Perl 6 - the future is here, just unevenly distributed

IRC log for #inline, 2014-11-27

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

All times shown according to UTC.

Time Nick Message
00:40 ingy leont: ping
00:40 leont pong
00:40 ingy hi
00:40 ingy I was just wondering something
00:41 ingy is there a module that will emulate having a clean perl (no non-core mods) except for your prereq mods?
00:41 ingy if not I'll write one now :)
00:42 leont /achieve?
00:42 leont I think not, and it sounds rather dirty
00:44 mohawk ingy, you could insert coderef into @INC which used Corelist
00:44 ingy mohawk: correct
00:45 leont Except it would fail over to the rest of @INC if not found, so you need to take that into account too
00:45 ingy well I at least want this for ZD release testing
00:45 ingy @INC=sub{…}
00:46 leont Probably the easiest way is to use %Config to set @INC to only include core dirs
00:46 mohawk leont, shurely it wouldn't fail over if you did "die"?
00:47 leont I said "need to take that into account" ;-)
00:47 ingy or replaced @INC entirely
00:47 ingy anyway, thx
00:47 mohawk it's not rocket surgery
00:47 ingy :)
00:47 leont @INC =(@Config{'installarchlib', 'installprivlib'}
00:47 leont );
00:48 ingy the more important part is allowing the prereqs (and their prereqs etc)
00:49 leont O right
00:49 mohawk to the metacpan, robin!
00:49 leont Yeah, guess you'll need a sub for that
00:49 leont Still don't get why you'd want this
00:50 ingy to do simple clean perl testing
00:51 mohawk dependency-checking mostly?
00:52 ingy more or less
00:58 ingy hahaha
00:58 ingy I found the perfect name
00:58 ingy that I already used
00:58 ingy for something useless
00:58 ingy in 2007
00:58 ingy inc.pm
00:59 mohawk use inc 'coreonly'?
00:59 mohawk use inc 'corelist'
00:59 ingy export PERL5OPT='-Minc=core,dist.ini'
01:00 ingy mohawk: same thing really
01:00 ingy except I'll support named grouplets
01:01 ingy and I could make it plugin
01:01 ingy not sure about that
01:01 mohawk ingy, did i mention i'm making EUMM handle multi *.xs under lib/?
01:01 ingy I zip something up and we can piss on it
01:01 mohawk seems like that might be relevant for inline-module
01:02 ingy I think it would be great to start seeing .xs files under lib
01:02 mohawk watch this space
01:02 ingy instead of XS.xs
01:02 mohawk it will involve some tricky tinkering by the user
01:03 mohawk they'll need to set MULTIXS=>1 in WriteMakefile
01:03 ingy Inline::Module punts all the work to Inline::C
01:04 ingy it has no concept of XS
01:04 mohawk indeed
01:04 mohawk well, just mentioning it
01:04 ingy er, no knowledge
01:04 ingy yeah I think it's great
01:05 mohawk it's been relatively easy
01:05 mohawk i think everyone's frightened of going inside EUMM
01:05 ingy I'm also a fan of having the source files install
01:05 ingy nod
01:06 mohawk and having lived inside it, i understand that fear ;-)
01:06 ingy schwern made a career of scaring people away
01:06 ingy I talked to schwern the other day
01:07 ingy he's hacking core evidently
01:07 * mohawk remains silent ;-)
01:07 ingy I wouldn't know (or care)
01:07 ingy p5p is not my rathole
01:08 mohawk nor mine really
01:09 mohawk it's fair to say EUMM is big (nearly 15k lines in lib/ExtUtils/* in the distro)
01:10 mohawk and a fairly complex bit of software
01:10 mohawk with lots, and lots, of technical debt
01:13 ingy well I'm glad you care for it
01:13 ingy don't let it eat you alive
01:14 mohawk ha ha
01:14 mohawk i'm definitely in charge of this thing
01:14 mohawk 18k lines under lib
01:14 mohawk yowser
01:15 mohawk i'm going to wrangle it in due course to be more data-structure orientated rather than "Makefile text fragments"
01:15 mohawk that's an adventure for another day
01:59 GitHub128 [inline-c-pm] wphillipmoore comment on issue #20: I'm building Inline-C 0.66 (haven't tried with 0.67 yet, but those tests... http://git.io/DExMOA
02:02 leont The thing you have to understand about eumm is that it's the older OO project in Perl
02:02 leont Some choices were unfortunate, but they really couldn't have known better
02:02 leont s/older/oldest/
02:03 leont It has lots of technical dept, but that's usually not the biggest challenge
02:03 leont It's usually that it has to be portable on systems you'd rather not want to think about
02:04 leont Lots of edge cases and stuff
03:21 mohawk leont, you should know by now i'm not judging the previous engineers who worked on EUMM
03:22 mohawk and remember i've fought the portability battle myself with the "perl in space" stuff ;-)
03:31 * ingy has evil thoughts
03:34 mohawk that's not news
04:06 ingy I figured out how to use `inc` locally
04:06 ingy at least one way
04:06 ingy not sure it's important, but wanted to at least know I could
04:08 ingy not even evil…
04:10 ingy use inc(); sub scope { local @INC = inc->inc(…); … }
05:01 ingy hrm, that's not useful, since `use` statements would execute at compile time
06:25 chansen joined #inline
07:27 ingy mohawk, leont: inc-0.02.tar.gz Released
07:27 ingy https://github.com/ingydotnet/inc-pm#readme
07:28 ingy it's just the basic extendable framework
07:28 ingy but I think it's pretty awesome usage-wise
10:33 mohawk neat idea

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