Improving the handling of PR:s
dan at rucci.org
Sat Jan 12 06:46:03 PST 2008
Peter Schuller wrote:
> * The committer may not have access to the hardware, or may not have a
> software setup that allows for testing. This means doing such testing
> suddenly requires a lot more effort.
Would it be possible to identify PRs which are related to specific
hardware vs PRs which are not related to hardware?
With respect to hardware specific PRs (Since these sound like they can
be especially troublesome if the right hardware isn't in the right hands)
- user A uploads his dmesg somewhere (if that output is enough)
- user A also submits his email address and requests for notifications.
- a hardware specific PR & Patch comes in with a committer who doesn't
have the necessary hardware to test the patch
- user A get a fun email stating the above with instructions on how
he/she can help.
- user A applys the patch and run the test case or whatever.
- user A sends some feedback.
- enough of this happens and the committer has more testing behind the
patch; the set of all users with this hardware rejoice.
Otherwise, I can't see how A can help out with these patches unless they
are actively searching gnats for each piece of hardware they have.
More information about the freebsd-current