Camelia, the Perl 6 bug

IRC log for #bioperl, 2011-05-19

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

All times shown according to UTC.

Time Nick Message
02:29 CIA-71 bioperl-live: Chris Fields master * r60401c1 / Bio/Root/Build.pm : uncomment versioning code, which sets the version in META.json/yml from dist_version - http://bit.ly/jEpLCE
02:29 CIA-71 bioperl-live: Chris Fields master * rd3d1fb2 / Changes : add a few changes - http://bit.ly/iTKD3E
02:29 CIA-71 bioperl-live: Chris Fields master * rcc02f59 / Changes : add notes about recent fix, and actually change date - http://bit.ly/mAg8uy
02:29 CIA-71 bioperl-live: Chris Fields master * r789e42f / Build.PL : Ace is deprecated, per LDS, not supported anymore, woe be those who want these modules to work - http://bit.ly/kTkcvQ
02:33 CIA-71 bioperl-live: Chris Fields master * r1d6bccb / Changes : perl 5.14.0 not pending anymore - http://bit.ly/iT4Fo0
02:48 dnewkirk left #bioperl
03:04 rbuels left #bioperl
03:05 rbuels joined #bioperl
03:26 scottcain joined #bioperl
03:32 ThePing joined #bioperl
03:32 ThePing left #bioperl
03:34 scottcain left #bioperl
03:39 svaksha left #bioperl
03:40 svaksha joined #bioperl
04:07 kai left #bioperl
04:08 kai joined #bioperl
04:08 kai left #bioperl
04:08 kai joined #bioperl
07:40 kai left #bioperl
07:49 kai joined #bioperl
08:23 kai left #bioperl
08:35 kai joined #bioperl
08:35 kai left #bioperl
08:35 kai joined #bioperl
09:10 ank left #bioperl
11:54 chadadavis joined #bioperl
11:56 scottcain joined #bioperl
12:35 scottcain_ joined #bioperl
12:36 scottcain left #bioperl
12:36 scottcain_ is now known as scottcain
12:38 Guest5712 left #bioperl
12:38 Guest5712 joined #bioperl
12:38 Guest5712 is now known as yang
12:52 leprevost joined #bioperl
13:48 CIA-71 bioperl-live: DaveMessina master * r022b7e7 / (Bio/Tools/Run/WrapperBase.pm t/Tools/Run/WrapperBase.t):
13:48 CIA-71 bioperl-live: - now actually checking if executable was found (and throwing if not).
13:48 CIA-71 bioperl-live: - also fixed test that looked like it was testing for this but wasn't.
13:48 CIA-71 bioperl-live: Signed-off-by: DaveMessina <online@davemessina.com> - http://bit.ly/jBg4eg
14:01 chadadavis left #bioperl
14:14 zenman left #bioperl
14:16 chadadavis joined #bioperl
14:18 ank joined #bioperl
14:22 carandraug joined #bioperl
14:53 carandraug left #bioperl
14:54 carandraug joined #bioperl
14:54 carandraug left #bioperl
14:55 carandraug joined #bioperl
15:08 chadadavis left #bioperl
15:17 scottcain left #bioperl
16:02 scottcain joined #bioperl
16:11 pyrimidine rbuels: FYI, I will be at Galaxy Conf all next week, so contact will be somewhat sporadic
16:12 rbuels pyrimidine: ok cool
16:59 dbolser pyrimidine: where is it at?
16:59 dbolser (geographically speaking)
17:09 pyrimidine dbolser: Lunteren (Netherlands)
17:09 pyrimidine http://galaxy.psu.edu/gcc2011/Home.html
17:14 dbolser ty
17:14 * dbolser bets Peter Cock is there
17:15 dbolser yikes... NBIC, my nemasisis
17:31 scottcain_ joined #bioperl
17:31 scottcain left #bioperl
17:31 scottcain_ is now known as scottcain
17:49 CIA-71 bioperl-live: Chris Fields master * r2e79747 / Bio/Root/Test.pm : run this in an eval block; tests are still failing in bioperl-run, but possibly do to misuse of executable() - http://bit.ly/iNrmt2
18:06 CIA-71 bioperl-run: DaveMessina master * r44559f1 / INSTALL :
18:06 CIA-71 bioperl-run: updating version numbers. Any way to have this auto-inserted via dist-version or similar?
18:06 CIA-71 bioperl-run: Signed-off-by: DaveMessina <online@davemessina.com> - http://bit.ly/jJKO1D
18:36 CIA-71 bioperl-run: DaveMessina master * r407187f / t/BWA.t :
18:36 CIA-71 bioperl-run: Bio::DB::Sam missing from required modules list was causing test failure instead of skip
18:36 CIA-71 bioperl-run: Signed-off-by: DaveMessina <online@davemessina.com> - http://bit.ly/jaKpXh
18:36 CIA-71 bioperl-run: DaveMessina master * r5b9baf9 / t/Blat.t :
18:36 CIA-71 bioperl-run: skip test count was off
18:36 CIA-71 bioperl-run: Signed-off-by: DaveMessina <online@davemessina.com> - http://bit.ly/m8k7yP
