Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2016-08-01

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
03:22 travis-ci RPerl build failed. bulk88 says 'WIP PERL_NO_GET_CONTEXT'
03:22 travis-ci https://travis-ci.org/wbraswell/rperl/builds/148765691 https://github.com/wbraswell/rperl/commit/b6e56565466c
03:27 willthechill joined #perl11
03:30 willthechill bulk88: I see you are trying to pass some Perl context info in lots of places but it is failing on travis
03:31 willthechill in typemap.rperl you have changed all T_PACKED to T_PACKETCXT but then you redefined T_PACKED
03:31 willthechill did you mean the new typemap definition to be T_PACKEDCXT instead of T_PACKED?
03:31 willthechill https://github.com/wbraswell/rperl/commit/b6e56565466c#diff-6e540fc00a94bb449dd9c96ff1125fd4R33
10:23 dalek joined #perl11
11:14 travis-ci perl11/cperl#1369 (smoke/gh190-mderef-loop - d92acc3 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/148839761
13:07 travis-ci perl11/cperl#1370 (smoke/gh190-mderef-loop - 2adedcd : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/148861635
13:16 mako joined #perl11
16:33 bulk88 willthechill yeah, thats a bug
16:33 bulk88 im currently in alabama and programming is kindda hazy for my head
16:41 willthechill bulk88: okie dokie
16:42 willthechill bulk88: I will wait for you to be back in full swing
18:22 travis-ci RPerl build failed. bulk88 says 'WIP PERL_NO_GET_CONTEXT'
18:22 travis-ci https://travis-ci.org/wbraswell/rperl/builds/148942355 https://github.com/wbraswell/rperl/compare/b6e56565466c...e6d1641e6a78
18:22 travis-ci perl11/cperl#1372 (smoke/maint-5.24c - 206c8da : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/148944328
19:01 willthechill bulk88: I see you fixed typemap.rperl to define T_PACKEDCXT but we are still failing on travis
19:01 willthechill does this code pass tests on your personal machines?  specifically we see failures starting in t/04_type_scalar.t
19:16 travis-ci perl11/cperl#1373 (maint-5.24c - 206c8da : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/148951284
19:18 bulk88 willthechill it does NOT pass on my machine
19:21 bulk88 #   Failed test 'main::RPerl__DataType__Integer__MODE_ID() lives'
19:21 bulk88 how do I debbug that?
19:21 bulk88 not ok 111 - main::RPerl__DataType__Number__MODE_ID() lives
19:21 bulk88 #   Failed test 'main::RPerl__DataType__Number__MODE_ID() lives'
19:21 bulk88 #   at t\04_type_scalar.t line 111.
19:21 bulk88 # died: Can't use an undefined value as a subroutine reference at t\04_type_scal
19:21 bulk88 ar.t line 109.
19:26 willthechill bulk88: just a min, will help you debug it
19:34 willthechill bulk88: okay here we go
19:35 willthechill the following code should display 0, which is the MODE_ID for PERLOPS_PERLTYPES
19:35 willthechill perl -e 'use RPerl::DataType::Integer_cpp; print main::RPerl__DataType__Integer__MODE_ID(); print "\n";'
19:35 willthechill now this code should display 2, which is the MODE_ID for CPPOPS_CPPTYPES
19:35 willthechill perl -e 'use RPerl::DataType::Integer_cpp; RPerl::DataType::Integer_cpp::cpp_load(); print main::RPerl__DataType__Integer__MODE_ID(); print "\n";'
19:35 willthechill oh wait I guess you also need CPPOPS_PERLTYPES because that is what you are currently working on...
19:36 willthechill well try those 2 first while I work on the other one
19:43 willthechill bulk88: okay this will do what you need in CPPOPS_PERLTYPES mode, which is displayed as MODE_ID of 1
19:43 willthechill perl -e 'use RPerl; rperltypes::types_enable("PERL"); use RPerl::DataType::Integer_cpp; RPerl::DataType::Integer_cpp::cpp_load(); print main::RPerl__DataType__Integer__MODE_ID(); print "\n";'
19:43 willthechill important!  you need to delete _Inline/ directory before running this
19:43 willthechill it definitely can and will stay in CPPOPS_CPPTYPES mode if it finds a cached version in _Inline/
19:44 willthechill (this is a limitation of how RPerl interacts w/ Inline::CPP and will have to be addressed at some point, but not today)
19:56 bulk88 perl -e 'use RPerl::DataType::Integer_cpp; print main::RPerl__DataType__Integer__MODE_ID(); print "\n";' gives me 0
19:58 willthechill yes that is correct
19:58 willthechill a MODE_ID of 0 is PERLOPS_PERLTYPES
19:58 willthechill MODE_ID of 1 is CPPOPS_PERLTYPES
19:59 willthechill MODE_ID of 2 is CPPOPS_CPPTYPES
20:00 bulk88 C:\sources\rperl>perl -Ilib -e "use RPerl::DataType::Integer_cpp; RPerl::DataType::Integer_cpp::cpp_load(); print main::RPerl__DataType__Integer__MODE_ID(); print \"\n\";"
20:00 bulk88 2
20:00 willthechill excellent
20:01 willthechill now make sure you delete _Inline/ folder before going for the next one
20:01 bulk88 your types_enable is before a use line, is that intentional?
