Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-07-20

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

All times shown according to UTC.

Time Nick Message
00:01 dukeleto rbuels: that is what I planned, to test the death wail
00:01 carandraug joined #bioperl
00:03 tyru joined #bioperl
00:03 tyru left #bioperl
00:07 deafferret commute ... biab
00:07 rbuels dukeleto: i think after that test is in there, we can merge it
00:12 dukeleto rbuels: shouldn't this go in t/Ontology/Relationship.t ?
00:13 dukeleto rbuels: maybe not, hmmm
00:14 rbuels dukeleto: those tests are probably nonoptimal.
00:14 rbuels but i have not closely looked at them.
00:15 dukeleto rbuels: i am testing the add_relationship method on the Bio::Ontology::Ontology object, so i think t/Ontology/Ontology.t is the best place for it
00:15 rbuels dukeleto: yes i think so too
00:16 rbuels dukeleto: be sure to pull, i perltidied a bit
00:16 * dukeleto pulls
00:27 dukeleto rbuels: i am going to convert solgenomics to an org. it is irreversible. pressing the big red button in 1 minute
00:27 rbuels dukeleto: sounds fine to me
00:27 * dukeleto imagines the bomb countdown sequence from Space Balls
00:36 rbuels damn, the das registry looks like ass
00:36 rbuels http://www.dasregistry.org/listServices.jsp​?organism=Drosophila%20melanogaster&CSN​ame=BDGP_5.4&CSTypes=Chromosome&cap​abilities=any&labels=any&cmd=find
00:44 dukeleto seems like adding a circular relationship makes the test suite go into an infinite loop. this seems bad.
00:51 rbuels dukeleto: heh.
00:52 rbuels cycle checking is Hard though.
00:52 rbuels and a big performance hit Will Not Do for large ontologies
01:08 CIA-95 bioperl-live: Duke Leto topic/leto_interpro_parse * r12aa177 / t/Ontology/OntologyEngine.t :
01:08 CIA-95 bioperl-live: Add tests for the exceptions thrown by invalid Bio::Ontology::Relationship's
01:08 CIA-95 bioperl-live: This adds tests for the functionality added in d1d9d1ab26ddcb8f92e3663035623e96e193c5f4 - http://bit.ly/aXdMjf
01:20 dukeleto rbuels: who shall merge the branch?
01:20 rbuels dukeleto: the branch shall be merged by rbuels, and the merger of the branch shall be him
01:20 dukeleto rbuels: we have a failing test
01:21 dukeleto rbuels: t/Ontology/IO/interpro.t
01:21 rbuels dukeleto: and the angels in the heavens wept hot tears of .... um, weeping
01:22 * rbuels looks
01:22 dukeleto rbuels: i am fixing it
01:24 dukeleto rbuels: the region term has no parent
01:24 dukeleto rbuels: should that be?
01:25 rbuels i think it should be .......
01:25 * rbuels checks
01:25 rbuels wow
01:25 rbuels git bisect run prove -I. -r   t/Ontology/IO/interpro.t  FTW
01:26 dukeleto rbuels: i think that is actually fine
01:26 dukeleto rbuels: it is a root term
01:26 rbuels dukeleto: ah. yes.
01:26 rbuels dukeleto: the tests just need to be updated
01:27 dukeleto rbuels: just pushed
01:27 CIA-95 bioperl-live: Duke Leto topic/leto_interpro_parse * rcfafb29 / t/Ontology/IO/interpro.t : Fix some failing InterPro tests, since the number of root terms changed - http://bit.ly/9HBkv5
01:27 * dukeleto runs a full test suite now
01:30 dukeleto rbuels: full test suite passes. merge my battleship
01:33 rbuels dukeleto: merging.
01:33 rbuels (and running tests)
01:34 dukeleto yay for testrunning
01:40 rbuels there's a test failure in trunk
01:40 * rbuels bisects
01:40 * rbuels thinks it is because of fangly
01:40 rbuels usually is.
01:40 rbuels he's not so hot at test running.
