Perl 6 - the future is here, just unevenly distributed

IRC log for #pr-challenge, 2015-01-29

| Channels | #pr-challenge index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:00 boreas joined #pr-challenge
00:29 bayashi joined #pr-challenge
00:55 boreas joined #pr-challenge
00:55 BRAD_ joined #pr-challenge
01:18 deven2 joined #pr-challenge
01:32 vairav joined #pr-challenge
01:42 cakirke joined #pr-challenge
01:49 deven_ joined #pr-challenge
01:52 punter joined #pr-challenge
01:57 boreas joined #pr-challenge
02:04 tbushell joined #pr-challenge
02:04 Su-Shee_ joined #pr-challenge
02:17 deven_ So a question for everyone -- any tips on converting Any::Moose code to Moo?
02:18 haarg it's mostly the same as converting something from Moose to Moo
02:19 haarg https://metacpan.org/pod/Moo#INCOMPATIBILITIES-WITH-MOOSE this will likely help
02:20 haarg at some point we should get a porting guide written, since it does come up.  and a porting guide would likely highlight the differences in an easier to read format even if you weren't actually porting anything.
02:20 haarg deven_: is there a specific module you are looking at?
02:20 Peter My first step would be to remove 7 characters and pray
02:20 boreas joined #pr-challenge
02:22 deven_ haarg: Yes, my assigned module is Hailo, and ether filed an issue report to convert Any::Moose to Moo.  I figured that would be a good PR to work on.
02:23 deven_ I've seen Moo and Moose code but have never gotten around to coding for either myself.
02:25 haarg that's going to be a somewhat tricky conversion, because it uses MouseX::Getopt
02:25 Peter deven_, If you need a replacement for Moose's types, try Type::Tiny
02:26 haarg often, the changes needed between Mo[ou]se and Moo are syntax level things
02:26 Peter A quick search gives this https://metacpan.org/pod/MooX::Options
02:26 haarg but Getopt doesn't have a direct equivalent - there are some MooX modules in the same problem space, but it will take a bit more work to translate
02:27 deven_ haarg: Is there a Moo equivalent to MouseX::Getopt?
02:27 deven_ Right now I'm still trying to get enough of a development environment setup on a fresh Fedora 21 install (my normal server is in flux due to hard drive issues)
02:28 haarg MooX::Options would be the first thing to look at.  like i say though, it's not a direct equivalent.
02:28 deven_ I still haven't gotten to the point of even being able to build and test the module as-is. :(
02:29 deven_ The github repo doesn't have a Makefile, but there's a dist.ini -- I assume I need to do a "dzil build" for it?
02:29 haarg yes
02:30 haarg although often 'prove -lr' will work for testing
02:31 deven_ At this point, I've cloned the github repo and I've been trying to get my environment setup.  Won't prove just try to run the tests?
02:31 Peter yep
02:33 deven_ I'm assuming the libs won't be ready to test with until running the build process.
02:33 kid51 joined #pr-challenge
02:34 haarg most of the time the tests can be run directly against what is in the repo without doing a build
02:34 deven_ hmm. Class::MOP::load_class is deprecated
02:37 doby joined #pr-challenge
02:38 deven_ updated MooseX-Role-Parameterized, maybe that will help
02:47 deven_ Now the "dzil build" command seems to run and generates a build in a "Hailo-0.73" directory.  I assume this was automatic?  The last version on CPAN is 0.72
02:49 deven_ When I went there and ran "perl Makefile.PL" and "make" and "make test", it seems to build okay, and tries to run the tests, but those keep failing from missing modules, which I'm still installing.
02:49 deven_ Should I use "cpanm Hailo" to make sure all the prereqs are installed?
02:51 deven_ haarg: Good to know. (running against repo)
02:52 haarg cpanm --installdeps Hailo
02:52 haarg or dzil listdeps | cpanm
02:53 deven_ haarg: oh, cool
02:53 Peter deven_, should also be able to do "dzil test"
02:54 deven_ Peter: After "dzil build" or instead of?
02:54 Peter to test
02:54 deven_ I really need to learn Dist::Zilla, but I only have one CPAN module to my name and not much practice. :)
02:54 Peter You don't need to manually build first
02:55 deven_ Peter: Does "dzil test" imply "dzil build" like "make test" often does?