18:41 ank left #bioperl
18:52 perl_splut joined #bioperl
18:53 pyrimidine yep, Peter will be there, is giving a talk
19:05 perl_splut interesting little problem on the list. Good thing the main text is in english, heheh
19:44 dukeleto pyrimidine: https://gist.github.com/981517
19:44 dukeleto pyrimidine: bug or feature?
19:46 pyrimidine dukeleto: not sure.  What is being passed to Root::IO?
19:46 dukeleto pyrimidine: a Path::Class::File
19:46 dukeleto pyrimidine: which is created by File::Spec
19:47 rbuels it's not created by File::Spec
19:48 dukeleto pyrimidine: rbuels is correct. It is created by Path::Class
19:48 pyrimidine well, Bio::Root::IO is a bit of a relic, wouldn't be surprised if something is being passed that should work but doesn't
19:49 rbuels dukeleto: the file in question is probably being passed to something bioperly with -file
19:50 pyrimidine rbuels: yeah, that's my guess, looking at the code
19:50 dukeleto rbuels: Bio::BLAST::Database does a use base 'Bio::PrimarySeqI';
19:50 dukeleto rbuels: i think that has something to do with it
19:50 rbuels dukeleto: it's probably mimosa's call to  Bio::SearchIO in the root controller
19:50 rbuels well, that's stringified
19:50 rbuels at least in master
19:51 pyrimidine can you do something like: -fh => $file->open('r')
19:51 dukeleto rbuels: Bio::BLAST doesn't use -file
19:51 dukeleto rbuels: i am working in the multiseq branch
19:51 dukeleto rbuels: one sec
19:52 dukeleto rbuels: now you can pull the multiseq branch
19:52 dukeleto rbuels: it is an odd error
19:53 dukeleto rbuels: it doesn't happen the first time, i have to blast a second time for the error to show up
19:53 dukeleto rbuels: i will look at the SearchIO call in the root controller
19:53 rbuels dukeleto: so then it's someplace in the cached report handling.
19:53 dukeleto rbuels: nope
19:53 pyrimidine just warning everyone, there are some real archaeological finds in Bio::Root::IO.  'Calling Dr. Jones'-type things
19:53 dukeleto rbuels: i have to blast multiple times, but not the same blast runs
19:54 rbuels hehe
19:54 rbuels calling dr. jones?
19:54 * rbuels doesn't get it
19:54 pyrimidine Indiana
19:54 rbuels oooh
19:54 rbuels right
19:54 dukeleto rbuels: not saying it isn't mimosa, but isn't the cached report stuff, at least i don't think
19:59 rbuels anyway, it would probably be nice for Bio::Root::IO to be able to handle objects that stringify to filenames
19:59 rbuels so probably add some "$foo" in there
19:59 rbuels at some point.
20:01 dukeleto pyrimidine: you want a failing test?
20:01 dukeleto pyrimidine: for Path::Class support, that is
20:01 pyrimidine dukeleto: sure
20:02 pyrimidine should be easy enough to add a test for
20:03 pyrimidine I still tend to think the easiest way is to pass this as a file handle, though.  We avoid the stringifying issue (since there isn't a global .Str or ~ operator ala Perl 6)
20:03 rbuels yeah there is
20:03 rbuels double quotes
20:03 * dukeleto lulz
20:04 pyrimidine oh. you mean overloading
20:04 perl_splut good catch on the while =~ m//g <-- missed the missing g
20:04 pyrimidine perl_splut: thx
20:04 dukeleto pyrimidine: you want them in t/Root/RootIO.t ?
20:05 pyrimidine dukeleto: yes.  We can add some code to skip them if Path::Class isn't present
20:06 dukeleto pyrimidine: sure. i will add that as soon as I fix mimosa :) I think i tracked down where this is happening...
20:08 dukeleto rbuels: it was Bio::GMOD::Blast::Graph
20:08 dukeleto rbuels: the outputfile param barfs on Path::Class objects
20:08 pyrimidine rbuels: what I meant by stringifying is an explicit global core operator or method one can use to stringify anything (without the overloading).
20:09 dukeleto rbuels: via the bioperl barf. I will just fix it in bioperl
20:09 rbuels dukeleto: no, because your fix won't get released for a long time
20:09 rbuels dukeleto: well, fix it in both bioperl and mimosa.
20:10 dukeleto rbuels: fixed it in mimosa already
20:10 * rbuels nods
20:10 dukeleto pyrimidine: sure there is, but it isn't easy to find
