Perl 6 - the future is here, just unevenly distributed

IRC log for #marpa, 2016-04-01

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

All times shown according to UTC.

Time Nick Message
00:36 idiosyncrat_ Marpa::R3 is in a repo on Github.
00:37 idiosyncrat_ The conversion to ExtUtils::MakeMaker is complete.
00:37 idiosyncrat_ All tests now succeed.
00:38 idiosyncrat_ ronsavage: Module::Build is now out of the Perl core, and was difficult with a module whose build is as complex as Marpa's.
00:39 idiosyncrat_ There are still various cleanups to do.
00:40 idiosyncrat_ choroba: not sure what call is returning the exception.
00:41 idiosyncrat_ choroba: IIRC you can be current position using standard methods, and also convert it to line,column with other standard SLIF recognizer methods.
01:40 ronsavage joined #marpa
01:40 ronsavage choroba: Re http://irclog.perlgeek.de/m​arpa/2016-03-31#i_12268274. savage.net.au/Perl-modules/h​tml/marpa.faq/faq.html#q117.
03:35 idiosyncrat_ Could somebody do me a favor and try to build Marpa::R3?
03:36 idiosyncrat_ I'll think there are likely to be problems about files missing in the MANIFEST, and it would save me time if someone would spot these for me.
03:36 idiosyncrat_ Once we know it's build-able, I'll add a few of the easiest features, and we're in the more fun territory.
03:41 idiosyncrat_ Use commit f707eedc1785352de4eee71f1528eba198e88a7b in the Marpa::R3 repo --
03:41 idiosyncrat_ All tests in the test suite pass on my development box.
03:42 idiosyncrat_ Thanks!  And ...
03:42 idiosyncrat_ Good night!
04:39 rns joined #marpa
04:45 jdurand joined #marpa
04:47 jdurand Re http://irclog.perlgeek.de/m​arpa/2016-03-30#i_12258441 - almost that - I'd go further, exploiting the fact that Marpa allows to read things that are NOT in physical input - this can be used to drive the lexer direction and to make sure there is no ambiguity
04:47 jdurand For example, to match the rule A ::= B - C I'd rewrite it to:
04:48 jdurand A::= B (BOK)
04:48 jdurand A::= C
04:48 jdurand with completion events on B and C
04:48 jdurand BOK is an internal lexeme, that will not affect the position when it is lexeme_read()'ed
04:49 jdurand Proof of concept in the pastebin
04:49 shadowpaste0 "jdurand" at 217.168.150.38 pasted "Ho do do exclusion with completion events and a fake lexeme" (96 lines) at http://fpaste.scsys.co.uk/509264
04:51 rns idiosyncrat: re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269163
04:51 shadowpaste0 "rns" at 217.168.150.38 pasted "perl Makefile.PL (first run from freshly cloned repo)" (6 lines) at http://fpaste.scsys.co.uk/509265
04:55 jdurand Re my paste : oups maybe an error in it, but the idea remains : use a fake lexeme to drive the lexer
05:00 rns on second run, Makefile.PL tries to run itself and fails with
05:00 rns Can't open perl script "Makefile.PL": No such file or directory
05:00 shadowpaste0 "rns" at 217.168.150.38 pasted "MARPA_USE_PERL_AUTOCONF=1 perl Makefile.PL (second run)" (97 lines) at http://fpaste.scsys.co.uk/509267
05:04 shadowpaste0 "rns" at 217.168.150.38 pasted "patch to remove code running Makefile.PL from itself" (21 lines) at http://fpaste.scsys.co.uk/509268
05:05 shadowpaste0 "jdurand" at 217.168.150.38 pasted "Ho do do exclusion with completion events and a fake lexeme - take 2" (96 lines) at http://fpaste.scsys.co.uk/509269
05:06 jdurand i.e. the exclusion rule is:
05:06 shadowpaste0 "rns" at 217.168.150.38 pasted "perl Makefile.PL after http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269315" (102 lines) at http://fpaste.scsys.co.uk/509270
05:07 jdurand exclusion_rule (MARKER)
