Camelia, the Perl 6 bug

IRC log for #bioperl, 2013-02-05

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

All times shown according to UTC.

Time Nick Message
00:05 MohsenT hi
00:33 MohsenT rbuels: hi
09:38 carandraug deafferret, following our talk last week, here's the patch https://github.com/bioperl/bioperl-live/pull/51
11:18 heeeeelp joined #bioperl
11:18 heeeeelp hello!
11:19 heeeeelp soooo, can someone please point me towards the best bioperl guide on t'web?
11:20 heeeeelp Also, why is strawberry Perl better than the normal Activestate?
11:22 heeeeelp .....and, how do i check wether bioperl is installed? I though i had done it, but 'BioPerl' is not a recognised commmand....'
11:51 heeeeelp i think need to install something called nmake, but the link is dead
12:18 heeeeelp also, after running 'perl Build test, you say 'All the tests should pass, but if they don't let us know" So, 3/320 failed - just letting you know as requested
12:34 heeeeelp Ok. I now have both perl and bioperl installed. I can confirm that perl is working, but i amnot sure how to confimr wether bioperl is working. I have have followed the guides that say create a file with this in: '#!/bin/perl -w   use Bio::Seq;' However, i get the following line (which makes me think that the bipperl installation has failed: 'Can't locate Bio/Seq.pm in @INC (@INC contains: C:/strawberry/perl/site/lib C:/s trawberr
12:34 heeeeelp etc,etc
12:35 heeeeelp do i need to be IN a ceratin folder to run bioperl? do i need to somehow map a path to the bioperl modules?
12:35 heeeeelp *certain
14:01 rbuels joined #bioperl
14:04 heeeeelp Hello?
14:49 rbuels joined #bioperl
16:27 MohsenT joined #bioperl
16:43 deafferret carandraug: sweet! looking
17:02 deafferret rbuels: can you make carandraug a collab in github? I appear to not have the power
17:02 deafferret any reason these other 3 pull requests are stalled? https://github.com/bioperl/bioperl-live/pulls
17:03 rbuels deafferret, carandraug: by the power of grayskull, i have the power.
17:04 * rbuels smashes open the github teams management page, inserts carandraug
17:04 rbuels it is done!
17:04 deafferret huh. i feel bad about these other 3 pull requests... i think I'm going to ram them through to
17:05 deafferret o
17:05 deafferret with the cake
17:45 carandraug deafferret, rbuels : thanks
17:45 deafferret thank YOU!
17:46 carandraug what should I do about the variable WIDTH that I mention on the comments?
17:49 carandraug my idea would be to remove them, and instead define the defauls with a hash on _initialize. Should make the parsing of parameters at least 1 line shorter and not make it accessible to the user.
17:50 carandraug also, if WIDTH is meant to be configures, then I should create BLOCK
18:01 deafferret carandraug: sounds like you know what you're doing. proceed.  :)
18:02 * deafferret doesn't have any useful opinions on these things, since he hasn't touched a fasta file for a couple of years
18:04 carandraug deafferret, it should only make a difference if Bio::SeqIO::fasta::WIDTH (it's a variable, not a method), is meant to be accessible by users. Is it?
18:05 carandraug the only usefulness of accessing it would be to set a default for all of the objects created rather than reusing one object
18:05 carandraug or so it seems to me. But bioperl is such a large thing, I'm afraid to touch it too much
18:06 deafferret carandraug: if there's no documentation claiming "use WIDTH like this" then I'd say it's an internal thing not intended for user input.
18:06 deafferret and if there's no tests demanding specific widths then I wouldn't worry about "breaking" it.
18:07 deafferret until someone points out a problem, at which time tests should be added
18:07 deafferret $0.02
18:07 deafferret "default for all the objects" sounds like a bad design. I like your plan of exposing it as an explicit (optional) argument for any given object
18:08 deafferret (I prefer a per-object API, not a class/package default)
18:10 looper joined #bioperl
18:13 carandraug deafferret, how much documentation do I need to look for to assume there's none claiming that? The POD documentation of the module is enough?
18:13 carandraug deafferret, also, can you guide me on running the tests?
18:19 deafferret carandraug: ~/src/bioperl-live$ prove -Ilib -r t/SeqIO
18:19 deafferret is how I ran the tests
18:22 deafferret at a glance I don't see any WIDTH stuff except what you just added?  ~/src/bioperl-live$ ack --perl -i WIDTH
18:23 carandraug deafferret, I didn't add it. It was already there
18:25 deafferret ah, right. sorry. you added block().  So width() IS documented. Are there any tests...? looking
18:26 carandraug deafferret, also, should we have checks that those values are indeed numbers and integers? And complain when a user tries to set invalid parameters? I can do "Bio::SeqIO->new(-format => 'fasta', -blok => 78)" and nothing will happen, not even warning message. Note the typo on blok
18:26 deafferret I see ZERO tests.
18:26 carandraug deafferret, width() the method is documented. WIDTH the variable is not I think
