I might be wrong, but isn't the sole purpose of revdep-rebuild to fix missing SLOTs? If it is, then revdep-rebuild shouldn't ever need to be run. Ergo some packages need fixing. And tomorrow is bugday (heck, it's already bugday in some timezones!)... Reproducible: Sometimes Steps to Reproduce: 1. find a package which lacks proper SLOTting and mask newer versions of it 2. emerge a package depending on the package in 1 3. remove the masks you set in 1 and update it 4. run the package you emerge in 2 Actual Results: missing libraries Expected Results: problems when emerging the package depending on a library package in the first place (ie you either shouldn't be able to finish step 2, or step 3 shouldn't have any effect)
No, it is to rebuild that packages that get broken when libraries are upgraded. ie. when a package A is built against library B and B is upgraded then A is broke. revdep-rebuild with detect that and re-merge A.
Additionally, not all packages support being SLOT'ed. In fact, most don't unless upstream intends for it, like gcc, python, etc.
Then shouldn't this be closed as "upstream"?
