Camelia, the Perl 6 bug

IRC log for #ironperl, 2009-03-16

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

All times shown according to UTC.

Time Nick Message
20:38 diakopter joined #ironperl
20:38 diakopter cj: I didn't disappear...
20:45 cj 11:35 [Users #ironperl]
20:45 cj 11:35 [Users #ironperl]11:35 [ cj] [ ilbot2]
20:45 * cj kicks irssi
20:45 diakopter eh
20:45 cj sure, that was on the 12th, but... ya' know... I haven't checked since then
20:46 diakopter so, I tried to write a PEG parser for Perl 6 in NPEG b/c I was getting frustrated with STD/viv's YAML output
20:46 diakopter and got fairly far on it
20:47 cj oh?
20:47 diakopter but it's obscenely slow
20:47 cj and then the world began to burn?
20:47 diakopter just like the packrack parser/combinator system I wrote in Perl 5 last year based on the combinators in MJD's Higher Order Perl book
20:47 diakopter (which was also obscenely slow)
20:47 diakopter <sigh>
20:53 diakopter but ultimately, even though TimToady's parser is technically a one-pass parser, it needs very tight integration
20:55 diakopter and I'm just not sure it's possible, actually
20:55 diakopter anymore
20:57 diakopter to use it from an environment that's not seamlessly integrated with C
20:58 diakopter ... because of the way the design of the language has adapted to the existing implementations
20:58 diakopter bbl
21:08 diakopter (back)
21:09 cj strange that you mention MJD... I've just been talking with him...
21:09 diakopter cj: let me know if you have any comments on the above.....?
21:09 diakopter oh?  heh;  I was chatting with him on irc
21:09 diakopter this morning
21:09 diakopter asking him if he wouldn't mind exchanging some email about Perl 6 and such.
21:10 diakopter he said that would be fine
21:10 diakopter I haven't written the first email to him though
21:11 cj how do you mean "seamlessly integrated with C"?
21:11 diakopter meaning, I need to interact with the parser (running in Perl 5) both synchronously and asychronously
21:11 cj I'm sure Larry would like to hear how you feel that the design of the language has made itself dependent on its implementations.  I have a feeling that one of the design goals is to avoid such dependence
21:12 diakopter no, that's not one of the design goals
21:13 cj no?  I've heard a number of times that there should be multiple competing implementations that adhere to a standard/test suite...
21:13 cj "specification" as it were
21:13 diakopter but the "grammar" (let's be honest, it's utterly unformalizable) is really just a highly complex program with tons of custom behavior.... not formalizable in a "grammar" at all
21:13 diakopter sure, sort of
21:14 diakopter yes, there should be the competing implementations, but there's really not
21:14 cj pugs, parrot, rakudo?
21:14 diakopter rakudo struggles to keep up with grammar/parser changes, and the author is currently bogged down in implementing large changes to it... could take months to sort out.
21:15 diakopter pugs has been dead for years
21:15 diakopter basically no one is contributing new code to parrot, nor has for months, really.
21:15 diakopter except the implementors of rakudo
21:15 diakopter well, allison has
21:15 diakopter by new code, I mean implementations of new functionality
21:16 diakopter featureset-expanding functionality
21:16 diakopter yes, lots of bugfixes go in
21:16 diakopter and documentation fixups
21:18 diakopter in my not-terribly-humble opinion, parrot is totally mired in bureaucracy and over-lofty design goals to make any progress with the scant resources it has (the few people willing to continue to hack on it a couple of hours per week), but there are still tens of thousands of hours of work remaining for parrot to get to where they set out to be
21:18 diakopter yes, the scant resources who are committed to the project are very diligent, very vigilant, and very smart
21:18 diakopter but there just aren't 100 of them
21:19 diakopter no one has added any new features to pugs for years
21:20 cj so what's the implementation folks are using?
21:20 diakopter jonathan and patrick continue to do heroic work with rakudo to make it fit with a stodgy, slow parrot
21:20 diakopter using?
21:20 cj no?
21:20 diakopter who uses any implementation, really?
21:20 diakopter there are couple of POC apps here and there
21:20 diakopter using rakudo
21:20 diakopter and the author of Pugs uses it for a lot of things
21:21 diakopter and various haskell folk use Pugs for its Perl 6-ish features
21:21 diakopter cirica 2007
21:21 diakopter circa
21:21 diakopter rakudo still can't do IO except for to/from files
21:23 diakopter but seriously, who else uses any of the Perl 6 implementations besides what I've mentioned
21:23 cj I've wondered that myself...
21:25 diakopter no one... but like I said there are a few POC app[lets]
21:25 diakopter gtg.. will be back online later tonight

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