Camelia, the Perl 6 bug

IRC log for #bioperl, 2010-06-03

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

All times shown according to UTC.

Time Nick Message
02:07 brandi joined #bioperl
02:07 brandi left #bioperl
02:28 flu_ joined #bioperl
02:28 buelsbot joined #bioperl
02:28 JunY joined #bioperl
03:00 dnewkirk joined #bioperl
09:01 flu_ joined #bioperl
09:10 bag_ joined #bioperl
09:52 bag_ joined #bioperl
10:46 JunY Hi, deafferret. Just saw your yesterday's messsages. Thx a million. I am learning github forking :D
11:11 kblin JunY: I'm sure there's a couple of folks around who chan help if you run into git problems
11:26 JunY yep, that is good. I am learning github now. so far, so good ~
12:43 brandi joined #bioperl
12:46 brandi left #bioperl
12:47 JunY Hi, rbuels. are you here?
12:48 JunY I assume I should finish all the coding work first, at least the coding work for a package, then I can upload my code uisng github.
12:48 JunY or, can I upload them by weeks?
12:52 * JunY slaps JunY around a bit with a large trout
13:11 JunY \me is enjoying a good cup of coffee :D
13:12 * JunY is enjoying a good cup of coffee :D
13:44 * deafferret waves
13:46 deafferret JunY: I tend to commit everything every couple of hours. That way I don't lose my work to a typo or power failure. You can push up to github daily or weekly so others can see what you're up to
13:47 deafferret don't know how picky rbuels is about super-clean-commits
13:47 * dnewkirk guesses rbuels is barely awake
14:40 kblin well, you can push to one branch regularly, and have a "review" branch that's squeaky clean
14:41 deafferret or just commit to your local constantly, then push a single big commit (rebase)
14:41 deafferret I *think* that's how that works
14:42 deafferret & commute
16:19 rbuels JunY: no, you should commit your work as you write it
16:20 rbuels JunY: in your fork of github
16:20 rbuels JunY: in a normal work day of coding, i usually commit between 10 and 40 times (depending on what is going on)
16:46 deafferret rbuels: then a rebase and push? or just push all 40?
16:46 rbuels deafferret: usually i rebase before pushing
16:46 deafferret so that means all 40 commits become a single commit, right? that's the point of rebase?
16:47 deafferret or... does rebase ensure that the push doesn't move 'master' back in time (undoing other commits) ?
16:47 rbuels deafferret: no, the main point of rebasing is to rebase.  meaning that you change the commits to be based on the latest upstream commit, so that you don't generate an extra 'merge' commit
16:47 rbuels deafferret: but, rebase can also merge commits and such
16:48 rbuels deafferret: sometimes i merge and clean up commits.
16:48 * deafferret remains unclear on best practice given lots of code reviewers
16:49 rbuels really, the best practice for review in git is to look at the diff of the branch versus master
16:49 rbuels i find.
16:49 deafferret I heard that Linux (for Linux kernel) wants *all your weeks of work in one commit, dammit* so it's easier to review
16:49 * rbuels finds an example
16:49 deafferret s/Linux/Linus/
16:49 rbuels deafferret: well, the linux kernel is a little odd, they are basically packaging the whole branch in one commit.  which is valid.
16:49 rbuels but i don't think it's necessary for us to do that
16:50 rbuels deafferret: here's an example
16:50 rbuels deafferret: go to http://github.com/bioperl/bioperl-live and hit 'Branch list'
16:50 rbuels deafferret: then hit 'Compare' next to my longest_orf branch
16:51 rbuels deafferret: and you can see the commits i've done on the branch, plus the diff of all the stuff that has been done on there
16:52 deafferret so once you merge that to master those 14 commits will interleave into all other commits in history, right?
16:52 rbuels deafferret: yeah, that's true
16:53 deafferret so 5 people each do 100 commits in 5 branches over a month, then once merged those 500 commits all look like they were happening at the same time in history
16:53 deafferret that's cool, if that's what people are expecting  :)
16:53 rbuels deafferret: well, you can see the branches in gitx or gitk or whatever
16:54 deafferret ya. works for me. thanks
16:55 rbuels deafferret: but of course, if developers want to squash commits together and clean up history, it's nice
16:55 * rbuels shrugs
16:59 deafferret rbuels: how would one do that if one were so inclined?
16:59 rbuels deafferret: mostly git rebase -i
17:11 JunY Thx got it~
17:15 JunY So if I am right, I work on the clone (bioperl-live) files. Then I upload it to my own fork every few minutes or hours
17:15 deafferret you    git commit;   every few hours. you    git push;    once or twice a day or whenever
17:18 JunY I just write "git commit", right?
17:18 rbuels JunY: read this: http://drupal.org/node/299067
