Brief description of your issue
Directly based on microsoft/winget-pkgs#407968 (comment).
It's safe to say that @Trenly knows more about the codebase factors for this than I do, so I'll mostly quote him. though a qualified guess of my own on a summary, is that it's a lack of official tests in winget-cli for that kind of scenario (Things that involve 2-letter language codes without countries) that necessitates further test environments for them.
More or less direct quotes, though shortened for focus:
- "(…) - I'm not seeing any code coverage in the CLI repo that ensures the two letter tags resolve correctly"
- "(…) - it's not that I'm worried it won't work. As you correctly pointed out - it does seem to be handled. My concern is that it is NOT tested anywhere in the CLI codebase, which makes the implementation undefined behavior
Things like -
- InstallLocaleRequirement: ["en"] - does it pass or silently reject en-US packages?
- ARP producing "en" as InstalledLocale during an upgrade flow is never exercised
- Raw GetDistanceOfLanguage("en", "en-US") - the Windows API score is never asserted in any tests"
And keep in mind, it is in fact supported, with the XnSoft.XnShell package in pkgs having become a prominent case of it.
Brief description of your issue
Directly based on microsoft/winget-pkgs#407968 (comment).
It's safe to say that @Trenly knows more about the codebase factors for this than I do, so I'll mostly quote him. though a qualified guess of my own on a summary, is that it's a lack of official tests in winget-cli for that kind of scenario (Things that involve 2-letter language codes without countries) that necessitates further test environments for them.
More or less direct quotes, though shortened for focus:
Things like -
And keep in mind, it is in fact supported, with the
XnSoft.XnShellpackage in pkgs having become a prominent case of it.