02:55 veryrusty joined #pr-challenge
02:55 Peter Kind of, it uses a temporary build directory iirc
02:57 Peter deven_, test will not build a dist
03:02 itcharlie for purpose of pr-challenge, I usually just `cpanm -L $dir_name $some_module_name`. It installed the module with all its dependencies in a directory.  Then I run ` export PERL5LIB=$dir_name:$PERL5LIB`   to include as one of the Perl 5 Libraries.
03:07 deven_ itcharlie: nice trick
03:07 itcharlie Thanks
03:08 deven_ Looks like I finally have the tests working with the prereqs installed now.  Took long enough!
03:10 deven_ It claims "All test successful" but there's a few questionable ones.
03:10 deven_ t/hailo/non_standard_plugin.t ............... skipped: A `Issuing rollback() due to DESTROY without explicit disconnect()' bug in Hailo
03:10 deven_ t/hailo/real_workload.t ..................... skipped: You need BRAIN=... with a path to a Hailo brain and LOG= with an irc log
03:11 deven_ t/storage/SQLite/file-exhaustive-all.t ...... skipped: Set TEST_EXHAUSTIVE_ALL= to run all exhaustive tests, these take a lot of time
03:11 boreas joined #pr-challenge
03:12 deven_ Not sure if I should worry about those.  The non_standard_plugin one looks like it might be a real bug needing some attention.  The real_workload one presumably involves building a real Markov chain database of some sort.
03:13 deven_ I'm trying the exhaustive ones -- 25,342 tests!  Sounds exhaustive indeed!
03:20 deven_ Enabling those test make the test runtime go from 1:41 to 6:27
03:23 deven_ Okay, so with this using Any::Moose, it's running the test suite twice, once with ANY_MOOSE=Mouse and again with ANY_MOOSE=Moose -- for obvious reasons.
03:25 deven_ haarg: Should I just change "use Any::Moose" to "use Moo" and see how bad the breakage is with that alone?
03:25 deven_ (obviously it won't just work)
03:26 haarg i suspect it will break too much to give you any useful results
03:29 raj4perl joined #pr-challenge
03:29 deven_ haarg: Suggestions for a first pass?
03:31 haarg deven_: the isa checks should all be replaced with Type::Tiny
03:33 deven_ haarg: example?
03:33 haarg so instead of `has attr => (..., isa => 'Bool', ...);`, have `use Types::Standard qw(Bool); ...; has attr => (..., isa => Bool, ...);`
03:38 haarg some other things i see in the code that Moo doesn't support: lazy_build, override/super, auto_deref
03:39 haarg and Getopt is the remaining problem i see
03:40 deven_ haarg: Will it still be feasible to use Moo with those unsupported features?  Rewrite to not use them?
03:40 rvandam joined #pr-challenge
03:41 haarg yes
03:42 haarg lazy_build is a short form of specifying a number of options, so it can be easily replaced with the long form.  check the Moose docs for more info.
03:43 haarg override/super does roughly the same thing as an around, so you should be able to replace it with that
03:44 haarg auto_deref, you can replace with a private getter, the add a wrapper sub that will deref
04:37 deven_ use Any::Moose 'X::StrictConstructor';
04:38 deven_ haarg: Can that just become: use MooX::StrictConstructor I hope?
04:38 Peter deven_, I've had problems with that
04:38 haarg it would likely work, but if not you could start by just removing that part
04:38 Peter For some reason, in some configurations on cpan testers, it failed when I used it
04:39 Peter I would just use it and see what happens, I gave up trying to reproduce
04:39 deven_ Peter: Any idea what was causing the problem?
04:39 Peter I think it was something to do with the interplay of Moo and Moose, but not sure
04:40 Peter MooX::StrictConstructor does something less "clean" than Moose::... does iirc
04:40 deven_ haarg: Should I list types individually or just do use Types::Standard qw(:all);
04:40 haarg i generally list them.  it's up to you.
04:40 deven_ ok
04:41 haarg it's using autoclean, so they'll get cleaned up in the end
04:41 Peter Does Hailo make any of its own types?
04:42 haarg not that i saw
04:43 deven_ default => sub { 0 } or default => sub {0}
04:44 deven_ anyone have stong opinion either way?
04:44 deven_ I'm leaning toward the spaces
04:45 haarg no opinion.  default => 0 works too
04:46 Peter haarg, that works in Moo now?