01:41 rbuels we will see what the bisect turns up though
01:41 rbuels yep
01:41 rbuels f81d406d6092019f2a4e86d7f173dd487e49f924
01:54 deafferret can you nopaste an actual bisect run?
01:58 * rbuels digs in his backscroll
01:58 dukeleto i saw fangly commiting directly to master and it looked a bit fishy
01:58 rbuels he is fishy in his practices.
01:59 * rbuels looks askance
01:59 dukeleto rbuels: which test is failing?
01:59 rbuels dukeleto: oh it's just test count foolishness
01:59 dukeleto rbuels: i think i don't have the deps installed to see it
01:59 rbuels dukeleto: because Assembly.t is a horror
01:59 rbuels dukeleto: i have broken it into 3 files
01:59 rbuels dukeleto: and deleted all the stupid subroutine refs and test count BS from it
02:00 deafferret loose cannon!
02:00 * dukeleto cleans his flame-thrower in anticipation of improving the bioperl test suite
02:02 dukeleto yeah, i didn't have Ace installed, so I didn't notice his failing tests
02:03 dukeleto rbuels: if bioperl supports back to 5.6.1, we have have to change the qr// in the test to a plain string
02:03 dukeleto we may have, rather
02:03 rbuels whoa, no qr in 5.6?
02:03 * rbuels did not know that
02:03 dukeleto rbuels: i am not sure, let me look
02:04 dukeleto rbuels: actually, they were added in 5.6.0
02:04 dukeleto good to know
02:07 * dukeleto runs ./Build installdeps for bioperl
02:13 CIA-95 bioperl-live: Robert Buels master * rb500f65 / (4 files): broke t/Assembly/Assembly.t into 3 sane test files (+13 more commits...) - http://bit.ly/chY9KB
02:13 rbuels dukeleto: ok the branch is merged and pushed
02:14 rbuels dukeleto: installdeps is not going to do much, most of the deps are optional
02:14 rbuels dukeleto: as an artifact of nobody ever breaking things off into more dists
02:42 pyrimidine joined #bioperl
02:46 pyrimidine rbuels, dukeleto: still around?
02:46 rbuels pyrimidine: dukeleto is out, but i am still here
02:46 pyrimidine cool
02:47 rbuels pyrimidine: there might be some additional cleanup needed on those test files, but I don't have the deps to test them fully
02:47 pyrimidine will check them out.
02:47 pyrimidine what was the test fail on trunk?
02:48 rbuels pyrimidine: oh it was just test count silliness
02:48 rbuels pyrimidine: again.
02:48 pyrimidine ;P
02:48 rbuels pyrimidine: which is why i hauled off and took a hammer and chisel to that .t file.
02:48 pyrimidine fangly?
02:48 rbuels yep.
02:48 pyrimidine grr
02:48 rbuels well it probably worked on his machine with all the deps.
02:48 rbuels probably.
02:48 * rbuels gives him the benefit of the doubt
02:49 pyrimidine well... this has happened before, but okay
02:51 pyrimidine glad to see dukeleto lending a hand.  just waiting to see what he thinks of the code base!
02:51 pyrimidine ;)
02:52 rbuels heh.
02:52 pyrimidine feel free to rip up Bio::Ontology*, btw.  It probably needs it.
02:52 rbuels seems to be holding up remarkably well actually
02:52 rbuels test coverage is of course not so hot, but those were retrofitted right?
02:53 rbuels can't really expect that much from retrofitted tests...
02:53 pyrimidine that seemed to happen quite a bit in the test suite, actually
02:53 pyrimidine SearchIO had quite a bit of that IIRC
02:53 rbuels well there didn't used to *be* any tests right?
02:54 pyrimidine yep
02:54 rbuels (er, hardly any)
02:54 pyrimidine so anything is better than nothing
02:54 rbuels yes indeed
02:54 pyrimidine as long as 'anything' is actually testing something
02:54 rbuels pyrimidine: do you have the deps required to run all the t/Assembly stuff?
