Last modified: 2013-07-18 21:36:34 UTC
We have a browser test for the ULS IME feature that contains a Malayam text string. https://gerrit.wikimedia.org/r/#/c/72365/ The test passes when run against test2wiki. But the linter in Jenkins reports a bogus failure and will not merge the branch to master. https://integration.wikimedia.org/ci/job/qa-browsertests-ruby1.9.3lint/423/console Can this be forced, or the linter updated or bypassed in order to merge this code?
Seems like a ruby 1.9.3 issue :] $ ruby1.8 -c uls_ime_steps.rb Syntax OK $ ruby1.9.3 -c uls_ime_steps.rb uls_ime_steps.rb:46: invalid multibyte char (US-ASCII) uls_ime_steps.rb:46: invalid multibyte char (US-ASCII) uls_ime_steps.rb:46: syntax error, unexpected $end, expecting keyword_end ...ge).input_method_enabled == 'ഇൻസ്ക്രിപ്?... ... ^ $ We are using the version from Ubuntu Precise: *** 1.9.3.0-1ubuntu2.7 0 When copy pasting the sequence above, I do not have that issue: $ irb1.9.3 irb(main):001:0> puts 'ഇൻസ്ക്രിപ്' ഇൻസ്ക്രിപ് => nil Turns out ruby 1.9 is properly supporting unicode \O/ It just need an hint about the encoding being unicode and not US-ASCII: # encoding: utf-8 puts 'ഇൻസ്ക്രിപ്റ്റ് 2 That pass :) Followup on Gerrit change