04:46 haarg yes
04:46 Peter Back when I last used it that was supposed to work, but nothing but refs did :(
04:49 deven_ haarg: Is that a recent change?
04:49 haarg default => $non_ref has worked since 1.001000 (mar 2013), although there were some bugs that have been fixed with with false values (1.002000, may 2013) and undef (1.004005, may 2014)
04:50 deven_ haarg: I was wondering why Moo would only support the sub, because it seemed easy enough to support both, at a glance..
04:51 deven_ haarg: So I should probably mark Moo 1.004005 as the minimum required version?  Or the current version?
04:52 haarg if you are using default => undef, then 1.004005 is the minimum
04:53 deven_ Hmm, when I drop the sub {}, I get an error: Invalid default 'HASH(0x1ce38b0)' for Hailo->storage_args not a coderef or a non-ref or code-convertible object at /usr/local/share/perl5/Method/Generate/Accessor.pm line 658.
04:54 haarg a hash is a ref
04:54 haarg or rather a hashref is a ref
04:54 haarg you have to use the sub {...} form for refs
04:54 deven_ oh, ok
04:54 haarg otherwise every object would get a ref to the same hash
04:55 haarg that's why the sub {} form was the only supported form initially
04:56 deven_ haarg: To make sure they were newly-generated unshared refs?
04:56 haarg yes
04:56 deven_ ok, that makes sense.
04:56 deven_ Certainly it's nice to be able to skip the sub {} for ints and strings.
04:56 deven_ well, scalars. :)
04:57 haarg refs are scalars :)
04:57 Peter haarg, I may have been trying to set undef...
04:57 deven_ Okay, fine, you knew what I meant!
04:57 haarg hehe
04:57 deven_ Stickler!
04:58 Peter I might try and port my stuff back to Moo, see what happens
04:59 Peter I don't *need* any Moose stuff
04:59 haarg i often use scalar like that too, but i think it's worth being specific in this case
04:59 Peter but I don't like an interface where extra args are silently gobbled up
05:00 Peter and I wanted 0 fails :P
05:00 deven_ haarg: You were right to point it out.  non-ref, better? :)
05:00 haarg i'd be interested in seeing what issues you ran into
05:00 haarg deven_: yarp
05:01 haarg i have a branch sitting around somewhere to improve MooX::StrictConstructor a bit, but i have no idea if it would help whatever case you ran into
05:01 Peter Ah, didn't know you were the author :)
05:02 haarg of strictconstructor, i'm not
05:02 haarg there were just some edge cases i wanted to round off
05:02 Peter Ah, you are Moo's author :)
05:03 haarg maintainer
05:03 Peter Ah, wish metacpan showed that info more easily
05:03 haarg i think i have the most commits at this point, but i didn't write it initially
05:04 Peter Just checked, it was mst
05:04 deven_ haarg: Hmm, this gets a syntax error: override _train_fh => sub {
05:04 deven_ in Hailo/Command.pm
05:04 haarg i mentioned that already
05:05 deven_ Is that a missing feature?
05:05 Peter no override
05:05 haarg Moo does not have override/super
05:05 deven_ Sorry, I didn't internalize which were the missing features.  I can use around?
05:05 haarg you should be able to use around
05:05 deven_ how would it translate?
05:06 deven_ Can I just change the override keyword to around?  or more needed?
05:06 haarg more will be needed
05:06 haarg have you read the docs for around and override/super?
05:08 itcharlie left #pr-challenge
05:08 deven_ looking now :)
05:11 deven_ haarg: Have I got this right?  Change "override" to "around" and then replace "super()" with "SUPER::_train_fh()"?
05:12 deven_ Actually, I probably need to pass the same args too, right?
05:12 haarg no, you won't use SUPER::
05:13 deven_ Oh, I need this $orig
05:14 deven_ $orig->($self, $fh, $fast, $filename);
05:14 deven_ sound right?
05:15 haarg basically, yeah
05:15 haarg i usually spell it $self->$orig($arg1, ...)
