| 1 |
On 02/04/14 16:01, Mike Frysinger wrote: |
| 2 |
> On Wed 02 Apr 2014 13:01:25 Samuli Suominen wrote: |
| 3 |
>> Problem 1: |
| 4 |
>> |
| 5 |
>> https://bugs.gentoo.org/show_bug.cgi?id=472766#c21 |
| 6 |
>> |
| 7 |
>> I'm not sure if wildcards are supported by /etc/sandbox.d/ files |
| 8 |
> they are not. however, path matching is based on prefixes, so there's always |
| 9 |
> an implicit glob at the end. would be reasonable to change the code to use |
| 10 |
> fnmatch. |
| 11 |
> |
| 12 |
> e.g. SANDBOX_PREDICT=/dev/dri/card probably works |
| 13 |
|
| 14 |
I hope SANDBOX_PREDICT="/dev/dri/card" with "" is OK too? |
| 15 |
|
| 16 |
> |
| 17 |
> however, i think we're relying on sandbox preventing bad code from doing bad |
| 18 |
> things. there really should be a way for the build to disable the logic in |
| 19 |
> the first place from kicking in. |
| 20 |
> -mike |
| 21 |
|
| 22 |
You are right |
| 23 |
I believe this started after a major mesa version bump, so I'd start |
| 24 |
looking for the culprit |
| 25 |
in Mesa's OpenCL code, but I have no idea howto go futher with the |
| 26 |
debugging... yet |
| 27 |
|
| 28 |
Meanwhile, =media-gfx/imagemagick-6.8.8.10[opencl] now installs the |
| 29 |
sandbox.d file, |
| 30 |
workaround is better here than nothing since this is affecting multiple |
| 31 |
binaries, |
| 32 |
packages :/ |
| 33 |
|
| 34 |
- Samuli |
Archives