Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-07-25

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

All times shown according to UTC.

Time Nick Message
06:28 vicash joined #pdl
07:41 mohawk joined #pdl
13:58 chm joined #pdl
13:59 chm mohawk: please see http://pdl.perl.org/?page=mailing-lists for links to subscribe
14:01 chm sivoais: where are we with "full release with warning" now that sf.net is back up?
14:03 chm mohawk: the key to dev release -> actual release is the "it's not got worse" part
14:03 chm mohawk, sivoais: which is a lot of the motivation to break out PDL into Core + others
14:04 chm mohawk, sivoais: I may have an idea for how to work around this better but still thinking...
14:05 chm mohawk: the purpose of members lists is to reduce spam and the like
14:06 chm whew!  I think that catches me up to today!
14:25 chm sivoais: I took a look at PDL::SV (really needs POD!)---very cool and shows the clarity and simplicity
14:26 chm of using method modifiers and Moo/Moose stuff to implement comprehensible code.  (Still needs POD :-)
16:58 sivoais yep, I need really need to put a POD in :-P
16:59 sivoais I can take a quick go at the full release with warning today
21:54 sivoais chm: one thing to note when working on PDL3... we need something that explicity enumerates cases
21:55 sivoais so that when the code is missing some functionality, there should be something like a linter to tell us
21:55 sivoais I'm looking at Gen/PP and see the following
21:56 sivoais <https://github.com/PDLPorters/pdl/blob/a3961b01fc4df1a6faee99f39f03b3a1778bd63a/Basic/Gen/PP.pm#L3230>
21:56 sivoais paste incoming:
21:56 sivoais # need special cases for
21:57 sivoais # a) bad values
21:57 sivoais # b) bad values + GlobalNew
21:57 sivoais # c) bad values + PMCode
21:57 sivoais # - perhaps I should have separate rules (but b and c produce the
21:57 sivoais # same output...)
21:59 sivoais essentially, it needs to remind the devs that they haven't told the compiler anything about if the code handles badvalues, NaNs, or other cases
22:01 sivoais this is something that is more in the realm of compilers, automated testing, and types
22:01 sivoais I'm trying to remember the name of the type of "random testing" that helps with this. The library QuickCheck for Haskell (and others) helps with this
22:02 sivoais pdurbin sent info on this before, but I forgot the name...
22:04 * sivoais looks up the twitter conv on this...
22:06 sivoais ah! "generative testing"
22:06 sivoais <https://twitter.com/philipdurbin/status/575452861779173376>
22:19 sivoais this blog post was one thing that got me rethinking about the above: <http://strangelyconsistent.org/blog/when-i-dont-even-feel-like-using-blame>
22:19 sivoais NaN problems
22:19 sivoais :-)
22:23 sivoais Also, really important to look at staged programming: <http://okmij.org/ftp/meta-programming/HPC.html>
22:23 sivoais PDL::PP implements that sort of, but needs to be more formal about it
22:38 sivoais I believe run4flat once complained about PDL::PP's inability to check code
22:44 sivoais hmm... Gen::PP doesn't handle HdrCode
22:45 sivoais nvm
22:45 sivoais looked at the wrong thing
23:18 opkick [pdl] zmughal pushed 1 new commit to sf390: http://git.io/vYzca
23:18 opkick pdl/sf390 110b7e8 Zakariyya Mughal: throw a warning when using comparison operator and badval is in {0, 1}...

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