This week in open source

Holy pull requests, Batman!

capybara-webkit

Lots of churn on the backend of capybara-webkit. Matthew Mongeau (halogenandtoast) added status codes (558c58b), which meant that the gem now passes another capybara spec (7ea51d5 and 71ca80d). Hongli Lai (FooBarWidget) gave us a working console.log (dc49e5f).

Hongli Lai (FooBarWidget) and Trevor Smith (trevorsmith) attempted to uncover memory and process issues. Hongli started with a patch to use Process#spawn if available (922d197). Trevor then claimed that IO.popen is actually what we need (d8abcbc and c362098). I tried to merge all this in but broke it. Please voice some knowledge on issue #137 so we can move this forward!

laptop

The Linux side of our laptop script got some attention from Ma Yiu Ming Patrick (PatrickMa), who fixed the README to contain accurate instructions (7fa6ede).

flutie

Continuing on the doc fixes theme, flutie saw a typo fix (application.css">6f28f4d) from Emilien Taque (etaque) and a markup fix (26896b2) from Matt Jankowski (mjankowski).

factory_bot1

Some small updates to factory_bot this week, as Brandon Keepers (bkeepers) removed ye olde find(:first), replacing it with just first (d56125d), Jeff Duti (jdutil) fixed a copy/paste issue in the docs (6bc3b87), and Joshua Clayton (joshuaclayton) configured the Travis-CI tests to hit the latest stable patch of each major version of Ruby (17d47bd and 5e501e8).

bourbon

The bourbon Sass plugin continues to improve. Chad Mazzola (ubuwaits) tried it on a new app and submitted his documentation changes based on that (2da2b19). Phil LaPier (plapier) updated the README based on some feedback from Chad (62bbcb2), added a transform-origin property (0943d14), added some amazing CSS3 timing functions (081dd3d), and then released version 0.1.5 of bourbon (ad4dbdc).

kumade

The simple Heroku deployer grows in complexity. kumade got a whitespace fix from its maintainer—remember to leave no trailing whitespace, people! (680fc82). The fantastic Marcos Tapajós (tapajos) took care of a few issues by making a deploy branch if it doesn’t already exist (988f9b3) and making a rake task for consistency (11366db). He also updated the docs to mention bundle exec (f7afe03).

Nick Quaranto (qrush) ran with the rake task and gave us an overridable kumade:before_asset_compilation task (9dda7ee, 691e456, and e613c29). He needs to do special processing before packaging assets, so this helped him out.

Nick then added the Japanese characters for “kumade” to the README before releasing version 0.2.2 of kumade.

cocaine

Our command runner, cocaine, got some mingw32, logger, and doc love: alextk (alextk) gave us more precise unix detection that can better differentiate unix from mingw32 (73c2cb3). Jon Yurek (jyurek) refactored atop that (4df6592 and 8478ed2) before adding a logger (44332c4). He and Marcos Tapajós (tapajos) updated the README (6fb2bf1 and accbc4f) before Jon released version 0.2.0 of cocaine (fb39757).

paperclip

Attaching files just got more tested and documented! paperclip saw a commit from Prem Sichanugrist (sikachu) adding REE to Travis-CI (ab7db5a) and from Chris Harper (chrisharper) adding Rubinius (aca668b).

Hold on, it would be a mistake to overlook Chris Harper’s pull request</. Ladies and gentlemen, the most surefire way to have us take a pull request, even if it causes the tests to break or is something we are not particularly interested in, is to submit a picture of yourself in a monkey suit:

So anyway, Harold Giménez (hgimenez) added a note on how to detach a file (b2a370e, 0a55c41, and e365326).

And Marcin Urbański (murbanski) rounded it off with a new feature: by specifying the :source_file_options you can control the options to convert for input processing (eab19f8).

Project name history can be found here.