Camelia, the Perl 6 bug

IRC log for #dbdi, 2010-06-26

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

All times shown according to UTC.

Time Nick Message
11:40 timbunce joined #dbdi
11:46 mberends joined #dbdi
11:46 timbunce mberends: hi!
11:46 mberends howdy!
11:46 cognominal joined #dbdi
11:47 mberends some implementation to date: http://github.com/mberends/fakedbi
11:47 mberends just based on DBI v1 of course
11:48 mberends symmetrically, I'll add #dbdi to my auto-join list :)
11:48 timbunce :)
11:48 timbunce I posted something a while (month) ago with some kind of plan. I'll find that, update it, and send it to the dbdi-dev mailing list.
11:49 timbunce are you familiar with the idea of DBDI?
11:49 mberends yes, I've read the roadmap docs
11:50 * timbunce reads your README
11:50 mberends I understand that DBDI will apply to Perl 5 and 6
11:50 timbunce so at first sight you're working at (or providing an interface at) the layer above where I'm focussed at the moment. So that's good.
11:51 mberends very good, even
11:51 timbunce not perl5 - at least I've no plans (other than perl5-on-perl6 I guess)
11:51 mberends oh, I misunderstood that point
11:51 timbunce I'll update the (very) old roadmaps docs in the DBI for the next release
11:52 timbunce Perl 5 DBI has become too hard to change. Best to leave it on long-term life support.
11:53 mberends fair enough.  the migration workload would be immense.
11:53 timbunce do you have plans to add sqlite support soon? That would make life easy for testers who wouldn't want to install mysql.
11:54 mberends yes, it's currently blocked on a Parrot NCI problem that should be solved by ash_'s GSoC project by August.
11:55 timbunce cool.
11:55 timbunce Would you prefer if I forked fakedbi on github and sent you pull requests or would you rather give me write access?
11:56 mberends I prefer you to have write access
11:56 mberends because I may be absent for several days at a time
11:56 timbunce anything you need from me to do that?
11:58 mberends found you on github. I'm watching the Perl5 on Android talk, will resume in a few minutes :)
11:58 timbunce :)
12:00 moritz_ timbunce: in general there's a whole lot of perl 6 projects on github to which commit bits are given out freely (and automated via a bot)
12:01 timbunce moritz_: I'd be happy to move java2perl6 there (it's currently on https://java2perl6.googlecode.com/svn/trunk) How can I integrate with the bot?
12:03 moritz_ timbunce: either by /msg'ing me the github API key, or by moving the github project to somebody whos API key is already register (mine or  'perl6' for example)
12:03 timbunce ok, thanks.
12:04 moritz_ (for the 'perl6' account you'd need to contact pmichaud, who holds the keys)
12:14 mberends joined #dbdi
12:55 mberends intermission at perworkshop.be, tux has just told us why he hates *every* database
13:20 moritz_ and is it the same reason that he maintains the CSV modules on CPAN? :-)
13:24 cognominal mberends, with only one track you will get more people attending your Perl 6 talk than in Calais :)
13:25 mberends moritz_: yes :) and he recommends perl5 hash lookup tables instead of joins
13:25 mberends cognominal: yes, some delegates have to resort to leaving early to be able to miss my talk ;)
13:36 cognominal :(
13:52 moritz_ timbunce: where do I get jdbclib from?
13:52 timbunce perl mkjdbc.pl
13:53 moritz_ mv: cannot stat `jdbclib': No such file or directory
13:53 timbunce I'm checking in a fix about now...
13:54 timbunce done
13:56 timbunce fails on DateTime still as that's not in rakudo yet. I'll TypeCast.pm that to Any for now
13:58 timbunce fails on some due to recursion - that may need a pre-pass to fix so we know when to predeclare a stub for the role.
13:59 timbunce (which is pretty much what stalled me last time I hacked on this - but rakudo has improved much since then)
13:59 moritz_ rakudo: say DateTime.new
13:59 p6eval rakudo 6769e1:  ( no output )
13:59 moritz_ rakudo: say DateTime.now
14:00 p6eval rakudo 6769e1: OUTPUT«2010-06-26T13:41:14+0000␤»
14:00 timbunce ah, I'd forgtooen to add it to the list of bultins. thanks.
14:01 timbunce dateTime isn't mentioned in S02 - should it be?
14:03 mberends timbunce: it's moved. http://perlcabal.org/syn/S32/Temporal.html
14:05 * timbunce wonders where 'builtin' ends and 'bundled' begins. I could probe for them but it doesn't seem worth it - this is just a migration bump.
14:07 timbunce If a 'use' isn't required then doesn't that count as 'builtin'?
14:07 mberends the synopses stuff should be all built in.  What's new is that much of it is implemented in Perl 6, so the build process makes a stage 1 compiler without them and then uses that to build a bigger version of itself with the P6 defined ones added in.
14:08 timbunce mberends: sure, but S02 says "Built-In Data Types" without listing DateTime. I'll raise it on #perl6
14:09 mberends ok. S02 should refer to the higher numbered docs
14:10 timbunce got to go, ttyl
14:10 mberends o/
14:11 timbunce one last thing... my last commit has java2perl6 working for mkjdbc.pl except for the failures due to recursion.
14:12 timbunce If anyone wants to look into those I'd be *delighted*!
14:12 mberends right, I'll give it a try!
14:49 moritz_ I've traced one of the infinite recursions
14:49 moritz_ it's module loading that recurses
14:49 moritz_ but I guess you knew that...
16:27 mberends joined #dbdi
20:26 timbunce moritz_: yeap. The trick is to know when to generate "use Foo;" vs something like "role Foo { ... }" (I guess)
21:06 timbunce joined #dbdi

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