Last modified: 2014-07-28 09:39:24 UTC
I noticed the test suite for Scribunto skip two tests because they can't find the file NormalizationTest.txt which comes from Unicode. We can get the file installed on Jenkins server either: - by installing the `unicode-data` package which provide a bziped version of it, maybe use puppet to decompress it - add the uncompressed NormalizationTest.txt file to integration/jenkins.git so it would ends up being available under /srv/deployment/integration/slave-scripts/ In both cases, it would be helpful to have a $wg variable to let us override the current hardcoded __DIR__ . "/NormalizationTest.txt"
Note actually including this file raises the number of tests run for Scribunto from 1378 to around 38000. Also, if I recall correctly, Scribunto (like MediaWiki) passes against 6.0.0, and not necessarily against the normalizations added in later versions of Unicode.
> Also, if I recall correctly, Scribunto (like MediaWiki) passes against 6.0.0, > and not necessarily against the normalizations added in later versions of > Unicode. Ubuntu Precise ships Unicode data 6.0.0 for now. So it seems we can reuse it.
*** Bug 68643 has been marked as a duplicate of this bug. ***