Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-01-03

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

All times shown according to UTC.

Time Nick Message
00:12 Molaf__ joined #pdl
02:41 frogonia joined #pdl
02:42 dystan joined #pdl
03:44 ilbot2 joined #pdl
03:44 Topic for #pdl is now Install PDL: http://pdl.perl.org/?page=install  | Book: http://pdl.perl.org/content/pdl-book-toc.html | Mailing list: http://pdl.perl.org/?page=mailing-lists | Pasting: http://scsys.co.uk:8001/pdl | Channel is logged by ilbot2: http://irclog.perlgeek.de/pdl/
09:14 Molaf joined #pdl
15:13 jberger run4flat: PDL's pod system has been relying on pod conventions and meta-model stuff that has not just been deprecated, they have been removed since 1997-8
15:14 jberger we need to identify which "odd" behaviors to keep and we should try to join the rest of the world on the other stuff
16:09 run4flat OK, I'm in
16:09 run4flat :-)
16:10 run4flat I suppose that explains why we've never removed PDL::Doc from the PDL system. :-)
16:13 test4flat joined #pdl
16:13 test4flat another test
16:13 test4flat try again
16:13 test4flat ah, no noise. :-(
16:19 test4flat joined #pdl
16:19 test4flat sounds!
16:20 run4flat cool
16:20 test4flat exit
16:20 test4flat ha!
16:38 jberger run4flat, having fun?
16:38 run4flat almost, except that it didn't make noise when you said that!
16:38 run4flat grrr!
16:38 run4flat oh well
16:39 jberger ok, I'm going to start a quick branch for this stuff
16:39 run4flat sweet
16:39 run4flat keep me posted
16:48 jberger hey, I'm getting an error here: https://github.com/PDLPorters/pdl/blob/master/Doc/Doc.pm#L71
16:48 jberger can you try to glean what is intended?
16:48 jberger while I'm working on other stuff
16:48 run4flat ok
16:53 run4flat I'm going to put my notes here
16:53 run4flat feel free to ignore until I ping you
16:53 run4flat I believe that the "Parmode" key means "Parse_mode"
16:55 run4flat I don't understand why both Mode and Parmode exist
16:56 run4flat OK, I think I see. =head1 sets Mode to NAME, and Parmode to NAME
16:57 run4flat If you the pod author has other directives below this, i.e. =head2 (which would be highly unconventional), it goes to "last"
16:57 run4flat I have no idea where that "last" takes us
16:57 run4flat it should probably simply pass and do nothing
16:57 run4flat i.e. replace "last;" with "# do nothing"
16:58 run4flat how did you trip that error?
16:58 run4flat ping jberger, see my answer two lines up
17:00 jberger oop, hi
17:00 jberger =head1 NAME
17:01 jberger which was a "lucky" accident in my testing
17:01 run4flat you just put =head1 NAME and that did it?
17:01 jberger yeah
17:01 run4flat i.e. nothing else? Except maybe =cut?
17:01 jberger no, I had some test text
17:01 jberger but that was all
17:02 run4flat can you paste your exact test string?
17:03 jberger drat, in my last round of cleaning I think I dumped the file on accident
17:03 jberger harumph
17:03 jberger I think I can recreate it
17:03 jberger my netbook takes quite a while to compile PDL
17:06 run4flat yes, it takes way to long
17:06 run4flat I don't think it's too big a deal
17:06 run4flat I think that the replacement I suggest should work
17:07 jberger I want to say that I used
17:07 jberger =head1 NAME
17:07 jberger ** blank
17:07 jberger Some I<test> text C<here>.
17:07 jberger ** blank
17:07 jberger =cut
17:08 jberger where the blanks are empty  lines of course (but my IRC client wont send them)
17:08 run4flat right. :-)
17:08 run4flat OK, I bet it choked on the =cut
17:08 run4flat if you replace that with a =head2 or =over, it would choke
17:08 run4flat if you insert another =head1, it should be fine
17:09 run4flat nice catch
17:09 jberger hmmm
17:09 jberger interesting
17:09 jberger I will test when I'm done compiling
17:09 run4flat ok
17:10 run4flat If removing the "last" statement fixes it, I would recommend not including that in the test suite
17:10 run4flat I can't imagine that bug being re-introduced
17:10 run4flat and it would obfuscate the test suite for anybody who might come to it in the future
17:16 jberger I'm not sure I'm creating a test suite
17:16 run4flat oh, ok
17:16 run4flat :-)
17:16 run4flat good then
17:16 jberger I have started a file, but I'm not sure what I'm testing for
17:16 jberger exactly
17:17 jberger by depending on Pod::Parser/Select we get much improved test coverage
17:17 jberger it just might not be what is expected :-P
17:17 run4flat if you wanted to write tests, and I'm not saying you should,
17:18 run4flat I would test that scan does what we expect
17:18 run4flat https://github.com/PDLPorters/pdl/blob/master/Doc/Doc.pm#L659
17:18 run4flat that should return a hash with the relevant stuff for the docs database
17:19 jberger I say this, not meaning that I'm trying to pass the buck, but I think you should do it
17:20 jberger I don't know what is expected, and I'm not sure I can take the time to learn
17:20 jberger make some simple test from the master branch
17:20 jberger test it against my branch
17:20 jberger and then if it all seems good, then merge my branch
17:21 run4flat fair enough