02:54 pyrimidine yes
02:54 pyrimidine I think
02:55 rbuels pyrimidine: because i'm fairly sure that the refactored scripts are a.) going to leave some garbage files around, and b.) have wrong test counts
02:55 pyrimidine they don't have any clean up?
02:55 rbuels pyrimidine: there was some hamhanded cleanup that didn't make it through the chiseling, i was hoping that something a little better might come out in the fixup
02:55 pyrimidine actually, now I think about it, there were a few files lurking about after tests were run
02:56 rbuels pyrimidine: perhaps a little use of oh.... File::Temp?
02:56 pyrimidine hmm, one would think
02:56 pyrimidine Sendu even added a bit o' sugar for that purpose
02:56 pyrimidine test_output_file()
02:56 rbuels ah.  sugar.
02:57 rbuels sugar often has unintended consequences of tight coupling that prove to be troublesome later on.
02:57 pyrimidine yes
02:57 pyrimidine that's one reason why I would like to quietly let Bio::Root::Build go
02:57 * rbuels looks around conspiratorially
02:58 * rbuels looks under deafferret's belly rolls for any gnomes that might be hiding
02:58 * pyrimidine thinks rbuels said 'jelly rolls', is hungry
02:58 pyrimidine Bio::Root::Test is just an easy way of pulling in Test::More/Warn/Exception
02:58 rbuels heh
02:59 pyrimidine with the aforementioned sugar
02:59 * pyrimidine actually spelled 'aforementioned' correctly there
03:00 pyrimidine look did it again!
03:00 rbuels wow.
03:00 rbuels well, it's phonetic
03:00 rbuels mostly.
03:00 pyrimidine true, will probably mispell something
03:00 * rbuels is not that impressed, really
03:00 pyrimidine ;)
03:00 rbuels you will have to try harder
03:00 pyrimidine heh
03:01 rbuels maybe something like 'monocoque'
03:01 rbuels oh, wait.
03:01 rbuels i just spelled it for you.
03:01 * rbuels hides in one of deafferret's rolls
03:02 rbuels ok, i closed leto's bug, it's merged in
03:02 pyrimidine ok
03:02 * rbuels deletes the topic/leto_interpro_parse branch
03:03 pyrimidine you've already added him to the team, correct?
03:03 rbuels yeah
03:03 pyrimidine good
03:03 pyrimidine after 1.6.2 we can proceed with a bit of experimentation on core -> modularizing
03:05 rbuels i think it would be best done by writing script that parses some kind of hand-editable specification file of where everything goes, then does all the file moving, dist creation, etc etc automatically.  otherwise you will be in git mv hell.
03:06 pyrimidine dukeleto had raised the idea of using git submodules, not sure if that would work here
03:06 rbuels if i can find some tuits, i would certainly not be shy about working on this, but the tuits are scarce.
03:06 pyrimidine same here
03:06 * rbuels knows nothing about git submodules
03:07 pyrimidine will be using quite a few for the release(s)
03:07 pyrimidine me neither
03:07 pyrimidine (re: submodules)
03:07 rbuels pyrimidine: we might think about using Dist::Zilla for the new dists.
03:07 pyrimidine okay
03:07 pyrimidine I'm up for anything to make releases easier
03:08 rbuels pyrimidine: i have only poked at it a little bit, but what i've done with it so far looks very good
03:08 rbuels pyrimidine: and it seems popular ...
03:08 * rbuels shrugs
03:08 rbuels anyway, i need to go
03:08 pyrimidine k
03:09 rbuels pyrimidine: let me know if you need me to clean up any messes in those Assembly tests
03:09 pyrimidine will do
03:09 rbuels pyrimidine: do you have the tuits to clean them up?  i don't want to foist them on you
03:09 pyrimidine may look at them, if I have the time tomorrow
03:09 rbuels in reality, i'm just being a whiner about installing deps
03:09 pyrimidine but if not, pop a message on IRC if you intend going that way
03:09 rbuels ok.
