Do not randomize chosen locale and use a UTF-8 supported locale.
(Closes: #925879, #1004950) Use an Estonian locale, which has unusual sort order compared to many european languages, which may catch some interesting issues. Testing more obscure locales and character sets is certainly desireable, but it is also very important for reprotest to pick the varied locale deterministically. Randomly choosing the locale, or choosing based on a hash of the inputs, might lead to picking a different locale while debugging patches for a potential fix, with misleading results if the issue is highly locale specific.
-
mentioned in merge request !17 (closed)
-
mentioned in issue #4 (closed)
Please register or sign in to comment