Perl 6 - the future is here, just unevenly distributed

IRC log for #pogl, 2017-02-19

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

All times shown according to UTC.

Time Nick Message
12:43 chm joined #pogl
15:46 chm Mithaldu: I think I've got the fix for the GLEW_VERSION_M_N and others.  Will push a devel release this PM.
15:46 chm Mithaldu: Where do we stand with working code for MacOSX?
15:47 Mithaldu aight, gonna have a look at that then
15:47 Mithaldu as for macos, fucked
15:48 Mithaldu i got xvfb to run on travis, but it does not provide a gl context
15:48 Mithaldu and i have no idea how we would get it to
15:48 Mithaldu otoh
15:48 chm Mithaldu: what about just compiling for a user?  The Travis-CI can come later...
15:48 Mithaldu a friend of CandyAngel confirmed the whole thing runs and tests on osx without skipping anything
15:49 chm Great to hear!  Maybe this quick developers release can be the next official release very soon...
15:53 Mithaldu you could ping me once you have the code up so i can have a look before you push it to cpan
15:54 Mithaldu also, any thoughts on the test.pl thing yet?
15:54 Mithaldu can understand if not, just a little curious, as there's a bunch of stuff in there that i don't quite understand
15:55 Mithaldu also some functions that when i tried to use the ::Modern variant of, with the same calling convention, just plain broke
15:55 chm As I mentioned, it was never a good test, just a demo that would run.
15:55 Mithaldu started returning 0 instead of something
15:56 chm The other downside is that it depends on lots of non-OpenGL::Modern modules
15:56 chm such as OpenGL::Array, OpenGL::RPN, OpenGL::Matrix, OpenGL::Image, and OpenGL::Shader (I think)
15:57 Mithaldu that honestly makes it perfectly suited for what we need
15:57 Mithaldu that is you can use it as a proving grounds for either replacing those modules with ::Modern variants, or excizing them
15:57 chm That doesn't make it suited as a test for OpenGL::Modern since some things are going to be different.
15:58 Mithaldu well no
15:58 Mithaldu that's the point
15:58 Mithaldu things will be different
15:58 chm If you want a litmus test for compatible wrapper routines, then it would be fine.
15:58 Mithaldu and it's easier to decide what should be different and how if you're looking at working code
16:01 chm Mithaldu: I just tried making a distribution and the result crashed on build with an error about cpanfile
16:01 chm Any idea what is wrong?  Where it is being called?  How to fix?
16:01 Mithaldu paste?
16:01 Mithaldu also do you have use ExtUtils::MakeMaker::CPANfile; ?
16:02 chm I have EMC
16:02 chm cpanfile is not available: /cygdrive/e/chm/pogl/OpenGL-Modern-0.02_01/cpanfile: No such file or directory at /home/chm/local64/lib/perl5/Module/CPANfile.pm line 47.
16:03 Mithaldu right, then i need your output
16:03 Mithaldu err, huh
16:03 Mithaldu that's odd
16:03 Mithaldu can you add Devel::Confess to mfPL?
16:03 chm That is the error.
16:04 chm I notice that cpanfile is not in the MANIFEST but I can't see anywhere in the code that calls it
16:04 Mithaldu it shouldn't be called
16:04 Mithaldu emc should read it from .
16:04 Mithaldu oh
16:04 Mithaldu wait
16:04 chm well something is using it
16:04 Mithaldu if it's not in the manifest then you're talking about inside the build
16:04 Mithaldu add it to the manifest
16:05 chm why wasn't it in the MANIFEST already?
16:05 Mithaldu brainfart
16:05 chm Let me try...
16:06 chm That seems to have been the problem.
16:06 chm Thanks
16:06 Mithaldu cheers
16:06 Mithaldu for explanation
16:07 Mithaldu cpanfile is the sole source of deps for the entire thing
16:07 chm In the future, I'm fine if you want to add things to OpenGL::Modern for the build,
16:07 Mithaldu that means if anything changes, edit that one
16:07 chm please make the changes to MANIFEST if needed.  I have cruft in my directory and
16:07 Mithaldu and mfpl will use the relevant modules to load that when doing anything, like generating the mymeta files
16:07 Mithaldu yeah sure
16:07 Mithaldu as i said, brainfart
16:07 chm usually ignore the distcheck garbage
16:11 chm Mithaldu: One thing I noticed is that glewCreateContext has an error about only OpenGL 1 0 or something
16:11 Mithaldu where?
16:12 chm I haven't verified it but did see an error return.  We should be able to check this with
16:12 chm the fixed variable accessors.
16:12 Mithaldu i asked where because i don't know in what context you saw that error
16:13 chm I'm not sure where I saw it.
16:14 Mithaldu did you see it in some output or in the code itself?
16:14 Mithaldu (that's how vague the statement was to me ;))
16:15 chm The statement was a vague as my knowledge unfortunately, :-(
16:15 Mithaldu haha, well, fine to ignore then, until you remember how you saw it :)
16:15 Mithaldu last time i checked the createcontext functions all swallow errors
16:16 Mithaldu kinda rude really
16:18 chm Yes, I don't really trust them.  The code is confusing as well.
16:18 chm I think your approach is safer and clearer to follow.
16:20 Mithaldu you mean the moral equivalent sample i gave in the issue about teasing out their errors?
16:21 chm I mean something with explicit and documented functionality rather than guesses from internal code in GLEW
16:23 Mithaldu well, in general that sounds like somehting i'd agree with
16:24 Mithaldu otoh there is the issue that glew abstracts away a LOT of crossplatform stuff in the createcontext call
16:24 Mithaldu to the point where i think it's safe to say if we don't get a context, the issue is at the level of "device or driver missing"
16:24 chm Well, 0.02_01 passes for cygwin/X11 and Strawberry Perl Portable.  I'm pushing to PAUSE now...
16:26 Mithaldu fwiw
16:26 Mithaldu might be good to rename some of the list thingers now that their purpose is becoming more clear
16:26 chm Assuming it tests well, we can plan on an official release by Monday PM.
16:26 Mithaldu lib/OpenGL/Modern/NameLists/MakefileAll.pm seems to be constants?
16:26 chm What is a list thingers?
16:26 Mithaldu lib/OpenGL/Modern/NameLists/Modern.pm is functions?
16:27 chm Seems reasonable but maybe wait a few releases.  Since they are internal, I don't think that will hurt things.
16:27 Ryoga joined #pogl
16:27 Mithaldu sure thing, i don't plan to touch it without input
16:27 Mithaldu just recommending things due to the whole "naming is hard" thing
16:27 Mithaldu hey Ryoga
16:28 Mithaldu also the approach of "just make 'em functions" looks ok to me
16:28 Ryoga Hi  Mithaldu
16:30 chm Mithaldu: Nice to see the Travis-CI confirmation when I pushed to master.  Thanks again.
16:30 Mithaldu most of that was corion
16:30 CandyAngel Heyas Ryoga *waves*
16:40 chm OpenGL::Modern 0.02_01 is out, https://sourceforge.net/p/pogl/mailman/message/35675454/
16:52 CandyAngel \o/ Yay!
17:01 Mithaldu chm: typo :) skipj
22:21 Ryoga joined #pogl

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