05:15 deven_ haarg:oh, that looks better
05:16 haarg but it does the same thing
05:16 Peter haarg, this is the sort of error http://www.cpantesters.org/cpan/report/3876f228-2ecc-11e4-8637-e0d24bc0d69e
05:16 deven_ ok, I changed it: $self->$orig($fh, $fast, $filename);
05:16 Peter at the beginning of the failures
05:16 Peter It is rejecting a valid param
05:17 Peter I never managed to reproduce it though
05:17 Peter But plenty failed on many setups
05:17 Peter http://www.cpantesters.org/distro/C/CloudFlare-Client.html?oncpan=1&distmat=1&version=0.03_6-TRIAL
05:18 Peter And oddly it only affected one class in that dist, when iirc all of them are used in tests and all used MooX::StrictConstructor
05:29 deven_ haarg: I don't need __PACKAGE__->meta->make_immutable with Moo, right?
05:29 haarg correct
05:29 deven_ cool
05:33 deven_ haarg: use Any::Moose '::Role'; changes to use Moo::Role?
05:33 haarg yep
05:33 Peter deven_, I hope you're going to blog about this with all the details after :P
05:34 Peter Could be a useful guide
05:34 deven_ Peter: Well, I don't have a blog yet!  But I hope to. :)
05:34 Peter :)
05:34 haarg Peter: that test report is odd
05:35 haarg i can't reproduce the issue either
05:35 deven_ haarg: I should still use Moo before use Moo::Role, no?
05:35 haarg no
05:35 Peter haarg, I've just noticed I still am using Moo in some of my tests :D
05:35 haarg and i don't see any obvious problems in the code in your module or strictconstructor that would lead to those errors
05:35 Peter s/tests/classes/
05:36 deven_ haarg: oh, ok!
05:37 Peter haarg, I have no clue, maybe it could be due to using Kavorka?
05:37 Peter it is the only "weird" module in that class
05:38 Peter And it isn't in the other classes
05:39 haarg Peter: that's it
05:40 haarg i downgraded kavorka and can reproduce the problem
05:40 Peter Ah, do you know which version it starts blowing up on?
05:41 haarg i only tested 0.033 because that's what is listed in the test report
05:41 Peter Ah, thanks so much for testing
05:42 Peter I can use Moo again!
05:42 deven_ haarg: isa => "Hailo"
05:42 deven_ Do I just drop the quotes?
05:43 haarg isa => InstanceOf["Hailo"]
05:43 haarg from Types::Standard
05:43 deven_ same for isa => DBI::db
05:43 deven_ ?
05:43 haarg yep
05:43 deven_ kk
05:44 Peter haarg, probably this is the cause https://rt.cpan.org/Public/Bug/Display.html?id=98666
05:45 haarg Peter: yes, i expect so
05:45 haarg newer Kavorka has a workaround but hopefully it can be updated to use the official api at some point
05:46 Peter Yes, I see the recent email in there from you :)
05:47 deven_ my $hailo = $self->hailo;
05:47 deven_ $hailo->learn($string);
05:47 deven_ haarg: Can't call method "learn" on an undefined value at t/lib/Hailo/Test.pm line 268.
05:47 deven_ Any guess why $self->hailo is undef now when it wasn't before I started making changes?
05:48 haarg Moo doesn't support lazy_build
05:48 haarg i'm guessing you didn't convert that
05:49 deven_ haarg: I missed that of course.  I knew there was something else!
05:49 deven_ and the auto_deref
05:54 srcorzine joined #pr-challenge
05:56 srcorzine left #pr-challenge
05:57 haarg deven_: http://paste.scsys.co.uk/460124 something like this should work for auto_deref
05:57 haarg change from @$out to %$out if using it on a hashref attribute
05:58 deven_ haarg: and for lazy_build?
05:58 haarg look up the Moose docs for it
05:58 Peter deven_, lazy_build is just a shortcut for using two options
05:58 haarg it will show what it is a shortcut for
05:59 deven_ haarg: found it: http://search.cpan.org/~stevan/Moose-0.46/lib/Moose/Cookbook/Recipe9.pod#The_lazy_build_shortcut
06:01 deven_ Peter: Looks to be a shortcut for 5 options
06:02 Peter my memory sucks, sorry
06:02 deven_ Peter: np :)
06:02 deven_ haarg: Why not add the lazy_build shortcut to Moo for convenience?
06:03 haarg because it's usually a bad idea
06:03 deven_ how so?
06:03 haarg you usually don't want the predicate and clearer
06:03 deven_ why so?
