Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-03-15

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

All times shown according to UTC.

Time Nick Message
00:45 brunov joined #bioperl
03:35 driveby_bot joined #bioperl
03:35 driveby_bot /home/svn-repositories/bioperl: r16922 (maj) : fix _fastize changing sequence alphabet upon gap-stripping
03:35 driveby_bot Diff: http://tinyurl.com/ygpgjy5
04:27 neL joined #bioperl
04:40 neL left #bioperl
07:22 Furiiku joined #bioperl
12:51 Furiiku joined #bioperl
14:26 melic joined #bioperl
15:27 brunov joined #bioperl
16:01 perl_splut joined #bioperl
16:12 pyrimidine joined #bioperl
17:33 pyrimidine rbuels: any more word on code (beyond it being foobar)?
17:34 rbuels what, who said it was foobar?  ;-)
17:34 pyrimidine s/foobar/fubar/ ;>
17:34 rbuels yeah, dukeleto and i hacked on it yesterday
17:34 rbuels actually had an epiphany
17:34 rbuels (but mostly failed at our tasks)
17:35 rbuels the nonintuitive epiphany is:  from-the-bottom is the only sane way to proceed with breaking it up
17:35 rbuels because:
17:36 rbuels a.) you can break a piece of code off the top, but as other pieces get broken off, you will have to go back and update the dependencies of everything you broke off the top, again and again
17:37 rbuels and b.) it's actually easier to break something off of the bottom, because it's much easier to trace dependencies *downward* (toward the root) than upward
17:37 rbuels i was thinking about formulating this in a long-form email to the list ...
17:38 rbuels so, the results of our hacking yesterday were 2 successful new dists (Bio-PhenoType and Bio-EntrezGene), and one unsuccessful one (Bio-Ontology)
17:40 rbuels we did our work in a fork of the bioperl-live repo, so it will need a little poking if we want to actually merge, but i don't really think it needs to be merged
17:42 pyrimidine ok
17:44 pyrimidine It would be easy to split off Bio::Root* and call it Bio-Root (or core), then have everything have that as a dependency
17:44 rbuels yep.
17:45 rbuels we just have to make sure that there is a place for people to pull 1.6.1 from svn
17:45 rbuels and make sure they know to pull from there
17:45 rbuels cause just pulling bioperl-live will stop working one bio-root is split off
17:45 rbuels s/one/once/
18:19 pyrimidine sorry, was away
18:19 pyrimidine My philosophy re: svn is
18:19 pyrimidine if one is reliant on cutting edge code, then expect to get cut
18:21 pyrimidine if we build a build dependency on Bio::Root into live, and then someone downloads it and it doesn't work, it makes me think they aren't actually doing things the proper way
18:25 driveby_bot joined #bioperl
18:25 driveby_bot /home/svn-repositories/bioperl: r16923 (nml5566) : fixed case change for Bio::SeqFeature::Tools::TypeMapper::FT_SO_map()
18:25 driveby_bot Diff: http://tinyurl.com/yg6v6e9
19:13 rbuels pyrimidine: ho ho, well yes they aren't doing things the proper way
19:14 rbuels pyrimidine: but they probably get irate because they have come to rely on that behavior ...  but the reply can just tell them to check out the 1.6.1 release tag, or whatever
19:14 * rbuels shrugs
19:17 pyrimidine yes, that's the best way to address it.  I don't think this will be an issue in the long run, once we work things out.
19:19 pyrimidine we'll bump trunk to 1.7 at some critical juncture, then start working things out on branches and merging back to bp-live
19:20 pyrimidine (well, whatever is left over will be merged back)
19:21 * rbuels volunteers the present as the so-called critical juncture
19:21 rbuels lol
19:21 rbuels really, the splitting up throws the present version numbers out the window
19:22 rbuels actually, that was a question leto and i were thinking about yesterday ....
19:22 rbuels all or most of the modules need $VERSION
19:22 rbuels so should we ....start them at 0.10 or something ...
19:22 rbuels or start them at '1.007000' ...
19:23 rbuels or '1.7.0',..... or whatever is the right thing for perl's screwed-up versioning
19:23 * rbuels looks
19:27 * rbuels votes for starting split-off modules at '1.700'
19:27 rbuels because 1.007000 is too many zeroes
19:28 rbuels or ... wait.... perldoc version suggests using qv('1.7.0')
19:30 rbuels bah.  we should just drop the third part of the version and make the versions numeric
19:30 * rbuels votes again for '1.700'
19:31 rbuels too bad deafferret is in mexico being held for ransom
19:43 pyrimidine I think Bio::Root* would retain the original versioning.  Everything else in bp-live would need to have something comparable or higher.
19:44 pyrimidine Everything new could have it's own.
19:46 rbuels that's reasonable
19:47 pyrimidine Bio::Graphics switched to a simple duple for versions (now at 2.01).  We could do the same with the others I suppose.
19:47 rbuels yeah i think that would be best
19:47 rbuels so split off Bio-Root and keep it 1.7.0 or whatever
19:47 pyrimidine yes
19:47 rbuels and other things split from bp-live could start at 1.700
19:47 rbuels or 1.70
19:48 rbuels (and any new modules can do version numbering however they want)
19:48 pyrimidine yes
19:48 rbuels ok, so the only question is, '1.70' or '1.700'
19:49 pyrimidine We just need to ensure that any current modules can still be compared to their old version numbers
19:49 rbuels ah.
19:49 pyrimidine so, if SearchIO were split off, we need to compare 1.6.1 against 1.7, and 1.7 should come out on top
19:49 pyrimidine (as newer)
19:49 rbuels perl -Mversion -E 'say qv(1.70) > qv("1.6.1")'
19:50 rbuels gets 1
19:50 pyrimidine okay, wfm
19:50 rbuels wfm?
19:50 pyrimidine works for me
19:50 rbuels enopurl
19:50 * rbuels nods
19:52 pyrimidine I think, for 1.7, we should bump the perl dep to 5.8.9 (or whatever the latest is for that series).  I would attempt 5.10 but I think we will have problems with some users with that.
19:52 pyrimidine (seeing as 5.12 is just around the corner)
19:53 rbuels sounds good to me
20:42 wessidepraxis joined #bioperl
20:47 wessidepraxis joined #bioperl
21:11 wessidepraxis joined #bioperl
21:29 etal joined #bioperl
21:33 pyrimidine left #bioperl

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