[GSoC] GSoC week 1 status report

[Clayton Wheeler]

Hi all,

I've put my first GSoC status report on my project blog:

http://csw.github.com/bioruby-maf/blog/2012/05/13/progress/

(The web version of this has 100% more hyperlinks, but here's a plain text version, too.)

This has been my first half-week of work on my Google Summer of Code project, and it’s off to an exciting start. The first order of business has been to get my development environment together; since I’ve been a microbiology student instead of a programmer for the last year, it’s taken some work. In that process, I’ve ended up making a few open source contributions just to get my tools working the way I want. I’m running GNU Emacs 24 and trying to take more advantage of it than I have in the past. I’ll have much more to say about this in a future post.

I’ve also started working on the BioRuby unit test failures under JRuby, as a way of familiarizing myself with the BioRuby code base as well as the community and its development processes. Right now, JRuby in 1.8 mode is showing 6 failures and 126 errors, which is hardly confidence-inspiring for people considering using JRuby with BioRuby. This is too bad, since JRuby has some definite advantages as a Ruby implementation. After looking into these failures, I’ve broken them down into a few categories:

	• temporary file permissions problems, likely due to some sort of Travis-CI environment issue
	• a bug in JRuby’s implementation of Open3.popen3 which I’m working up a bug report for
	• an odd autoload problem I’ve filed JRUBY-6658 for and sent an accompanying RubySpec patch for
	• a problem with libxml-jruby, which appears unmaintained, for which I’ve submitted a BioRuby patch plus JRUBY-6662
	• and a small test case bug relating to floating point handling, which I’ve submitted apatch for.

Once these are resolved, JRuby should be passing the BioRuby unit tests in 1.8 mode, and closer to passing in 1.9 mode. (There are a few extra failures under 1.9 that I haven’t sorted through yet.)

I’ve also gotten a start on my project itself, creating the bioruby-maf Github repository with a project skeleton and writing my first Cucumber feature for it. This is, in fact, my first Cucumber feature ever. However, I did spend a few cross-country flights reading the RSpec and Cucumber books last week; between that and cribbing from Pjotr’s code I feel like I have some idea what I’m doing. Just assembling that feature has been useful, too, since I’ve had to get several of the existing MAF tools running on my machine. In fact, my test MAF data and the FASTA version of it are courtesy of bx-python, which will be my reference implementation in many respects.

Clayton Wheeler
cswh at umich.edu