DOPEWARS 2.1 UPDATE

We have to assume that maintainers might even not be aware of this failure in those cases. Till noticed that there's a large number of packages which failed to build in the F28 mass rebuild but no bug was opened for them. retire packages which have a last successful build with. It's too late to retire packages for F28 right now. We should to this after the mass rebuild in every release cycle. In both cases bugs MUST be opened, and we should wait 4 weeks, and retire packages automatically if those bugs are still in NEW state. they FTBFS in a mass rebuild and are are uninstallable because of missing dependencies fc27-or-earlier as the lass successful build) a few weeks after the F29 mass rebuild we would retire all packages that have. I would like to propose a more nuanced approach: We voted to retire them after 6 weeks, but I now think this would be premature. For those that succeeded, I think we should follow the course described in the previous paragraph and close bugs.įinally, we have a large group of packages which still fail to build. Some are still ongoing, but so far the tally is 79 successful, 565 failed. I fired off koji builds for all packages on the list in #1555378. If maintainers of individual packages want to create updates, that's of course OK.)

(That applies only to any kind of automated updates. There were good reasons to have the mass rebuild and it certainly would be better to use the rebuilt packages, but we are almost at final freeze and it's better to avoid churn. My reasoning is that if we were to automatically create bodhi updates for all those packages, without maintainer oversight, we risk regressions. After some consideration, I think that we should just leave those packages alone and close the FTBFS-in-F28 bugs for them.

DOPEWARS 2.1 UPDATE

There's a large group of packages for which there was a successful build for F28, but no update was ever submitted. I didn't differentiate between updates there were already pushed to stable and not, because that'd have required additional scripting and didn't seem important enough. I also closed bugs for those packages where a successful build for F28 was done and an update exists in bodhi, but the maintainer forgot to add the bug number to update description. 4 packages were listed erroneously, because either they were freshly added and do not have an F28 branch (rust-hyper), or they were unretired during the move to pagure ( ).

Dopewars 2.1