Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2016-12-23

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

All times shown according to UTC.

Time Nick Message
00:22 travis-ci perl11/cperl#2043 (smoke/master - 2de4b72 : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/186220583
10:51 travis-ci perl11/cperl#2044 (smoke/master - 6d9e459 : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/186303460
10:53 willthechill joined #perl11
10:56 willthechill joined #perl11
12:12 anton joined #perl11
18:20 bpmedley joined #perl11
19:02 willthechill joined #perl11
21:47 mako93r joined #perl11
21:48 mako93r Yay, even though I cheated, have a on this: https://ci.appveyor.com/project/93r/rperl-g7jdg/build/1.0.58
21:49 mako93r Had to turn off diff checking. This takes a longer time of debugging.
21:49 mako93r willthechill: You there?
21:49 willthechill yes I'm here
21:49 willthechill the diff checking is where astyle comes into play
21:50 mako93r Yes
21:50 willthechill wow great job!
21:50 mako93r the problem is the version of astyle under appveyor and under my local system are exactly the same
21:51 willthechill okay what does that mean?
21:53 mako93r I hardly can tell you what the issue is. This needs longer and more intensive debugging. For the moment it's only meaning appveyor can be made running successfully.
21:55 willthechill and you said you logged into the appveyor vm to run astyle comman directly?
21:58 mako93r I installed astyle via 'choco install astyle' and did a 'astyle --version' afterwards. I got a positive response with msg 'Artistic Style Version 2.05.1'
21:58 mako93r No direct login to appveyor needed.
21:59 willthechill okay then I believe the next step is to log into appveyor directly and run the astyle command to see why it is producing different output
22:05 mako93r This is one option, but I noticed something else. Just an example: Operator01NamedAtan2 (one I implemented) has this line in its 'cpp.CPPOPS_CPPTYPES' file:
22:05 mako93r #ifndef __CPP__INCLUDED__RPerl__Test__Operator01NamedAtan2__program_00_good_cpp
22:06 willthechill okay
22:06 willthechill yes those ifdef codes are necessary for proper compilation
22:06 willthechill can't be changed
22:06 mako93r Under my system without setting PERL5LIB to "blib;lib" it looks like this:
22:07 mako93r #ifndef __CPP__INCLUDED__C__Strawberry__Perl__Site__Lib__RPerl__Test__Operator01NamedAtan2__program_00_good_cpp
22:08 mako93r something like this. looks like path issues
22:08 willthechill hmm okay
22:08 willthechill yes that could be the problem!
22:09 willthechill that would cause astyle to format differently and also cause the diff check to fail even if astyle did not format differently
22:09 mako93r I could solve this problem by setting env var PERL5LIB to "blib;lib" before I did a build via the usual "perl Makefile.PL; dmake; dmake test; dmake install"
22:10 mako93r so the paths changed
22:10 willthechill okay that should be acceptable?
22:10 willthechill bulk88: are you here?
22:10 mako93r so therefore the '#ifndef'
22:11 willthechill bulk88 is the person who would know best about changing the dmake command and windows env vars, etc.
22:11 mako93r Yes, if it's time we need a cowboy/expert, bulk88 is the one.
22:13 mako93r BTW: when talking about appveyor: does your appveyor config allow the use of an appveyor.yml file or does it ignore it?
22:14 mako93r would be nice to know
22:16 willthechill mako93r: go ahead and make the PERL5LIB change to see if it solves the path problem
22:16 willthechill unfortunately, I have no idea about the appveyor.yml file
22:16 willthechill is there some way I can check that for you?
22:17 mako93r Yes, you can. Just connect to 'ci.appveyor.com'. You can login via your github account.
22:18 willthechill okay now what?
22:20 willthechill mako93r: here is our current appveyor test script
22:20 mako93r now you should a list a projects. RPerl should be listed
22:20 willthechill choco install strawberryperl --forceX86 && set PATH=C:\strawberry\c\bin;C:\strawberry\perl\site\bin;C:\strawberry\perl\bin;C:\windows\system32;C:\windows; && gcc -v && g++ -v && perl -V && dmake -V || perl Makefile.PL && perl -e"require CPAN;CPAN::Shell->notest('install','App::cpanminus');" && cpanm -n -v --installdeps . && set RPERL_DEBUG=1 && set RPERL_VERBOSE=1 && dmake test
22:20 willthechill yes I'm looking at the rperl project settings now, that's where I got that test script command I just pasted
22:21 mako93r fo
22:22 willthechill fo?
22:22 mako93r from 'settings' > 'environment' > 'install script'?
22:22 mako93r last one was a typo
22:25 mako93r What I really want to know is this: Is the setting 'settings' > 'general' > 'Ignore appveyor.yml' enabled or disabled?
22:27 willthechill ignore is NOT checked
22:27 willthechill so it will presumably utilize the appveyor.yml file
22:27 willthechill and no I do not believe we already have one
22:28 willthechill no appveyor.yml AFAIK
22:28 willthechill the paste above is from settings -> test -> script -> cmd
22:29 willthechill I assumed that you already knew the same build command we were using from the very beginning, the one I pasted above?
22:32 mako93r It's really not same anymore. But if this option 'Ignore appveyor.yml' is unchecked that's all I need to know and the best of all you don't need to do anything else.
22:32 mako93r I can come up with an 'appveyor.yml'
22:33 mako93r and your appveyor config should take care
22:33 mako93r that's what I wanted to know
22:34 mako93r don't bother
22:36 mako93r if your appveyor config pays attention to 'appveyor.yml' it could be an easy job for me
22:42 mako93r Made pull request. Let's see what will get.
22:43 mako93r Lean back and relax. :-)
22:49 bulk88 mako93r appveyor doesn't recognize appveyor.yml unless it is a github integrated appveyor project
