This week in open source

Mike Burns


shoulda is version 3.0.0! Someone needed to cut it, and Joe Ferris (jferris) stepped up to make the release (4f0c363).


A new version of high_voltage is out this week, with contributions from Matthew Horan (mhoran). Matt fixed the content_path hack on PagesController such that you can subclass and override the content path properly (0d84952, 9c7c0ad, 8884f70, and d4ba056, e4b1ce9). He also added Rails 3.1 compatibility (7556fcd). Joe Ferris (jferris) pushed version 1.1.1 out (09c2da4, a60cc5b, and a1f2476).


What’s new in version 1.4.0 of bourbon? Glad I asked myself, in writing. Phil LaPier (plapier) changed buttons from display: inline to display: inline-block, removing text decoration on anchor buttons (8bea3a7). He executed the “extract mixin” Bourbon refactoring pattern on render-gradients (9fd50d1), moving the linear and radial gradients to the root (92ecae3), all so he could support more gradients in the border-image mixin (0f91303). Transparent fallback colors are now allowed (b409982 and 5d86ebd), a new user-select mixin was added (b00d334), and finally Phil bumped the version (f9ee911).


A new version of paperclip is out—version 2.7.0 was pushed by Mike Burns (mike-burns) to Rubygems (760a4e0).

Mike—that’s me—flipped out at a failing test and reverted the :swallow_stderr support that was pull in last week ([04c9514]( “Revert "Paperclip#run now respects swallow_stderr setting. Fix for #741”

This reverts commit 58671ebfffe3b03177d9ad248737932441fe4016.

I am reverting this because the tests broke and because the code itself looks hideous.“)). The tests didn’t work and the code was ugly. As a general request: if you’re adding code to an ugly codebase, add beautiful code!

Prem Sichanugrist (sikachu) updated the docs to point out that awk-sdk must be a specific version (22a2a85). He also added support for Rails 3.2 (77c5268).

Tom Hughes (tomhughes) noted that the fingerprint field is not cleared when an attachment is deleted, though all other fields are…and he fixed that (189ee9c). Pat George (pcg79) finally fixed the failure message for validates_attachment_size, which was written very incorrectly, for years (45b7dac).


Nothing new on kumade. Gabe Berke-Williams (gabebw) stopped ignoring the Gemfile.lock, despite what Yehuda suggests (a2a2278); simplified the name of a scenario (ffd1aef); and set up Travis CI to run a full integration test (04d9f6b).


The shoulda-matchers project had a very ancient list of contribution guidelines so I, Mike Burns (mike-burns), updated them to reflect the fact that pull requests do not actually suck (81cd02b).


More tiny commits on bourne from Gabe Berke-Williams (gabebw), where he added Travis CI support (55248d6) and fixed some code formatting (f31aa72). Yup.


The only news for capybara-webkit this week is that Joe Ferris (jferris, a.k.a. JƒF) does not want bug reports about system-specific compilation issues (6b724ef). We simply can’t fix them.


Nothing new on clearance, but Dan Croak (croaky) did get it working under Travis CI (3430e88).


Tiny cleanups on appraisal as Gabe Berke-Williams (gabebw) changed some parens around and removed some of those awful if modifiers (7a5af59), and Brian Cardarella (bcardarella) updated the README to use actual English (1883b80).