John Snow <js...@redhat.com> writes: > On 5/18/21 5:28 AM, Markus Armbruster wrote: >> QAPIError not having .info don't actually exist before this patch. > > It's defined by QAPISourceError now, I just missed this spot in > test-qapi. It isn't used in practice until now, however.
I had QAPIError mentally filed under abstract types / didn't bother to formally make it one with decorators. Just as well, because it's not staying abstract: this patch creates instances.