18:26 deafferret carandraug: that's univerally true of BioPerl. So unless you what to tackle that "problem" writ large, don't worry about it
18:27 deafferret carandraug: sure -- the important thing is that you don't break the documented use cases. Which you didn't, afaict. Implementation should be improvable at will.  :)
18:27 carandraug deafferret, but is it by design? Because parsing all the options would slow down the code?
18:28 deafferret carandraug: yup. I would consider a global BioPerl change to explode on unknown arguments an improvement. Others might disagree.
18:28 deafferret that would be a long running thread in bioperl-l, if you want to open that can of worms :)
18:29 deafferret ... if I were tackling that I'd tackle it with Moose, which would not be bioperl-l -- it would be Biome -- a Moosey rewrite of BioPerl
18:29 deafferret py<TAB>++
18:29 deafferret https://github.com/cjfields/biome
18:29 carandraug deafferret, I do not have the time to go add tests to all parameters of all parameters
18:30 deafferret carandraug: ya, you wouldn't have to. You'd just add 5 tests showing successful exceptions being thrown in a few cases
18:30 carandraug all parameters of all packages I mean. But I could add them as I write them
18:31 deafferret carandraug: no, i.... do you have skype?
18:31 carandraug deafferret, when I mean tests for parameteres I mean input cheking
18:31 deafferret this is a lot more typing than I really want to do right now ;)
18:31 deafferret where are you? I'm central US.
18:31 carandraug deafferret, no, I do not have skype
18:31 carandraug deafferret, Ireland
18:32 carandraug deafferret, I don't even have a cam or mic. I use a old desktop that the university was going to throw away
18:33 deafferret ah. so .... if you changed BioPerl fundamentally to throw exceptions whenever invalid arguments are given, you wouldn't add 30,000 tests for that, you'd just add 5 or whatever to show that its working as expected in the case of direct class arguments and inherited arguments.
18:33 deafferret ... this gets into the whole "everything in BioPerl is an Interface" quagmire, which you may or may not like as a design decision
18:34 deafferret carandraug: http://www.bioperl.org/wiki/Advanced_​BioPerl#Rationale_of_Interface_Design
18:35 deafferret (Biome, in contrast, doesn't do that. To its credit, imho.)  :)
18:35 * carandraug is reading
18:36 deafferret If you're pro-Moose than you might consider Biome a jillion times cleaner than bioperl-l
18:37 deafferret if you're pro-perl6 than you might consider bioperl6 an adventure in the right direction of the future of bio* in perl :)
18:37 carandraug deafferret, I don't really have a preference. I haven't coded in perl in some time. Been doing more image processing lately in Octave. And in Octave, the top of each function is to validade every single parameter
18:39 carandraug deafferret, sometimes takes me a longer to validate all the input than to write the actual function
18:40 deafferret yup. Moose thinks it helps accelerate development by handling that sort of scaffolding Correctly so you don't have to worry about it
18:41 carandraug well, that's not really slowing down the development since it's not done at all
18:41 carandraug I though the problem was the gazillion of packages in one project
18:44 deafferret well, "problem" depends on who you ask. -shrug-   Yes, the monolithic nature of BioPerl has been bemoaned for years.
18:44 deafferret that's another effort currently underway
18:45 deafferret http://www.bioperl.org/wiki/Google_Summ​er_of_Code#Major_BioPerl_reorganization
18:45 carandraug deafferret, I always thought that the problem was bioperl-live. Then I read that it was already split into bioperl-live, db and run
18:45 deafferret err... 2011 at least.  :)
18:46 deafferret umm... there are 32 parts so far: https://github.com/bioperl/
18:46 carandraug deafferret, what happened with that GSoC project?
18:46 deafferret dunno. you might want to ask rbuels
18:46 carandraug deafferret, yeah. But none is officially released is it?
18:47 deafferret I consider github.com "officially released."  :)
18:47 deafferret pushing master to CPAN is another project. time consuming, someone just has to step up and do it every however often then volunteer to do so.
18:48 rbuels carandraug, deafferret: it made quite a lot of progress in splitting some things out
18:48 carandraug deafferret, someone should say that to debian packagers
18:48 rbuels carandraug, deafferret: Bio::FeatureIO was originally split out by that project i think
18:48 deafferret rbuels++ # dy-NO-mite!!
18:49 carandraug rbuels, but Bio::FeatureIO is in both bioperl-featureIO and bioperl-live. If I have a fix, where should I apply it?
18:49 rbuels carandraug: guesss.
18:49 * rbuels chuckles
18:50 rbuels carandraug: bioperl-featureio
18:50 carandraug rbuels, then can I remove those from bioperl-live?
18:51 rbuels erm ...
18:51 rbuels carandraug: ask pyrimidine
18:52 carandraug I find the huge codebase and the project fame, a bit intimidating