17:22 jberger basically, I think the PDL::Pod:: stuff though very different from the Pod:: stuff stayed close enough in API that we dodged a bullet
17:22 run4flat thank god
17:22 jberger the PDL::PodParser I think is patchable easily to be a subclass of Pod::Select, which is what I'm working on now
17:23 run4flat ok
17:23 jberger PDL::Doc only uses those instances
17:23 run4flat I'll see if I can write some tests
17:23 run4flat yeah, thankfully it is fairly tight
17:23 jberger though it does dip into the meta-model hash
17:23 run4flat docs stuff isn't strewn all over the place
17:23 jberger but I THINK only on the keys that PDL creates
17:23 run4flat yeah, that was my impression, to
17:24 run4flat easiest way is to ack an un-cleansed version of the PDL source
17:33 jberger hmmm, next problem
17:33 run4flat k
17:34 jberger PDL::Doc has two classes which are supposed to behave like handles
17:34 jberger but they aren't actually handles
17:34 jberger and the old system didn't care, but the new one dies with not a globref
17:34 run4flat Yes, I suspect those can be done away with
17:34 jberger I was just going to reimplement with a tied handle
17:34 run4flat can't we open a file handle to a string
17:34 jberger does that work
17:34 run4flat ?
17:35 jberger probably
17:35 run4flat which Perl introduced that?
17:35 jberger I think 5.8 can handle it, which is what chm said to target
17:35 run4flat yeah, 5.8
17:35 run4flat http://perldoc.perl.org/perlfaq5.html#How-can-I-open-a-filehandle-to-a-string?
17:35 jberger Since v5.8.0, Perl has built using PerlIO by default.  Unless
17:35 jberger you've changed this (such as building Perl with "Configure
17:35 jberger -Uuseperlio"), you can open filehandles directly to Perl
17:35 jberger scalars via:
17:36 jberger from perldoc -f open
17:36 jberger ok so we should be good
17:36 run4flat yep
17:36 run4flat Yeah! We get to use the shiny new stuffs!
17:41 run4flat I fear that PDL dipped into the SYMS hash
17:43 shadowpaste "run4flat" at 217.168.150.38 pasted "First notes on testing the docs database system" (8 lines) at http://paste.scsys.co.uk/221697
17:43 run4flat Alright, it's not awesome, but it's a start
17:43 run4flat some of the docs stuff should be refactored, I think
17:43 run4flat This should be run by Chris, as he was the one who wanted to push this along. :-)
17:46 opkick joined #pdl
17:46 opkick [pdl01] jberger created pod (+1 new commit): http://git.io/gSYkWQ
17:46 opkick pdl/pod 6245954 Joel Berger: removed PDL::Pod:: classes in favor of core modules Pod::...
17:46 opkick left #pdl
17:47 jberger I thought SYMHASH was a PDL key?
17:48 run4flat SYMHASH probably is, but I don't know that SYM is
17:48 run4flat sorry, SYMS
17:49 run4flat See, for example, https://github.com/PDLPorters/pdl/blob/master/Doc/Doc.pm#L500
17:49 jberger as far as I can tell, these are the old field names
17:49 jberger https://metacpan.org/source/MAREKR/Pod-Parser-1.51/CHANGES#L674
17:50 run4flat thanks, I had misplaced that link
17:50 run4flat ah, you are correct
17:51 jberger SYMS is actually a field on PDL::Doc which is not a subclass of PDL::PodParser
17:51 jberger so that should be a-ok
17:51 run4flat ok
17:51 run4flat just never explicitly declared anywhere, as you nko
17:51 run4flat know
17:51 jberger right
17:51 run4flat that makes it hard to be sure
17:52 run4flat but ack tells me we're ok
17:52 jberger but its used on the instance of PDL::Doc
17:52 jberger anyway, make doctest succeeds on my branch now
17:52 jberger ooops I forgot to kill the 'last'
17:52 run4flat haha
17:53 jberger you know what really helped
17:53 jberger I tacked -MCarp::Always into the makefile directive for make doctest
17:53 jberger brilliant!
17:54 jberger full stacktrace on all errors
17:54 run4flat haha
17:54 run4flat excellent
17:54 run4flat I love that module
17:54 jberger I've never needed that before, but I've always heard the Perl greybeards mention that they love it
17:54 * jberger just called run4flat a greybeard probably
17:54 run4flat it's fantastic for debugging
17:54 * run4flat chuckles
17:55 * run4flat actually wheezes
17:55 * run4flat claims it was in an attempt to not annoy his labmates
17:56 run4flat jberger also just claimed to have a few grey hairs himself
17:56 opkick joined #pdl
17:56 opkick [pdl01] jberger pushed 1 new commit to pod: http://git.io/mkGhlQ
17:56 opkick pdl/pod 69260d2 Joel Berger: removed problematic "last" directive
17:56 opkick left #pdl
17:56 jberger hahah
17:56 jberger ok play with that a little bit
17:57 jberger the thing to really remember is that PDL::Pod::Parser is replaced by Pod::Select not just Pod::Parser
18:04 run4flat yep
18:04 run4flat alright, boss is in, I gotta go
18:04 run4flat cya!
18:04 jberger 0/
18:09 jberger the only thing I haven't removed is the big warning in the documenation saying that using the Pod:: classes won't work
18:09 jberger I wish it said WHY they wouldn't
18:34 opkick joined #pdl
18:34 opkick [pdl01] jberger pushed 1 new commit to pod: http://git.io/KX6pog
18:34 opkick pdl/pod c4117fd Joel Berger: removed warning about Pod::Parser incompatibility
18:34 opkick left #pdl

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