06:03 Peter And that's why I thought it was two options
06:04 haarg most people only ever use or think about the lazy and builder parts of it
06:04 Peter Required should be implicit :P
06:04 haarg Moo supports is => 'lazy' as a shortcut for is => 'ro', lazy => 1, builder => "_build_$attr"
06:05 haarg which is generally a much more useful shortcut
06:05 Peter :)
06:06 haarg clearing an attribute is just usually not useful, and the extra mutability adds to the complexity of things
06:06 deven_ haarg: So for these lazy_build examples, maybe I should use is => 'lazy' instead?
06:07 haarg maybe
06:07 deven_ haarg: Does Moose also support is => 'lazy' or is that Moo-specific?
06:07 haarg but you'll also need to specify the predicate and clearer
06:07 haarg Moose does not, but MooseX::AttributeShortcuts does
06:07 deven_ I should see if the predicate and clearer are defined?
06:08 haarg lazy_build => 1 means they are defined
06:08 deven_ ah
06:09 deven_ so replace lazy_build => 1 with is => "lazy" and add require => 1 and the predicate and clearer?
06:09 Peter haarg, they might not be used
06:09 haarg Peter: maybe not internally
06:09 haarg but they are part of the api
06:10 Peter hmm, yeah, but surely they shouldn't really be used externally?
06:10 Peter unless documented as public
06:10 haarg probably not.  this is why lazy_build was a bad idea.
06:10 haarg deven_: lazy+required doesn't make much sense
06:11 Peter deven_ could ask ether maybe?
06:11 Peter other whoever it was
06:11 Peter too tired for memories :S
06:11 deven_ haarg: The link I found above shows required => 1 as part of the equivalence for lazy_build
06:12 deven_ I'm not sure I follow why, if there's a builder.
06:12 Peter having a builder surely implies required?
06:12 Peter or rather, makes it pointless
06:12 deven_ Peter: Looks like it does make it pointless.
06:13 deven_ haarg: I don't see any sign that the predicate and clearer are used.  Maybe try without them and see if it blows up?
06:13 Peter deven_, haarg means other modules could use them
06:13 Peter hmm, lets check the reverse deps
06:14 Peter only 4
06:14 haarg it doesn't really hurt to add them since they already exist
06:14 Peter deven_, yeah, listen to haarg, it is less effort :)
06:15 haarg try to keep things as close to the original before looking at removing things like that
06:15 haarg deven_: also here's a better link https://metacpan.org/pod/distribution/Moose/lib/Moose/Cookbook/Basics/BinaryTree_BuilderAndLazyBuild.pod#The-lazy_build-shortcut
06:15 haarg the link you posted is from 2008
06:18 deven_ haarg: hmm, ok.  That's what Google found. :)
06:18 haarg google is unfortunately usually a pretty poor source for perl information
06:18 deven_ unfortunate
06:25 deven_ so with has "_${k}" => { lazy_build => 1 ...}
06:25 deven_ is => "lazy", predicate => "has_${k}", clearer => "clear_${k}", right?
06:26 deven_ (I don't know why the heck he put braces around {k}
06:27 rvandam joined #pr-challenge
06:28 Peter Hmm, the attr starts with _
06:28 Peter I think the docs said that would make the pred and clearer _ too?
06:28 Peter "If your attribute starts with an underscore, Moose is smart and will do the right thing with the predicate and clearer, making them both start with an underscore. The builder method always starts with an underscore."
06:29 Peter Unless moo does that too.... in which case I'll show myself out
06:30 Peter I'm sure haarg knows :)
06:31 haarg you should be able to do predicate => 1, clearer => 1
06:32 haarg Moo will generate names compatible with lazy_build
06:32 Peter and build => 1 presumably too?
06:32 Peter *builder
06:33 haarg yes
06:34 haarg although if you use is => 'lazy', that is implied
06:59 sue joined #pr-challenge
07:11 boreas joined #pr-challenge
07:14 deven_ haarg: Oh, the predicate => 1 and clearer => 1 is nice, thanks!
07:16 reneeb joined #pr-challenge
07:29 deven_ haarg: I'm quitting for the night, but if you're interested to see the current state: https://github.com/deven/hailo/tree/moo
07:30 deven_ haarg: Here's the commit with all my changes so far: https://github.com/deven/hailo/commit/57faf778f58e57b533288d1f66d471fe1c7ae9e5
07:31 deven_ Many of the test cases are still failing, but I don't know why at this point.
