Failures on Overnight build sweep

[Chris Johns]

On 2020-04-06 22:06, Christian Mauderer wrote:
> I checked it: Chris pulled that one in together with the reverted commit
> from Sebastian today at about 7:00 UTC. So both patches are included in
> current RSB master.
> 
> Sorry for missing that step. It's something normally not really
> necessary during normal development (except if we really want to have a
> commit for every RTEMS commit) and I'm still not aware of a lot of stuff
> that is touched by the release.

I think the rules to guide us are ...

If anyone pushes a patch to rtems.git or rtems-libbsd.git branch 
5-freebsd-12 the corresponding RSB configs need to be updated. If you 
think the patch does not warrant an update of these configurations then 
please do not push the patch and please consider not posting the patches 
for review. This helps Joel, Gedare and I figure out where we stand in 
getting the release stable so a branch can be made.

There is a small window of difference between the git builds from the 
RSB and a release. The RSB git builds use the hash in the config file 
while a release uses the sources held in the release URL and the release 
scripts that create the release source packages for rtems.git and 
rtems-libbsd.git use the branch's HEAD. This is the key reason we 
currently need to keep the RSB configs at the HEAD.

Once we branch these restrictions can be relaxed so the sooner this 
happens the sooner the restrictions can be relaxed.

Thanks
Chris