Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2014-03-22

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

All times shown according to UTC.

Time Nick Message
14:01 lucs joined #marpa
19:43 jeffreykegler joined #marpa
20:50 jeffreykegler I'm getting a few failures on the latest bleeding edge developer's Perl: 5.19.10.  For the moment I'm assuming it's a Perl issue ...
20:50 jeffreykegler but if anyone has further insights, they'd be appreciated: http://www.cpantesters.org/cpan/report/2eb13546-b082-11e3-b80a-8c184f1022b2 is a typical example of these failure reports.
21:34 yxhuvvd btw I'm reimplementing the parser in Ruby from the paper you released. It's far from usable, but at least it feels nice to realize that the palindrom parser I use for testing fails to parse 'abaaba' since the grammar I used apparently only works for odd numbers :D
21:36 yxhuvvd now if that leo paper wasn't hidden behind a paywall :/
21:37 jeffreykegler Interesting and a very welcome development\
21:40 jeffreykegler1 joined #marpa
21:40 jeffreykegler1 Two comments: does Ruby allow interfaces with C libraries?  Can you write it as an interface to Libmarpa?  (Or perhaps you're doing it as a programming exercise.)
21:40 jeffreykegler1 Second comment: the paper is being totally rewritten at this point -- I changed the algorithm to not use LR(0) states, which makes it much simpler and more flexible.
21:43 yxhuvvd There is a lib called FFI, so that would be totally possible I suppose, given it was possible to extract it in a usable way. my first approach was to interface with it, but I ended that after not being able to figure out how to build it (or even how to install it. I'm *really* not a perl guy).
21:44 yxhuvvd even tried to look for the source and see if I could build it manually but ./configure crashed a few thousand lines in so I gave up that.
21:45 jeffreykegler1 The way to install Libmarpa is to install Marpa::R2 -- you don't have to use Marpa::R2, just install it
21:45 yxhuvvd as for the algorithm with lr(0) states, I find them pretty straightforward so I'll keep them for now.
21:45 jeffreykegler1 Think of Perl as your test language
21:46 yxhuvvd well, that is what I tried, and failed, to install. personal guess is that ubuntu have too new build tools, but I'm not certain that is the case so it may be my error as well :)
21:52 jeffreykegler1 yxhuvvd: Also, somewhere in the marpa-parser Google group archives, Peter Stuifzand did a port of the library to GO, and he describes how he built it.
21:52 yxhuvvd interesting. I'll look for it
21:53 jeffreykegler1 Anyway, running the build procedure leaves a fully built library somewhere
21:53 jeffreykegler1 On this Linux box, the library is named auto/Marpa/R2/R2.so
21:53 yxhuvvd the thing is, I ended up with an error when trying to install it.
21:54 jeffreykegler1 You could not install Marpa::R2?
21:54 yxhuvvd right. maybe by doing wrong though.
21:55 jeffreykegler1 Yes, installing Marpa::R2 is something you'll want to do first, if only as a test -- you don't actually have to use it after it's installed.
21:56 ronsavage joined #marpa
22:04 yxhuvvd ah. he links to the lib. that explains how it can be so short and simple :)
22:06 ronsavage jeffreykegler: Did you see http://www.nntp.perl.org/group/perl.perl5.porters/2014/03/msg213874.html
22:12 jeffreykegler1 ronsavage: Thanks.  It looks like the problem is a regression in the development Perl.
22:47 jeffreykegler joined #marpa
23:02 jeffreykegler1 joined #marpa
23:18 jeffreykegler joined #marpa
23:45 jeffreykegler1 joined #marpa
23:55 jeffreykegler joined #marpa

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