Re: [gentoo-dev] RFC: git-r3, additional clone types

[hasufell]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Michał Górny:
> Dnia 2014-02-24, o godz. 22:40:34 hasufell <hasufell@g.o>
> napisał(a):
>
>>> Mode upgrades and downgrades ----------------------------
>>>
>>> Mode is not associated persistently with a repository.
>>> Therefore, using a repository in a different mode than it was
>>> used before (e.g. due to different ebuild or user preference
>>> change) results in mixed-mode repository. [...]
>>
>> I am not sure if I understand this correctly, but it sounds like
>> it would be safer to rm -rf the old checkout completely when
>> switching modes? I wouldn't expect people to randomly switch
>> around, so why bother.
>
> I think a semi-random switching may occur when two different
> ebuilds use the same repository and have different 'minimal' modes
> set. For example, ebuild A uses 'git log' magically, while the
> other is completely fine with shallow clone.
>
> Otherwise, we'd require users to ensure sync of modes between
> different ebuilds using the same repo.
>

Wouldn't it be best to lock that repo to the highest minimal mode of
those ebuilds? Maybe with some kind of lock file. But actually I feel
this is over-engineering a theoretical problem. The only cases I know
of shared repositories between ebuilds is when they have the same
maintainer who is aware of the situation.
-----BEGIN PGP SIGNATURE-----

iQEcBAEBCgAGBQJTDI10AAoJEFpvPKfnPDWzvBIH/jcTtTUHsCqqx3rEdoBGZc74
lKE/UIw9ew7+3FcuMVrp++GAP9ro+6kxyDcpz7dJzntiiW3wSjHadEx9S5arQqJZ
qkKx8fHAAR0tOVxP+jERWcETv5PCF6GYI29VLI5IdzX4gAlXxFBrJCow7es+dbJM
QNQKH9xpFNWz2TJHw2xkf4GWwlQbtpqlGuPfTMQ4MTHDEeJJF6h8bm3QgfaLxl3h
LHfjHmbmGtSVHKI6ieK48WEQZX0wXntbAzne15nLYE04VuDowfPrsjHdleDgfz5T
DU3zbIt6ZqLMqErnAVH9chopUVqK/P63czmCrenAwrJ4DozvEZGKeoPnz9IIZdk=
=WtMs
-----END PGP SIGNATURE-----