17:18 rbuels JunY: it's a simple introduction to what version control is, how you usually use it, and why it's important
17:19 rbuels JunY: you need to read the git documentation
17:19 rbuels JunY: and understand what commits are, what branches are, what tags are,
17:19 rbuels JunY: what the parent of a commit is,
17:19 rbuels JunY: etc.
17:19 rbuels JunY: version control systems like git are the most important tools in software development.
17:19 deafferret rbuels is the parent of the commit of my heart
17:19 rbuels JunY: so you need to learn them well
17:20 JunY ok, great
17:20 rbuels JunY: very very well.
17:20 * rbuels rebases deafferret's heart onto a chicken
17:21 JunY Or, can I do like this?  I write my own code in a seperate file. Test it. Then paste it to the bioperl-live. Then upload it?
17:22 rbuels JunY: no.
17:22 rbuels JunY: work directly in bioperl-live.
17:22 rbuels JunY: use git.
17:22 rbuels JunY: git diff.  git status.  git commit.
17:22 rbuels JunY: learn what all those commands do.
17:23 JunY Yes Sir!!!
17:23 rbuels JunY: also, read the bioperl guide to using git
17:23 JunY yep, but the bioperl guide is not complete ...
17:23 JunY But I think the link you give me is really good. I will learn that
17:23 rbuels JunY: no, it kind of assumes that you understand the basics of git already
17:23 JunY ok, i see. I will read them
17:24 JunY BTW, how are other students' working progress? Do they have so many questions like I do? :D
17:27 rbuels JunY: you are actually being very good about communicating, which is good
17:28 JunY thx. it is hard to see other students' working progress... their blogs didnot update for the last few days...
17:29 rbuels JunY: yeah i know.  i need to send out some emails about that.
17:29 JunY Anyway, I will keep going. I am not very happy with my working progress. But Now everything is getting clear. I have started the real coding work now, hopefully, I will reach the milestone within the next two weeks...
17:30 rbuels JunY: learning git is the foundation.  it will be very difficult to make good progress until you learn it well
17:32 JunY yep, I get it. Then I should spend some time on learning github. It is just for the last few weeks, all my time was to learn things, I cannot wait to write some code now... My hands are getting itchy about it :D
17:33 JunY THx for your advice, rbuels. I need to go home for dinner now. C ya next time!
17:33 rbuels JunY: github and git are different things.
17:34 rbuels JunY: github is just a site that hosts git repositories and provides some web tools for them.
17:34 JunY ok, i see
17:34 rbuels ok good night
17:37 deafferret what timezone is JunY in?
17:38 rbuels i think he's in ireland
17:40 deafferret Aye Begora!
17:40 * deafferret kisses rbuels' Blarney stone
17:41 * deafferret has exhausted all his Ireland references    :o)
18:20 dave_messina joined #bioperl
18:25 kai joined #bioperl
18:25 kai joined #bioperl
18:33 dnewkirk joined #bioperl
18:48 buelsbot joined #bioperl
19:13 deafferret the rebirth of the buelsbot!
19:14 rbuels which does nothing
19:14 rbuels lol
19:21 bag_ joined #bioperl
20:13 deafferret i am disappointed that uselessness amuses you
20:30 * dnewkirk promises himself to never use ubuntu again
20:31 deafferret dnewkirk: wut? why?
20:31 * deafferret hugs Ubuntu
20:34 dnewkirk Always run into little bugs, in part because I switched to 10.04 LTS without enough testing
20:34 * dnewkirk prefers rpm based distros
20:34 * dnewkirk hugs Fedora 13
20:34 * flu hugs his Ubuntu while giving his openSuSe an evil glare
20:35 * deafferret pees on Solaris
20:35 dnewkirk ).)
20:35 dnewkirk 0.0
20:36 dnewkirk Solaris is amazing
20:36 deafferret though $this_ubuntu_box IS suffering some weird memory leak right now
20:36 deafferret dnewkirk: I've only been exposed to poorly admined 5+ years out of date Solarii that make my life icky
20:37 deafferret same for AIX
20:37 dnewkirk ahh
20:37 * deafferret prefers Windows (choke!) to old Solaris
20:38 kblin dnewkirk: I've yet to find a system that just works :)
20:38 deafferret OS X?  :)
20:38 flu 40 seconds, not bad
20:39 kblin deafferret: only got a 10.4 PPC, that misses prototypes from some of the header files
20:39 kblin so compiling programs is an awful lot of fun
20:40 deafferret oh. ya. I buy the latest OSX and stuff Just Works  :)
20:40 kblin right...
20:40 deafferret not cheap, but less suffering
20:40 kblin as long as stuff == stuff apple wants you to use
20:40 kblin deafferret: 10.6 still ships a 3.0.x samba, right?
20:41 deafferret dunno. never done samba on OSX
20:41 deafferret i can check if you tell me what to type
20:41 deafferret lappie is <---- over there
20:42 kblin smbclient --version should get it
20:43 deafferret OS X 10.6.3  smbclient 3.0.28a-apple
20:43 kblin fun
20:43 deafferret sarcasm?   ;)
20:43 kblin that's ancient
