eo
locale is overloaded between tests and pseudolocale generation
#6873
Milestone
eo
locale is overloaded between tests and pseudolocale generation
#6873
Description
After #6406, we expect
NEVER_LOCALE = "eo"
to be, well, never available. It's how we test worst-case scenarios for the locale resolver, such as if an instance's default or even only configured locale loses support. This has been enforced in CI, including onfreedomofpress/securedrop@develop
, since May 2022.After #6690, however,
eo
is available as the "pseudolocale" used to catch unmarked (untranslatable) source strings and test variable string widths. This has been available from Weblate onfreedomofpress/securedrop-i18n@i18n
since January 2023.In effect, we've been operating in a partition between tooling and tests: the one violates the other's assumptions, but they only come into contact during the release/QA/localization cycle.
Steps to Reproduce
Expected Behavior
Tests pass.
Actual Behavior
Failures:
Comments
There are a few moving parts here:
UnknownLocalError
on for Occitan (oc
) #6869 will let us address the confusion I raised in enable automatically-generated "pseudolocale" #6690 (comment) but never addressed: The values ofNEVER_LOCALE
and the pseudolocale should each (independently) be able to be whatever we want.i18n
branches and/or repository should run full CI workflow #6876The text was updated successfully, but these errors were encountered: