Build is broken

[tobie]

No description provided.

[marcoscaceres]

Seems that the processResponse() method has gone AWOL.

[silviapfeiffer]

Ah, good to know it's not just me: https://www.w3.org/Tools/respec/respec-w3c-common is broken, too, right?

[marcoscaceres]

@silviapfeiffer yes, it's broken too.

See, for example: http://picture.responsiveimages.org/

[darobin]

Found something else not covered by the test suite :(

[tobie]

Well, it is covered by the tests, the travis build is failing on master.

[marcoscaceres]

did @darobin do a Marcos and pressed the green button before Travis said it was ok? 😸

[darobin]

No @tobie, the Travis build is failing on the ui branch — which is a feature branch and unstable — but passing on gh-pages (further: the failure on the ui branch is just one of the tests not looking for the right thing and so tripping over the UI changes). I know, I checked. I haven't had a chance to investigate over the weekend, but there is a problem with the build script. That's what's not checked by the test suite: it runs the dev code, but not the build. The first order of business tomorrow morning (after fixing the build) is to change the build script to be able to produce arbitrary builds (not just releases) and then use that from the test script to ensure tests always run against built code (which I should've done from the start).

[tobie]

My bad! That sounds like a great idea.

Fixed the travis status image on the README file in #276.

[darobin]

Well, it's only a great idea in the sense that not running the tests against the built code was stupendously daft in the first place :)

The build is now fixed; and running the command line tests now generates a build (can't easily do the same when launching from the browser).