20:10 pyrimidine dukeleto: what would that be?
20:11 dukeleto pyrimidine: sprintf ?
20:11 dukeleto pyrimidine: or y/// :)
20:11 dukeleto pyrimidine: there are many
20:11 dukeleto pyrimidine: tr///
20:12 dukeleto pyrimidine: anything that forces string context
20:13 pyrimidine but what do you get with an object that doesn't overload stringification?  Just the address, correct?
20:15 dukeleto pyrimidine: yeah, which isn't what you want
20:15 pyrimidine I'm thinking of something along the lines of $foo.Str in Perl 6, where one can define a Str method to stringify an object
20:15 dukeleto pyrimidine: how is that different from overloading "" ?
20:15 dukeleto pyrimidine: seemingly, if someone didn't overload .Str in perl 6, the same problem occurs, right?
20:16 dukeleto pyrimidine: ~ just calls .Str on objects in Perl 6, iirc
20:16 pyrimidine I think there is a subtle difference in Perl6 with Str, can't recall though
20:16 pyrimidine but yes, I think you get the address if Str isn't defined
20:16 dukeleto pyrimidine: the subtle difference is a function of time :)
20:16 pyrimidine yep
20:19 pyrimidine IIRC the problem I have run into with overloading is it tends to bite you when you least expect it (violates that whole 'law of least surprises' thing).  But at the same time it's damn convenient :)
20:20 pyrimidine I think this cropped up with Bio::Annotation at some point
20:20 pyrimidine (namely, it abusing overloading to an obscene degree)
20:21 dukeleto pyrimidine: indeed. overloading is lovely 80% of the time, and causes maddeningly subtle bugs the other 20%
20:26 dukeleto pyrimidine: overloading is a huge performance penalty
20:26 pyrimidine yeah, I ripped that code out
20:26 dukeleto pyrimidine: i learned that when i started maintaining Math::MatrixReal
20:26 pyrimidine (of Bio::Annotation)
20:26 pyrimidine someone should let Lincoln know that, with Bio::DB::SeqFeature
20:26 dukeleto pyrimidine: it is fine for end-users to write it, but forcing everyone to use it under the hood is another thing entirely
20:27 pyrimidine agreed
20:43 CIA-71 bioperl-run: Chris Fields master * r59e7fbf / lib/Bio/Tools/Run/RepeatMasker.pm : don't call executable() - http://bit.ly/jmCqg2
20:59 scottcain left #bioperl
21:02 CIA-71 bioperl-live: Jonathan "Duke" Leto master * r67d124f / t/Root/RootIO.t : With little effort, our hero valiantly slays THE DEPENDENCY HOBGOBLIN! (TM) - http://bit.ly/jLLGQp
21:03 dukeleto rbuels: ^^^
21:03 * rbuels chuckles
21:03 CIA-71 bioperl-run: Chris Fields master * r4a685f9 / (Build.PL t/MCS.t): this module requires Config::Any; skip these tests if Config::Any isn't present - http://bit.ly/jjxpuB
21:04 rbuels nobody expects the dependency hobgoblin
21:06 CIA-71 bioperl-run: Chris Fields master * r8e4f7f6 / (Build.PL t/Match.t): this module requires Config::Any; skip these tests if Config::Any isn't present - http://bit.ly/jvXJAx
21:07 pyrimidine dukeleto++
21:07 * pyrimidine -> home
21:52 Yksi joined #bioperl
21:53 Yksi left #bioperl
21:57 carandraug left #bioperl
22:16 leprevost left #bioperl
22:51 CIA-71 bioperl-live: Jonathan "Duke" Leto topic/leto_path_class * re2a22d4 / t/Root/RootIO.t : Add some tests for Bio::Root::IO->new taking Path::Class objects as input - http://bit.ly/mPPRSX
23:04 CIA-71 bioperl-live: Jonathan "Duke" Leto topic/leto_path_class * rd6f4efa / t/SearchIO/blast.t : Perltidy t/SearchIO/blast.t, no functional changes - http://bit.ly/kQOL34
23:04 CIA-71 bioperl-live: Jonathan "Duke" Leto topic/leto_path_class * r466e3c8 / t/SearchIO/blast.t : Add warnings to t/SearchIO/blast.t - http://bit.ly/lYHzqe
23:04 CIA-71 bioperl-live: Jonathan "Duke" Leto topic/leto_path_class * r96ce0ea / t/SearchIO/blast.t : Add some tests for Bio::SearchIO->new understanding Path::Class input - http://bit.ly/mwb6o7
23:27 CIA-71 bioperl-live: Jonathan "Duke" Leto topic/leto_path_class * rf43bb4c / t/SearchIO/blast.t : Add more tests for Bio::SearchIO + Path::Class objects - http://bit.ly/kzdC9X

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