20:43 deafferret why use samba?
20:43 kblin but arguably the last pre GPLv3 version
20:44 kblin some people work with windows-using biologists :)
20:44 deafferret what-using what-now?   ;)
20:44 kblin and for some reasons, they expect me to share data with them
20:44 flu they should be shown the true path of enlightenment
20:44 deafferret tell them security protocols don't allow that
20:45 deafferret "call me when you have source code for me to fiddle"
20:45 kblin flu: "come over with your USB pen drive"?
20:45 deafferret "leave me alone I'm napping (and drunk)"
20:46 dnewkirk I used to be a Windows using biologist :)
20:46 kblin anyway
20:46 kblin I'm unconvinced of that "best os" thing
20:47 deafferret my beliefs do not require that you be convinced
20:47 deafferret hmm... Matrix quote fail
20:47 deafferret almost, but not quite
21:07 spekki01 so im having some issues with bioperl-hmmer3, the basic rundown is i can't seem to get a full description when i try to parse it out of a hmmsearch result file here is the basic code http://codepad.org/9DH2tfAA ive stripped out a bunch for simplicity. But the first part is the hmmsearch result file with full descriptions, next part is the code i use for getting the description (line 35), and the last part is what it looks like when i print
21:07 spekki01 out those variables.
21:08 spekki01 am i grabbing the description wrong?
21:10 kblin hm, hang on
21:12 spekki01 i tried looking at it myself and seeing if i could figure out the code that did it but im not familiar with it enough to know where to look i think :(
21:16 kblin you found a parser bug
21:16 dnewkirk out of curiosity, what would happen if you plug in my @description?
21:16 kblin the parser seems to stop reading the description on the first space
21:16 spekki01 lemme give it a go one sec
21:17 kblin I've added a test case to check for this
21:17 dnewkirk thanks kblin
21:19 kblin it's going to be a bit involved to fix, the logic to parse a hit line is to split on " "
21:20 kblin so I need to fix this to use something a bit more clever
21:32 kblin spekki01: got a fix for you, I think
21:33 spekki01 kk cool, ive been trying that @description but i dont think it works
21:33 spekki01 its seems to give me the same truncated line
21:34 kblin yeah, the parser stops reading at the first space of the description
21:34 spekki01 just out of curiosity what code did you go and edit to fix the issue?
21:36 spekki01 I wanna try have a go at fixing bugs i find as a learning experience so i get more familiar with the bioperl modules and how they work
21:40 kblin http://github.com/bioperl/bioperl-hmmer3/com​mit/52c3bf8f258aea09f23b1ea9b0e71c7609073f0e
21:41 spekki01 ok checking it out
21:44 spekki01 those changes are already commited right, so i can just use git to get that whole module again.
21:46 kblin yes
21:47 kblin I've also just applied the same fix to hits below the inclusion threshold. not sure if those already work, but I'd rather be consistent
21:47 kblin thanks for finding that one
21:48 spekki01 np thanks for fixing it, one day ill get to fixing them and just subbmitting a pull request to you guys
21:48 kblin looking forward to that :)
21:49 rbuels spekki01++ # heart o' gold
21:50 spekki01 lol just paying it back, this channels saved my ass so many times might as well help back when when im able.
21:52 deafferret kblin++
21:53 deafferret http://bioperl.org/wiki/IRC
21:53 deafferret kblin: you should add yourself  ^
21:53 deafferret spekki01: you should add yourself ^^
21:54 deafferret kblin: you need a gravatar for github.com  :)
21:54 spekki01 sure
21:55 kblin deafferret: I thought I had one
21:55 kblin it's just lousy :)
21:55 * deafferret squints
21:56 spekki01 i don't think ive seen one on you kblin
21:56 deafferret you da invisible man?
21:56 kblin it's a countour detection algorithm run on my real portrait, I'll add a real one tomorrow :)
21:57 deafferret awesome hat required   http://github.com/jhannah
21:58 deafferret hmm... so this is dead then?   http://github.com/kblin/bioperl-hmmer3
21:58 deafferret since you're committing straight to bioperl/ version now?
22:02 kblin deafferret: the plan was to post stuff like bug-fixes straight to bioperl/bioperl-hmmer3 and stuff I'd like to get review on to my fork
22:03 deafferret ah. your git-fu is strong
22:04 spekki01 lol
22:05 spekki01 if only i could say that stuff in RL and people would get it and laugh
22:05 deafferret kblin++ # irssi army, unite!
22:05 kblin but it's trivial to keep my personal repo updated
22:10 kblin anyway, bedtime for me. see you around tomorrow, folks
22:15 deafferret g'night
22:15 deafferret sudo kblin Control-a d
22:26 spekki01 night all
22:31 dnewkirk night
22:51 ama joined #bioperl
23:07 at5 joined #bioperl
23:07 at5 h

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