| Summary: | Forking all 'stable' profiles into 'testing' variants to support stable USE masking | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Michał Górny <mgorny> |
| Component: | Eclasses | Assignee: | Python Gentoo Team <python> |
| Status: | RESOLVED WONTFIX | ||
| Severity: | enhancement | CC: | base-system, dilfridge, pacho, qa |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| See Also: | https://bugs.gentoo.org/show_bug.cgi?id=439130 | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
Hmm, maybe qa@ or base-system@ has any comments on that? Well, we would not have needed stable masking for that. Unnecessary complication caused by bureaucracy and inertia. IMHO the "canonical" way would be to introduce a new set of profiles 11.0 which requires EAPI=5, and deprecate the 10.0 profiles. (What do we have that version number 10.0 for anyway.) (In reply to comment #3) > IMHO the "canonical" way would be to introduce a new set of profiles 11.0 > which requires EAPI=5, and deprecate the 10.0 profiles. > > (What do we have that version number 10.0 for anyway.) As a side note, however, this would still not allow for using the stable mask files in the base profile and in the main directory... (the same way as your solution). The Council has decided to add new profiles with EAPI=5. |
The idea is quite simple: 1) take all of our major profiles for stable arches, and consider them 'stable' profiles, 2) fork them into 'testing' sub-profiles, likely with a base 'testing' profile, 3) ..., 4) PROFIT! The stable keyword users will still use the current ('stable') profiles, while ~arch users will be encouraged to switch to the testing profiles. The use: when stabilizing a package with optional dependencies which we don't want to go stable, we just mask them in the base (stable) profile and unmask in the 'testing' profile. Now, who shall we CC?