22:49 bulk88 there is a difference between a git and a github project on appveyor
22:50 bulk88 github projects require you give admin perms on GH to appveyor
22:51 willthechill I use my github account to authenticate when I log in to appveyor
22:51 mako93r bulk88: that should be the case for RPerl
22:51 willthechill and appveyor is registered as a webhook in my github account under the rperl repo
22:51 willthechill so I'm pretty sure that, at least in MY github account, appveyor does indeed count as "github integrated"
22:51 willthechill if not, please correct me
22:52 mako93r willthechill: I made the same experience under my account of github
22:52 willthechill okay
22:52 willthechill mako93r: are you planning to create an appveyor.yml file which will replace the long build command I pasted above?
22:53 mako93r willthechill: so it looks like you're right
22:53 mako93r willthechill: yes, i'm planning excatly this
22:53 willthechill okay great
22:53 willthechill and what am I right about?
22:55 mako93r about appveyor config related to your github appveyor webhooks
22:55 mako93r looks good at moment
22:56 mako93r they are running
22:56 willthechill okay great, so do I need to do anything right this moment?
22:57 mako93r just lean back and relax, nothing else ;-)
22:57 willthechill haha not much relaxing, but yes that sounds great.  :-)
22:57 willthechill I am working on oath2 right now, yay!  :-P
22:57 mako93r may take some time. 15-25 minutes
22:58 mako93r we will see
22:58 willthechill so you are trying it now with the attempted PERL5LIB fix?
22:58 willthechill with the diff check tests re-enabled?
22:59 mako93r with PERL5LIB fix but with diff checks disabled
22:59 willthechill oh well how will you know if PERL5LIB fix does anything then?
23:02 mako93r experience and experiment. I tried it with it and without it. It made a difference at least on my local system.
23:02 willthechill ah okay then that's fine
23:05 mako93r but that hasn't to mean anything
23:05 mako93r we will see
23:05 mako93r tests still running
23:05 mako93r 't/13' remaining
23:06 willthechill cool
23:06 mako93r Cross finger! :-D
23:15 mako93r Yeeeehaawww! Looks very good. Travis's not quite finished. But anyway, my promise fulfilled (Even though I cheated a little bit)
23:15 willthechill okay great so what does that mean?
23:17 mako93r for you? Nothing! for me? I just proved that it is possible to make RPerl running under appveyor
23:17 mako93r but there's still a lot of debugging to do
23:17 willthechill okay sorry I'm just not following along very well I guess
23:18 willthechill what is it that you just now did, specifically?
23:18 willthechill I mean, what did you do different than before, which makes you happy now?
23:19 mako93r the main thing was that I changed appveyor's config (that alone made some difference)
23:20 mako93r and changed 't/13' so it skips reference files
23:20 mako93r not the best thing
23:21 mako93r but I promised (somehow) to come up with something that works somehow
23:21 willthechill okay so now the next step is enable astyle diff check?
23:24 willthechill I mean, isn't it possible that astyle diff check is now fixed thanks to changing PERL5LIB?
23:26 mako93r I need to check more of RPerl's use of 'astyle'. Is 'RPerl/Generator.pm' the only place in RPerl's codebase where 'astyle' is via perl's 'system()' function.
23:26 mako93r *is called* I mean
23:28 mako93r Worked: https://github.com/wbraswell/rperl/pull/72
23:28 willthechill nice!  :-)
23:28 mako93r Is this the first appveyor run working without failures?
23:28 willthechill astyle is called in Generator.pm and also Compiler.pm
23:29 willthechill working without failures?  well I think bulk88 and I turned off about half the tests one time and it worked...
23:29 willthechill nobody has ever got it to run with 100% of tests enabled
23:30 willthechill and even now the GMP tests are disabled because we need to figure out how to install it properly
23:31 mako93r But anyway: Did I do something right, at least partially?
23:31 willthechill well I can only assume so!
23:31 willthechill you understand more about what is happening here than I do!
23:31 willthechill that is why I keep asking what is happening
23:32 willthechill I don't understand how the windows stuff really works, but I trust in your abilities, haha!  :-)
23:33 mako93r Maybe!!! That's the curse of science. You never for sure. There's no 100%. Network Rule #8 says: "It is always more complicated than you think".
23:34 willthechill well any progress is good progress at this point
23:34 mako93r Have a look at RFC 1925 and you what I'm talking about.
23:34 willthechill haha okay maybe later, deep in oauth2 right now!
23:36 mako93r I'm leaving out words in my messages. That means I'm too much in party mode. Some bottle of beer too much. :-D BTW: Is my pull request of any use for you?
23:37 willthechill checking now
23:40 willthechill merged!
23:41 willthechill it is certainly the highest number of tests passed on appveyor so far!  :-)
23:41 willthechill great job!
23:41 willthechill mako93r++
23:42 willthechill let's just make sure we finish the job and remove that code from Generator.pm ASAP
23:47 mako93r Absolutely, let's see how this stuff behaves when future commits are coming
23:47 mako93r there's still a lot of work to do
23:48 willthechill very true, very true!
23:49 mako93r but anyway, I fulfilled my promise more or less to come up with a working appveyor (it just has to be considered experimental). even though friday is over
23:50 willthechill okay great job!  :-D
23:51 mako93r Yeehaw! And now back to my usual operator development. Do you plan a release for christmas?
23:53 willthechill yes of course, we shall have a Christmas and also New  Years release!  :-)

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