18:52 * deafferret votes for "add your fixes to bioperl-featureIO, remove any bioperl-featureIO code from bioperl-live, makes sure all the tests in both repos still pass"
18:53 rbuels or that.
18:53 rbuels i would check with pyrimidine to see where he is with it though
18:53 rbuels here, i'll write the email
18:55 rbuels sent
18:56 carandraug rbuels, I think we sent at the same time
18:56 rbuels oh well
18:57 * rbuels guarantees his email is uglier
18:57 carandraug though I sent to the whole mailing list
18:57 rbuels that's fine too i guess
19:45 audy left #bioperl
19:47 audy joined #bioperl
19:47 audy left #bioperl
19:50 carandraug is 5.010 features acceptable?
19:51 deafferret uh... that's probably a questions for the mailing list
19:51 deafferret if it's not in there already
19:51 carandraug deafferret, I want to use state
19:51 deafferret or discussed already
19:51 deafferret you can search the mailing list archives -- I assume that's come up a few times over the last few years
19:54 deafferret ffs bioperl-live/Bio/Root/Test.pm
19:54 * deafferret boos
19:56 carandraug deafferret, what?
19:56 deafferret libraries should be in lib/ not in the root of repos
19:56 * deafferret boos
20:09 sl33v3_ joined #bioperl
20:13 * rbuels watches amused as deafferret and carandraug explore the layers of cruft
20:14 * deafferret shrugs
20:15 deafferret personally, I'd probably spend my time in Biome nowadays :)
20:15 deafferret or BioGroovy ;)
20:19 deafferret but the opinions of Biotraitors aren't terribly important. I sold the fuck out :)
20:30 ChanServ joined #bioperl
20:37 ChanServ joined #bioperl
20:39 ChanServ joined #bioperl
21:17 MohsenT joined #bioperl
21:34 MohsenT hi
21:34 MohsenT i cant install BioPerl
21:38 rbuels MohsenT: what operating system are you using?
21:39 MohsenT windows
21:39 rbuels lol
21:39 MohsenT kobes server is down
21:39 MohsenT i cant add it to repo
21:39 rbuels well i don't know anything about windows, have you tried following http://www.bioperl.org/wiki/I​nstalling_BioPerl_on_Windows?
21:39 MohsenT yes
21:40 MohsenT it mentions adding repositories from some servers
21:40 MohsenT seems like kobes is down
21:40 MohsenT error 500
21:40 MohsenT not reachable
21:40 MohsenT and manually usung CPAN
21:40 MohsenT using*
21:41 MohsenT gives status 512
21:41 rbuels :-\
21:41 MohsenT it seems it is a couple of days that kobes is down
21:42 MohsenT :(
21:42 rbuels MohsenT: maybe try cygwin?
21:43 rbuels probably be easier.
21:43 rbuels if you know how to operate a unix command line.
21:43 rbuels http://www.bioperl.org/wiki/Installing​_BioPerl_on_Windows#Bioperl_in_Cygwin
21:44 MohsenT hmmm
21:44 MohsenT tnx
21:45 MohsenT i'll try it
21:45 MohsenT :)
21:46 ChanServ joined #bioperl
22:01 deafferret what is kobes?
22:01 deafferret <-- also knows nothing about windows
22:04 rbuels deafferret: some package repository for windows perl packages
22:05 rbuels "what would you like to build this chest of drawers with?"  "cheese!"
22:05 * deafferret reps five with the pimps
22:06 * rbuels likewise has 5 Reputation
22:06 * rbuels equips Robe of Charisma
22:06 * rbuels now has 6 Reputation
22:07 rbuels i put on my robe and wizard hat.
22:07 deafferret Put the Range Rover on them 24's I skate good
22:08 deafferret Twenty junkies beaming up to Scotty in my Enterprise
22:09 rbuels (enterprise!)
22:10 deafferret Took the white and turned it green. Now that's what I call chemistry!
22:20 rbuels deafferret: fun exercise, look up the tail number on the helicopter
22:20 deafferret 420?
22:20 rbuels lol no they couldn't afford to paint a custom number on the helo
22:22 deafferret I only see N128
22:22 deafferret at 2:16
22:23 deafferret oh... N128KR
22:23 deafferret 2:16 + 2:24
22:24 rbuels oh, used to be a helo tour company
22:24 rbuels now it's http://www.helilease.com/
22:25 deafferret http://www.flickr.com/photos/71sbeetle/2289555083/
22:26 rbuels cute little R44
22:26 rbuels there's a heli tour outfit out here in boston that has a black one and an orange one
22:26 rbuels i see them all the time
22:26 rbuels or maybe it's just the orange one they have
22:27 rbuels i don't think i've seen the black one hover and look around at the buildings
22:27 rbuels maybe i should get jordan tower to manage me
22:27 deafferret http://www.eee.dgualdo.it/vari/n128kr-pad-001.htm
22:27 deafferret "unknown damage 28.05.11 "
22:28 deafferret murder, burglary, and battery (presumably)
22:29 deafferret ya, that heli isn't nearly as bad ass when you see the whole thing
22:29 rbuels hehehe
22:31 deafferret http://www.youtube.com/user/jtower?feature=watch - click "Play all". There's only 196 of them.
22:32 deafferret (sadly)
22:33 deafferret oh dear. Brisco needs some gun safety refresher courses
22:33 deafferret http://www.youtube.com/watch?v=bhz142NAqw0
22:33 * deafferret prefers cake songs
22:34 * deafferret always parks his Escalade in his trailer park
22:35 rbuels maybe it's cause brisco can't see.

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