Richard Brežák

Noy really, what #251 gets us is a all or nothing support for access control. Ideally one wouldn't have to run 2 instances for public and provate repositories.

I put this under the future milestone as for this to really make sense, #176 needs to be in place. If there are any objections to this, let me know.

This is probably blocked by #173 but I'll verify if it can or cannot be done without it.

Anything more than this is blocked by #173 and further discussion is needed, as buildbot does not allow partial UI replacements as far as I can tell, therefore we cannot...

This is superseded by #255 as now evaluation failures appear as separate builds. 

Ah yeah if the build doesn't fail it doesn't show. That's my mistake.

Known bug, expected and harmless failure, but yes that system could be improved

We need it to be configurable per repo, so it falls under the "per reps settings" thing which ive mentioned in the past. Also we could just ignore any eval...

This seems like a infra/nix issue not buildbot-nix issue. There may be a way to detect this class of failures and to retry automatically though.

I think as a long term thing, handling this case would be nice. Computers suck, stuff happens, sometimes temporarily, would be nice to not involve the human, let the computer...