07:31 tbushell joined #pr-challenge
07:33 deven_ night
07:49 reneeb joined #pr-challenge
08:15 vairav joined #pr-challenge
08:19 neilb joined #pr-challenge
08:48 veryrusty joined #pr-challenge
08:58 sawyer joined #pr-challenge
09:09 neilb joined #pr-challenge
09:11 Relequestual joined #pr-challenge
09:24 G_SabinoMullane joined #pr-challenge
09:34 tbushell joined #pr-challenge
10:10 raj4perl joined #pr-challenge
10:18 punter joined #pr-challenge
10:31 sue_ joined #pr-challenge
11:44 shorty_mu joined #pr-challenge
11:49 reneeb joined #pr-challenge
11:54 cakirke joined #pr-challenge
12:14 kid51 joined #pr-challenge
12:39 punter joined #pr-challenge
12:40 punter joined #pr-challenge
12:56 Relequestual joined #pr-challenge
13:07 boreas joined #pr-challenge
13:17 boreas joined #pr-challenge
13:51 raj4perl joined #pr-challenge
14:01 BooK wow, the cpan prc mailing list setup is weird: having the private subscription management link added at the end of every email is disturbing
14:01 BooK especially when it also uses the List-* headers
14:05 StylusEater joined #pr-challenge
14:13 marderh joined #pr-challenge
14:14 gryphon joined #pr-challenge
14:23 jnap joined #pr-challenge
14:25 Relequestual joined #pr-challenge
14:26 vroom joined #pr-challenge
14:29 neilb BooK: I’ll have a look at the config later — I might be able to jigger with that
14:38 BooK I'm just afraid someone might top post too hastily and leak some info
14:39 deven joined #pr-challenge
14:57 sue_ joined #pr-challenge
15:12 tinypig joined #pr-challenge
15:59 dolmen joined #pr-challenge
15:59 dolmen https://github.com/blog/1943-how-to-write-the-perfect-pull-request
16:07 vairav joined #pr-challenge
16:15 gryphon Just personal preference, but I hate emoji. They get in the way of the words.
16:16 gryphon Of course, I'm also a hypocrite. :)
16:16 shorty_mu See, I have no way with words. I think they're nice ;)
16:17 deven haarg: Thanks for catching that HashRef bug.  I'm sure there's probably more than that.  Did you try running the tests or were you just reading the commit diffs?
16:19 deven It was 2 AM when I was making that change, and didn't stop to contemplate what I was dereferencing. :)
16:24 Peter What does that emoji example mean?
16:25 Peter I mean, what is a sparkly star meant to mean
16:52 sahadev joined #pr-challenge
17:18 boreas joined #pr-challenge
17:20 deven Peter: What sparkly star?
17:39 rvandam joined #pr-challenge
17:40 G_SabinoMullane The one in the github blog above. "Use emoji to clarify tone"
17:40 G_SabinoMullane -1 to emoji in that context
17:40 manchicken_ G_SabinoMullane: You ever play the game Mass Effect?
17:41 G_SabinoMullane Nope
17:42 manchicken_ They have a race in that game, the Elcor, which communicate verbally using a ton of clarifying words to convey feelings and emotions since their natural way of conveying those things is through pheromones. That reminded me of the Elcor.
17:45 neilb joined #pr-challenge
17:52 vairav joined #pr-challenge
18:50 gryphon I use words to clarify tone. I'm weird that way, I guess.
18:51 thrig weightlift with books?
18:51 gryphon heh
19:06 fredericop joined #pr-challenge
19:30 mvuets joined #pr-challenge
19:39 Peter manchicken_, conjecture, communicating with Elcor is likely tedious, sarcastic undertones
20:02 mvuets neilb: what's the completeness rate?
20:03 neilb 36.6%
20:03 neilb I reckon this weekend will see a surge, of both completes and skips, with some quits as well
20:03 mvuets and do you know the acceptance rate?
20:03 mvuets i.e. how many of those 36.6 got actually merged?
20:03 neilb not yet, no. Thinking I’ll write a script for that in eaerly Feb
20:04 neilb when I’m planning to look at various stats, write things up, etc
20:04 * mvuets just being curious if YVES looked after PR he got (o:
20:05 mvuets neilb++, that sounds to be fanstastic!
