Closed Blaisorblade closed 11 years ago
Why ignore the test failures?
I modified the commit to not enable test failures. The current problem is that continuous integration will keep complaining about the same failures, so we'll end up ignoring them. But the more reasonable solution is to fix or disable individual tests.
I would like to run this script locally, but with cabal-dev instead of cabal. Can you add some magic (environment variable? cli parameter?) to enable this?
Done. ./allTests.sh
still works and uses cabal, but you just need to set CABAL=cabal-dev
in the environment to have your desired result.
Example (under bash):
$ CABAL=cabal-dev ./allTests.sh
(The above sets the environment variable mentioned only for the duration of the command).
Merged together with Tillmann. This blocked #39.
As discussed by email, the script for continuous integration testing should be on the repo to be version-controlled and to allow developers to run it locally. Fix #26.