20:01 willthechill yes intentional
20:02 willthechill types_enable() sets one of the gcc -D compiler flags
20:02 bulk88 you know the use executes before hand?
20:02 willthechill so it must be before 'use' and 'cpp_load'
20:02 willthechill oh right
20:02 willthechill good point
20:02 willthechill is is only the cpp_load() that must be done AFTER types_enable()
20:02 willthechill so yeah
20:03 bulk88 BEGIN block or something u need, if its relaly needed, otherwise reorder for snaity reasons
20:03 willthechill no need for sanity here, this is just a hand-coded debug test
20:03 willthechill ;-)
20:03 willthechill no we don't need to worry about a BEGIN block
20:03 willthechill if it somehow helps you to debug it by moving the use to the beginning then that's fine
20:03 willthechill the point is that we see an output of MODE_ID being 1
20:04 willthechill that is CPPOPS_PERLTYPES mode
20:04 willthechill and is what you are currently debugging
20:04 willthechill although I assume we will see an error instead of MODE_ID 1, because that's kindof the point
20:05 willthechill we WANT to see MODE_ID of 1 because that means your new T_PACKEDCXT stuff is fixed
20:05 willthechill or at least less broken, haha!
20:07 bulk88 C:\sources\rperl>perl -Ilib -e "use RPerl; rperltypes::types_enable(\"PERL\"); u
20:07 bulk88 se RPerl::DataType::Integer_cpp; RPerl::DataType::Integer_cpp::cpp_load(); print
20:07 bulk88 main::RPerl__DataType__Integer__MODE_ID(); print \"\n\";"
20:07 bulk88 0
20:09 willthechill okay well if you get a 0 there instead of a 1, then hopefully you also got a bunch of useful new error messages
20:09 willthechill if not then you probably need to set your RPERL_DEBUG=1 & RPERL_VERBOSE=1 env vars
20:27 travis-ci perl11/cperl#1374 (maint-5.22c - 679e3f1 : Father Chrysostomos): The build passed. https://travis-ci.org/perl11/cperl/builds/148955487
20:29 travis-ci perl11/cperl#1374 (maint-5.24c - 488286f : Father Chrysostomos): The build passed. https://travis-ci.org/perl11/cperl/builds/148955498
20:29 willthechill kentnl: any updates for me?
20:30 willthechill :-)
20:30 willthechill mako: how goes it?
20:32 bulk88 willthechill nearly all the XSUBs are missing from, the .c file made in  _inlione/build
20:33 bulk88 so the pTHX_ conditional arguyment is makign whatever findss the C/C++ funcs, and wraps them, makes those funcs disappper
20:34 willthechill bulk88: not sure I totally follow you
20:34 willthechill you are saying we are missing XSUBS from the files in the _Inline/build/ directory
20:35 willthechill then from there the pTHX changes you have recently made is causing some errors?
20:36 willthechill I did not design Inline and I don't write XS by hand, so I don't exactly understand what you're saying, sorry  :-P
20:48 mako willthechill: not so well. I couldn't come up with anything useful. Had a job interview and not much time. This should change the next days.
20:49 willthechill mako: sounds good!
20:49 willthechill slow and steady wins the race
20:49 willthechill :-)
20:53 mako You're, captain. :-)
20:53 mako You're right, captain! That is.
20:54 willthechill *Starfleet salute*
20:55 willthechill :-D
20:56 mako Applies somehow to RPerl, too: "To boldly go where no Camel has gone before." :-D
20:57 willthechill haha nice one!  :-)
20:58 willthechill the official Space Scout mission: "... and to boldy proceed as none have done before"
20:58 willthechill (that's just the last part)
21:06 mako That means we've got a lot of work to do.
21:07 willthechill oh yes, very much so
21:07 willthechill got some VERY exciting stuff coming up in the near future
21:07 willthechill some stuff that will really put RPerl on the map globally
21:07 willthechill but that doesn't mean I get to skip work on the small stuff today, haha
21:13 mako Yeah, I've already seen you and bulk88 coping with some XS and Inline::CPP issues. At least it looks like that.
21:14 willthechill yes that is today's mud pit  ;-)
21:18 mako Like coping with the pig in the mud pit. And the pig liked it. :-D
21:20 willthechill yes except in this case the pig is XS and is somewhat less-than-conscious of our struggle, haha!
21:34 mako Dealing with XS can be very annoying. That's the reason why I used 'FFI::Raw' to access code written in C.
21:44 mako left #perl11
22:03 willthechill joined #perl11
22:17 esaym153 joined #perl11
22:26 willthechill joined #perl11
22:59 travis-ci perl11/cperl#1375 (smoke/gh190-mderef-loop - 22c2b74 : Father Chrysostomos): The build passed. https://travis-ci.org/perl11/cperl/builds/149030600

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary