| 1 |
On 01/03/2013 02:09 AM, Dirkjan Ochtman wrote: |
| 2 |
> In any case, we probably shouldn't spend a whole lot of effort on this |
| 3 |
> given the somewhat-impending git migration, which neatly solves this |
| 4 |
> problem. Maybe there's some low-hanging fruit in the commit ordering, |
| 5 |
> though? I was thinking it might be possible to have the Manifest |
| 6 |
> signed before committing the ebuilds, but it's entirely likely that |
| 7 |
> CVS keywords get in the way of that... |
| 8 |
|
| 9 |
The CVS keyword expansion causes the ebuild digest to mutate during the |
| 10 |
commit. If we repoman could predict correctly emulate the CVS keywords |
| 11 |
expansion on the client side, then it could generate a correct Manifest |
| 12 |
in advance. However, that seems difficult given that the CVS keyword |
| 13 |
expansion contains a timestamp with 1 second precision. |
| 14 |
-- |
| 15 |
Thanks, |
| 16 |
Zac |
Archives