| 1 |
On Sat, Dec 29, 2012 at 6:54 PM, Alexander Berntsen |
| 2 |
<alexander@××××××.net> wrote: |
| 3 |
> -----BEGIN PGP SIGNED MESSAGE----- |
| 4 |
> Hash: SHA256 |
| 5 |
> |
| 6 |
> All packages should have local descriptions of what the bindist |
| 7 |
> USE-flag specifically does. This should be a policy when writing |
| 8 |
> ebuilds that include it. |
| 9 |
> ... |
| 10 |
> I have filed bug reports for the affected packages. However, to avoid |
| 11 |
> this becoming an issue in the future I think it would be fruitful to |
| 12 |
> have a policy or something to make sure all ebuilds that make use of the |
| 13 |
> bindist USE-flag at the very least describe *what* toggling it does in |
| 14 |
> the context of that specific package, and preferably also *why* the |
| 15 |
> flag exists in the first place. This will summarise the problem and |
| 16 |
> the solution that USE=+bindist offers for the end-user. I don't find |
| 17 |
> it acceptable to make the details "left as an exercise" to the user. |
| 18 |
|
| 19 |
Once all packages currently using the flag have local descriptions, we |
| 20 |
could remove bindist from profiles/use.desc, converting it from a |
| 21 |
global to a local use flag. This would cause repoman to reject any new |
| 22 |
instances without a local description. |
| 23 |
|
| 24 |
That would be much more effective than a policy that may be |
| 25 |
(accidentally) ignored. |
Archives