James Addison

Question: is there a way to call `as_validator_add_issue` while overriding the default severity for a rule? (idea: `release-time-missing` is usually `error`-level, but we could downgrade it to `warning` for snapshot...

Hm. There's a possibly-unicode-related error appearing in these builds; I did encounter that locally, but suspected that it might be a difference in my local build environment. It appears not....

> Hm. There's a possibly-unicode-related error appearing in these builds; I did encounter that locally, but suspected that it might be a difference in my local build environment. It appears...

> I'm also unsure why the Fedora Rawhide continuous integration build appears as successful here; it should have failed, since the same problem occurred there according to the logs. `continue-on-error`...

> > Hm. There's a possibly-unicode-related error appearing in these builds; I did encounter that locally, but suspected that it might be a difference in my local build environment. It...

My next guess would be that something about the argument-passing related to `va_list` could be problematic. I might try refactoring to remove that.

> My next guess would be that something about the argument-passing related to `va_list` could be problematic. I might try refactoring to remove that. Yes, this appears to have resolved...

@ximion I think that this changeset now adheres more-closely to what you suggested in #570 and here: specifically that although `release-time-missing` should not be reported as an _error_ for `snapshot`...

@ximion my communication on this thread was somewhat verbose, so I'll try to be concise: I'd be glad for any code review on the approach currently coded here (approximately: refactor...

Sorry to nag: is this likely to be reviewable in the nearish future? I'm very much not a competent C++ developer, so if the changes initially seem rough/unacceptable, I'd probably...