| 1 |
El lun, 08-02-2016 a las 12:12 -0600, William Hubbs escribió: |
| 2 |
> As one of the maintainers of sys-fs/udev, I am very conflicted about |
| 3 |
> this. |
| 4 |
> |
| 5 |
> I tend to agree with Kent that we need to be absolutely sure before |
| 6 |
> we |
| 7 |
> switch the default that eudev will maintain feature parity with udev, |
| 8 |
> now and in the future, e.g. when a new release of udev hits, a new |
| 9 |
> release of eudev must happen asap that supports all of the features |
| 10 |
> of |
| 11 |
> the new udev. |
| 12 |
> |
| 13 |
> I also think mgorny's arguments against doing this must be |
| 14 |
> considered. |
| 15 |
> |
| 16 |
> Thanks, |
| 17 |
> |
| 18 |
> William |
| 19 |
> |
| 20 |
|
| 21 |
I agree with mgorny on this. |
| 22 |
|
| 23 |
Specially I wonder about how fast the fixes from udev will get merged |
| 24 |
in eudev :/. That can be not a problem currently as most non-systemd |
| 25 |
users are using the splitted udev (a bit like the setup other |
| 26 |
distributions like ubuntu were using, using udev from systemd as if the |
| 27 |
default provider for udev was sys-apps/systemd even for running with |
| 28 |
openRC). |
| 29 |
|
| 30 |
But this can be more problematic when most non-systemd users will be |
| 31 |
using a forked udev based on an older version that, then, won't contain |
| 32 |
all the available bugfixes. For example, currently, if I don't |
| 33 |
misremember eudev-3.1.5 is like systemd-220 and, then, it won't include |
| 34 |
most fixes since 220 to 228. |
| 35 |
|
| 36 |
Please take care I am not blaming on eudev maintainers or something |
| 37 |
like that, it's simply that I think it's safer from a user point of |
| 38 |
view (that relies on the hardware being recognized and working properly |
| 39 |
on a first install) to ensure they get a newer/fixed version than one |
| 40 |
that can have problems to be kept on sync. |
Archives