciencedirect.com/scie​nce/article/pii/030439759190180A
20:44 flaviu sciencedirect has a confusing website, but the link at the top "Download full text in PDF" allows you to read it
20:45 flaviu I don't think it's possible to link directly to the pdf though, it looks like it's time-limited.
21:23 sirdancealot joined #marpa
22:55 jeffreykegler joined #marpa
22:55 jeffreykegler flaviu: Thanks!  Could you submit a Github PR with those two fixes?
22:57 flaviu Of course, just a few moments
23:04 jeffreykegler flaviu: Got it.  Thanks!
23:31 ronsavage jeffreykegler: Re: http://irclog.perlgeek.de/​marpa/2014-12-26#i_9853946. I've added a mutator next_few_limit().
23:33 ronsavage jeffreykegler: Thanx for the github patches. I didn't ever receive the email, but I've merged in your ideas. And I fear every one of my Perl modules has the same (little) typo in the installation instructions :-).
23:53 flaviu joined #marpa

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