| 1 |
-----BEGIN PGP SIGNED MESSAGE----- |
| 2 |
Hash: SHA256 |
| 3 |
|
| 4 |
On 06/04/13 10:33 AM, Alexis Ballier wrote: |
| 5 |
> On Fri, 05 Apr 2013 22:18:22 -0400 Ian Stakenvicius |
| 6 |
> <axs@g.o> wrote: |
| 7 |
> |
| 8 |
>> |
| 9 |
>> Revbump -- very important in this case, as the slot-operator dep |
| 10 |
>> (iirc) does not take effect to allow sub-slot-triggered until |
| 11 |
>> after a version with the slot-operator has been emerged. |
| 12 |
>> |
| 13 |
>> So we want users to re-emerge packages either at the same time |
| 14 |
>> libpng-1.6 hits the tree, or beforehand so that they will be |
| 15 |
>> triggered for rebuild when libpng-1.6 hits. |
| 16 |
>> |
| 17 |
> |
| 18 |
> |
| 19 |
> so we force two rebuilds instead of one ? |
| 20 |
> |
| 21 |
|
| 22 |
Either we time it so they just rebuild at the same time (ie when the |
| 23 |
package is unmasked or stable-keyworded), or we commit the revbump |
| 24 |
earlier which would force a superfluous rebuild. |
| 25 |
|
| 26 |
Technically there could (in theory at least) be a way of rewriting the |
| 27 |
vdb to have the correct sub-slot entries (i had a script that did this |
| 28 |
during my EAPI=4-slot-abi testing) but this isn't particularily safe |
| 29 |
and would probably cause more issues for end users than the libpng-1.6 |
| 30 |
bump WITHOUT slot-operatored rdeps. |
| 31 |
|
| 32 |
|
| 33 |
-----BEGIN PGP SIGNATURE----- |
| 34 |
Version: GnuPG v2.0.19 (GNU/Linux) |
| 35 |
|
| 36 |
iF4EAREIAAYFAlFiChMACgkQ2ugaI38ACPBtYAD/YWxxprE1szh2meJVBt16Q+8x |
| 37 |
it+AvHEbMiRetCHchoUA/R0Nkw0Tg6zrx0jO/RgA5U4/H6GGWZUO27VYz8TFu5ae |
| 38 |
=yIRc |
| 39 |
-----END PGP SIGNATURE----- |
Archives