Perl 6 - the future is here, just unevenly distributed

IRC log for #inline, 2015-01-07

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

All times shown according to UTC.

Time Nick Message
06:35 willthechill joined #inline
12:11 mohawk evidently he thinks it's fantastic
12:12 mohawk i will put that davido___ unconditionally endorses it on the PR
17:42 davido___ mohawk : Sorry, been busy elsewhere.  Which PR should I be looking at?
17:43 mohawk https://github.com/Perl-Toolchain-Gang/ExtUtils-MakeMaker/pull/195
17:43 mohawk davido___, hope your busy stuff has been resolved to your satisfaction
17:44 davido___ Not yet resolved, just breaking away so nobody panics and writes my epitaph.
17:44 mohawk ha ha
17:44 mohawk well, i hope it's going well enough, then
17:45 davido___ I like the premise of #195.
17:45 mohawk me too
17:45 mohawk rehsack is saying "oh it's the wrong solution", so that will need addressing
17:47 davido___ It would help if there's an explanation of why it's wrong, but if he has a legitimate beef, I'm sure there's some common ground somewhere.
17:47 mohawk it would be great if he did
17:48 davido___ As for EUCG... my heart is in it, but my time isn't, at the moment.  If you want, tsee would probably grant co-maint / commit bit to you.
17:49 mohawk davido___, please could you recommend to him now that he do that?
17:49 davido___ I would suggest one thing though:  Look at Inline::CPP's Makefile.PL, and https://github.com/daoswald/Inline-CPP/blob/master/inc/ILCPPConfig/CompilerGuess.pm, to see those areas where Inline::CPP still has to work to get a useful answer.
17:49 mohawk i also want the ability to release it to CPAN
17:49 mohawk davido___, don't worry, i'm going to inhale all of I::CPP's logic
17:49 mohawk then it will just be able to use EUCG, as a one-liner
17:50 davido___ If you want to see Inline::CPP's "old" logic, that's in https://github.com/daoswald/Inline-CPP/blob/master/inc/ILCPPConfig/OldCompilerGuess.pm
17:51 mohawk if you have a moment, could you get in comms with tsee and recommend me getting CPAN co-maint, and a commit bit on his repo?
17:51 mohawk i'll make/add to an issue on that repo now with that note
17:51 mohawk (the OldCompGuess thing)
17:51 davido___ Yes, doing that now.
17:52 mohawk nice - making an issue now
17:53 davido___ mohawk: So there are three Inline::CPP files to look at, the "old", the "new", and the Makefile.PL
17:53 mohawk https://github.com/tsee/extutils-cppguess/issues/7
17:53 mohawk sorry, where are the old and new located?
17:54 mohawk obviously i know where the Makefile.PL is located :-)
17:55 davido___ They're both in inc/ within the Inline::CPP filestructure
17:55 davido___ See https://github.com/daoswald/Inline-CPP/tree/master/inc/ILCPPConfig
17:55 davido___ is your github mohawk?
17:55 davido___ sorry, i mean cpan
17:56 davido___ github is mohawk2
17:56 mohawk correct
17:56 mohawk no, ETJ on PAUSE/CPAN
17:57 davido___ Ok, I sent him a message.
17:58 davido___ I'll do my best to pay attention to the progress.
17:58 mohawk thanks, and thanks
17:58 davido___ weigh in on issues, etc.
17:58 mohawk you are a hero
18:02 davido___ sorry i haven't been around much.  my intentions are good though. lol
18:03 mohawk ha ha
18:09 davido___ Let me know if you haven't heard from tsee within a day.  I'll pester again.  He seems to be a pretty reasonable person though.
18:09 mohawk grin
18:10 mohawk what's the best method of comms to ping you?
18:11 davido___ daoswald@gmail.com
18:11 mohawk davido___, i'm also intending to update I::CPP in concordance with EUCG, obviously
18:11 mohawk i assume you'll see PRs there ok?
18:11 mohawk i'll email you now
18:11 davido___ Yeah, I'll get the PR's.
18:11 davido___ I should just put them through pretty quickly as long as they don't break Inline::Module.
18:12 davido___ lots of moving parts :)
18:13 mohawk i'm imagining dev releases for now
18:13 mohawk does Inline::Module dep on I::CPP?
18:14 davido___ Not exactly a dependency.
18:15 davido___ The distribution created using Inline::Module will have an authoring dependency on Inline::CPP if appropriate, and in those cases it's currently bundled along with eucg.  That can change someday.
18:15 mohawk ok
18:15 davido___ actually the cpan dist will have Inline::CPP and eucg bundled too, so that it can build out on the target system.
18:15 mohawk one assumes that dev-releasing ICPP and EUCG won't risk breaking that
18:15 davido___ and again, that can change someday.
18:16 davido___ yeah, dev releases shouldn't.  In fact, stable releases shouldn't.  But the easiest way to be sure is to do a test release with Alt::Math::Prime::FastSieve::Inline, or whatever it's called. :)
18:17 davido___ that's kind of the Inline::Module proof of concept for C++.
18:18 davido___ I replied to your email, by the way.
18:19 mohawk yup, just received!
18:20 mohawk which C++ compilers does it currently work on?
18:20 mohawk rehsack was talking about how my PR definitely couldn't work on HP-UX and aCC
18:24 davido___ Inline::CPP should work on gnu/g++, clang++, and MSVC++.  This should include on OS-X platforms, and both old and modern BSD's.
18:24 davido___ BSD systems made a shift a year ago from g++ to clang++, for example.  So there we have to actually look at the BSD version number to determine which compiler, if I recall.
18:25 davido___ OSX uses gnu/g++, but requires slightly different library names, I think.
18:25 davido___ MSVC++ is its own mess. :)
18:26 davido___ I don't think C++ XS extensions will ever be as portable as C, but great progress has been made.
18:27 mohawk i agree there will be portability issues
18:28 mohawk however, i don't see any fundamental reason why that (at least in terms of build options) can't be inhaled into EUCG
18:31 davido___ I agree.  All that sort of thing should be pulled into EUCG.
18:31 davido___ You know, the truth is that there are a finite number of possibilities for compiler and libraries....
18:32 davido___ One could move away from a logic-based approach to a lookup table approach, and put the table into a json document that can be edited.
18:33 davido___ Come up with five criteria that can be used to make the decision across all platforms, and then drop the answers into a lookup table, if that makes sense.
18:34 mohawk i'm agnostic on that
18:35 mohawk my first priority will be to make it "just work", without needing any spin-control after
18:35 davido___ Yeah, that's most important.
18:35 mohawk how it does its thang inside, well, that's a separate thing
18:35 mohawk i like the idea of data-driven, but it may not be 100% possible
18:35 davido___ My thought was that if we can narrow down the sources of information needed to make decisions, and then separate concerns, adding platforms becomes easier to maintain.
18:36 davido___ anyway, got to run again. :( haha
18:39 mohawk enjoy!
19:58 ingy davido___: note we added a (bundle => 0) option if you don't want to bundle EUCG etc
19:59 ingy though I'm not a fan of it
19:59 ingy well I'm not against it either
20:00 ingy tmtowtdi
20:16 willthechill joined #inline

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