05:07 jdurand exclusion_rule ::= wanted | not_wanted
05:07 jdurand with completion events on wanted and not_wanted
05:07 rns after the above patch, perl Makefile.PL runs ok, see above -- http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269322
05:08 jdurand Have a nice day (ps: thnx rns for the tests -: I have to leave)
05:10 rns jdurand: have a nice day, ps: using events to set zero-length markers -- https://metacpan.org/pod/distribut​ion/Marpa-R2/pod/Event.pod#Markers -- is way cool
05:12 shadowpaste0 "rns" at 217.168.150.38 pasted "MARPA_USE_PERL_AUTOCONF=1 make" (7 lines) at http://fpaste.scsys.co.uk/509271
05:14 rns then make fails, see above. has something to do with using Config::AutoConf obviously, assuming your dev box uses configure.
05:15 rns Config::AutoConf is 0.311, perl 5, version 22, subversion 1 (v5.22.1) built for cygwin-thread-multi-64int
05:16 ronsavage joined #marpa
05:17 jdurand joined #marpa
05:18 jdurand Re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269334 - yes but without explicit lexemes the parse tree value with have ambiguities
05:19 jdurand so in my idea using a nulled event has no advantage over a completion event, well I think - I'll clone the R3 repo tonight
05:20 rns ok, great. I'll have to look closer into your proofs of concept later this day.
05:22 rns re Marpa::R3 the same thing on windows with Config::AutoConf -- make file missing in engine/perl_ac_build
05:52 ronsavage joined #marpa
07:42 rns with cgywin's configure, Marpa::R3 complains about missing general_pattern.xsh:
07:42 shadowpaste0 "rns" at 217.168.150.38 pasted "perl Makefile.PL && make with cygwin's configure" (337 lines) at http://fpaste.scsys.co.uk/509277
07:43 rns "/usr/bin/perl.exe" "/usr/lib/perl5/5.22/ExtUtils/xsubpp"  -typemap "/usr/lib/perl5/5.22/ExtUtils/typemap" xs/R3.xs > xs/R3.xsc && mv xs/R3.xsc xs/R3.c
07:43 rns Cannot open 'general_pattern.xsh': No such file or directory in R3.xs, line 7071
08:40 teatime joined #marpa
09:23 rns idiosyncrat: re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269761 -- hopefully fixed in https://github.com/jeffreykegler/Marpa--R3/pull/1
09:38 rns idiosyncrat: re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269364 -- hopefully fixed in https://github.com/jeffreykegler/Marpa--R3/pull/2
09:39 koo7 joined #marpa
11:04 kaare_ joined #marpa
15:25 idiosyncrat_ joined #marpa
15:26 idiosyncrat_ rns: Thanks for the two pull requests!
15:28 idiosyncrat_ rns: re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12269334 -- yes, using events as zero-length markers was part of my intent for them & Jean-Damien was the pioneer in their use.
18:00 jdurand joined #marpa
18:00 jdurand Ron... where did you the BNF for regexp from...?
18:07 jdurand make releng on Marpa--R3 ok on debian GNU/Linux
18:27 choroba joined #marpa
18:42 jdurand Hqs qnyone see the Ubuntu buzz for bash on Windows ? Despite the fact that bash on Windows is not a new thing (-;) - I remember well having in mind the project to write a bash parser in perl - thus making bash "everywhere" - since (almost) all portability issues with OS are then handled by perl -;
18:43 jdurand "Has anyone" ? Sorry - moving from Swiss-French keyboard (QWERTZ) to French (AZERTY)!
18:45 choroba ronsavage: Re http://irclog.perlgeek.de/m​arpa/2016-04-01#i_12268906 Thanks, that's it!
18:48 jdurand In Marpa::R3 I also have a pb with MARPA_USE_PERL_AUTOCONF=1: Running [/home/jdurand/perl5/perlbrew/perl​s/perl-5.23.7-debug/bin/perl5.23.7 Makefile.PL]... Can't open perl script "Makefile.PL":  No such file or directory
22:29 idiosyncrat_ joined #marpa

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