20:06 mvuets I am sure in 12 month your ranking-assigning algorithm will be perfect (-:
20:07 neilb heh
20:07 mvuets By the way, talked today to someone who started considering to join the challenge. He might end up needing to catch up on January first. Right?
20:08 neilb nope, I’ve got several people lined up to start in february
20:08 neilb He’ll just get 11 rosettes for the year instead of 12 :-)
20:11 mvuets I see (-: Good to know.
20:13 Peter neilb, I haven't chimed in yet but will be completed :)
20:13 neilb \o/
20:15 Peter I've done a bunch of small PRs on mine, doing my part to skew the stats ;)
20:16 thrig a challenge for Blog Data to solve
20:17 Peter mvuets, got any advice for convincing someone to sign up?
20:17 thrig beer? chocolate? breaking kneecaps?
20:18 mvuets Peter: I outlined some opportunities that I find interesting in my post: http://cpanpr.tumblr.com/post/107684484713/month-zero
20:19 mvuets Peter: at the last AmsterdamX.pm meetup I tried to sell it to an audience of ~30..40 people, but seems nobody bought it )-:
20:20 Peter This guy was asking about CPAN / making modules etc, so I thought this would be a good opportunity for him :)
20:20 mvuets After all I think it's not something you should actually try _convincing_.
20:20 Peter Indeed, selling it is more what I meant
20:23 mvuets Peter: not sure if this is a good opportunity, but it's definitely a good bunch of excellent people glad to assist.
20:24 Peter mvuets, everything's always a good opportunity if you put in the effort to make it one
20:24 mvuets If he has no particular project/dist in mind and has willingness to get hands dirty---this lottery is a great kick off.
20:24 Peter yeah, i'm one of those guys ;)
20:25 mvuets Haha, happy with your prize, BTW? (-:
20:25 Peter Prize?
20:26 mvuets Yeah, keeping the analogy with the lottery. Happy with the dist you won?
20:27 Peter Yeah, prompted me to learn a bit about async stuff, POE etc
20:27 Peter And when it gets all fixed up I think I'll start using it
20:27 Peter Or something better if I find it :P
20:29 Peter mvuets, the end of the first paragraph of your post says you need to do 144 PRs :P
20:31 mvuets Peter: haha, good spot!
20:32 Mike__B Any #pr-challenge ppl on FOSDEM this weekend?
20:33 * mvuets o
20:33 Mike__B I'll be there, but only on Saturday, unfortunately I'll miss Larry Wall on Sunday :D
20:33 Peter neilb, any thoughts on expanding the challenge to include bitbucket repos too?
20:33 Mike__B But Saturday has the perl devroom.
20:34 neilb are there any dists on bitbucket? ;-)
20:34 mvuets Mike__B: I am pretty sure those talks will be recorded.
20:34 Peter Mine are, plus someone made a D::Z module for it, so there's at least one more :P
20:34 Mike__B https://fosdem.org/2015/schedule/track/perl/ https://fosdem.org/2015/schedule/event/get_ready_to_party/
20:35 neilb Get ready to PaRty — it’s a reference to the challenge!
20:35 mvuets neilb: haha
20:37 mvuets Peter++ typo fixed, thanks. Hope more people will apply now (-:
20:38 Peter I've never had a reason to rebase until this
20:38 Peter I've always wanted too
20:39 Mike__B And of course, this Sunday we'll get our new assignments!! Looking forward.
20:40 mvuets Mike__B: is it why you are not coming to FOSDEM on Sunday? (-:
20:40 Mike__B Of course :D
20:44 Peter I still can't get perlbrew to build anything older than 5.20
20:45 Peter Anyone else run into this kind of thing?
20:47 Mike__B Does it not BUILD or just not test succesfully?
20:47 Mike__B I like plenv better :D
20:49 Peter Mike__B, tests blow up
20:50 Peter Tried all the subversions too
20:50 Peter trying on a vm now
20:50 Mike__B I had the problem before on debian that tests fail due to a lib on my system.
20:50 Peter it is debian
20:50 Peter well, ubuntu
20:52 Mike__B Try plenv :D
20:52 Peter probably doesn't support fish?
20:52 Mike__B or... don't test :D
20:53 Peter Gotta assume something will blow up sometime then?
20:54 Mike__B Oftentimes it's just the test suite failing.