03:09 * rbuels has to go
03:09 rbuels ttyl
03:10 pyrimidine o/
05:32 bag joined #bioperl
06:09 CIA-95 bioperl-live: Florent Angly master * rbe8dd8d / Bio/Assembly/IO/ace.pm :
06:09 CIA-95 bioperl-live: Removed warning when Bio::Assembly::IO::ace parsing quality strings
06:09 CIA-95 bioperl-live: starting with a whitespace - http://bit.ly/9JlxGn
07:21 CIA-95 bioperl-live: Florent Angly master * r783e04c / (Bio/Root/IO.pm t/Root/RootIO.t): Bio::Root::IO insert() with no backup file - http://bit.ly/anvkTO
11:50 Guest11142 joined #bioperl
11:51 deafferret purl: rbuels is also a bot fly larvae with a penchant for jhannah's flabby parts
12:48 genehack rbuels: I'll pop in here from time to time today, if you have a chance to talk about stuff I could be helping with...
13:15 Guest20981 joined #bioperl
14:01 brandi joined #bioperl
14:01 brandi left #bioperl
14:43 brandi joined #bioperl
14:44 brandi left #bioperl
16:02 jhamilton joined #bioperl
16:09 JunY joined #bioperl
17:42 dukeleto hola
17:45 * rbuels waves
17:46 * rbuels waves at genehack
17:47 deafferret <John Stewart as Queen Elizabeth>Halohhhh!</>
18:28 kblin dukeleto: so we meet again.. :) hi there :)
18:29 dukeleto kblin: hola
18:30 kblin dukeleto: coming to the mentor summit this year?
18:32 rbuels kblin: you know him?
18:33 kblin #gsoc :)
18:33 * rbuels nods
18:35 dukeleto kblin: probably. did I meet you there last year?
18:35 kblin I was there in 2009 and 2010, but I think we chatted in 2009
19:54 pyrimidine joined #bioperl
20:03 * pyrimidine thinks it's funny NCBI is stil advertising a course they no longer support financially
20:03 pyrimidine *still
20:03 pyrimidine http://www.ncbi.nlm.nih.gov/Clas​s/PowerTools/eutils/course.html
20:12 deafferret eutils is dead?
20:14 pyrimidine no, just the training
20:15 deafferret if enough people sign up it pays for itself, right?
20:15 pyrimidine well, the problem is they used to sort of tour universities to run the training
20:16 pyrimidine was a victim of budget cuts, I assume
20:22 deafferret oh. ya. the only NCBI events I've been to locally (two) were free to attendees. dunno if the local univ paid a big block or not
20:23 pyrimidine I think it was gov-subsidized
20:44 pyrimidine rbuels: have convinced hilmar to move BioSQL over to github
20:52 deafferret github++
21:09 risha joined #bioperl
21:21 bag joined #bioperl
21:44 bag__ joined #bioperl
21:46 dukeleto pyrimidine: that sounds promising
21:47 pyrimidine looking at possibly doing it this Friday
21:48 pyrimidine same migration tool used for bioperl (thanks, btw for that)
21:57 CIA-95 bioperl-live: Duke Leto master * rbcde3be / (MANIFEST MANIFEST.SKIP): Update MANIFEST and MANIFEST.SKIP - http://bit.ly/ajLwsi
21:58 CIA-95 bioperl-live: Chris Fields master * r29cf230 / Bio/Tools/EUtilities/EUtilParameters.pm : allow POST for ID lists longer than 200 or terms longer than 300 (with optional setting); beginnings of carryover, but NYI - http://bit.ly/9zNRYk
22:00 CIA-95 bioperl-live: Chris Fields topic/eutils_carryover * r29cf230 / Bio/Tools/EUtilities/EUtilParameters.pm : allow POST for ID lists longer than 200 or terms longer than 300 (with optional setting); beginnings of carryover, but NYI - http://bit.ly/9zNRYk
22:22 ank joined #bioperl
22:36 pyrimidine left #bioperl
23:06 brandi joined #bioperl
23:06 brandi left #bioperl

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