20:54 Peter I used to be able to build them all on the same machine
21:02 Peter Here's a build log curl.io/get/3tzlkecu/3c378ae456aa411ecee30d77e44cb2b04519e150
21:25 Peter Fails on a different ubuntu 14.10 machine too
21:27 deven neilb: I don't know if this is a reasonable request (might be too much effort), but it would be really cool to see some additional things in http://rawgit.com/CPAN-PRC/resources/master/january.html like the status (Complete, Skip, Quit, etc.), date of that status, links to the actual PRs for that module, and links to blog post(s) about the PR (if any)...
21:27 deven It will be interesting to see how things play out as the deadline approaches!
21:32 mvuets joined #pr-challenge
21:33 deven neilb: Of course, info about the PR response (e.g. accepted) would be interesting too, but probably especially hard to compile.
21:34 neilb yeah, I can just find out “merged or not”, but I’m not going to do linguistic analysis of any comments, for example.
21:34 neilb Not this year, anyway.
21:34 deven neilb: That would be a bit much. :)
21:35 deven But having clickable links to the PR would allow anyone interested to follow up on the status of any module they care about.
21:36 deven And blog posts would just be cool to have indexed somewhere, though maybe just a list of them somewhere would work better.  (I know, just Google it!)
21:36 deven After all, some people might blog repeatedly about their challenge for the month -- others once or not at all.
21:36 neilb yeah, will do the clickable links in a later version of the assignments page(s)
21:37 neilb yeah, a feed of the blog posts too
21:37 deven neilb: Will you be able to backport that to January?
21:37 neilb yeah
21:37 deven Yeah, basically an RSS feed of all known CPAN PRC blog posts would be sweet.
21:41 deven haarg: I hope I didn't bite off more than I can chew with this Any::Moose -> Moo conversion! :)
21:41 kanashiro joined #pr-challenge
21:42 deven haarg: One of the cryptic errors from the tests right now: Undef did not pass type constraint "Int" (in $args->{"order"}) at (eval 474) line 397
21:46 deven haarg: Since the debugger didn't seem to work to find where it was failing, I tried some old-school debugging, printing messages from the code with a few judicious "around" wrappers.  Found that the error (on t/02_trivial.t) is occurring in the top-level "Hailo.pm" module, when learn() calls _storage(), which calls order(), which is returning undef, causing the type constraint failure.  No idea yet why it's undef.
21:47 Peter deven, get the old version of the module
21:47 Peter See if it is also undef in that
21:48 deven Peter: Yeah, I'm planning to trace the original code to see if it's undef there too, but it likely isn't.  It had a type constraint already, which surely would have triggered if it was also undef.
21:48 Peter If so, maybe Moose Ints are allowed to be undef and Types::Standard ones aren't
21:49 deven Peter: Well, I did wonder if that was a possibility.  Does anyone know if that's the case?
21:49 Peter should be easy enough to write a little test
21:49 Peter use moose, make class with int attribute, try to set undef
21:50 Peter https://metacpan.org/pod/distribution/Moose/lib/Moose/Manual/Types.pod
21:50 Peter Looks like it will probably not take undef, based on that hierarchy
21:51 thrig please, take my undef.
22:01 deven I think it's just failing to build the storage object properly due to a bug elsewhere.
22:11 veryrusty joined #pr-challenge
22:16 ether neilb: I noticed in your blog you were going to change to considering reverse deps only from other authors - Module-CPANTS-Analyse has code for that that you can steal
22:17 neilb I’ll have a look, but I doubt it does: I have a DB with all of the dependency graph. I just need to write a more complex query :-)
22:19 PerlJam joined #pr-challenge
22:39 ether ah ok fair :)
22:50 sahadev left #pr-challenge
23:02 absolut_todd joined #pr-challenge
23:11 neilb ether: of course I may find that my query-fu isn’t up to it, and I’ll just rip it off … :-)
23:12 ribasushi if only there were toolkits to help with complex query composition... ;)
23:12 ether see the big 'S' in the upper right on an RT ticket page -- that's the link for "this is spam"
23:12 ether that's preferred over 'rejected', so it doesn't affect the statistics for a dist
23:13 ether on dists you control, spam tickets will be deleted outright; for everything else, you increase the vote count by 1
23:14 neilb hey, who let ribasushi in here?!
23:15 neilb ether: ah, didn’t know that, thanks.

| Channels | #pr-challenge index | Today | | Search | Google Search | Plain-Text | summary