On 10/14/2015 01:15 AM, Markus Armbruster wrote: > Eric Blake <ebl...@redhat.com> writes: > >> Commit ac88219a had several TODO markers about whether we needed >> to automatically create the corresponding array type alongside >> any other type. It turns out that most of the time, we don't! >> >> As part of lazy creation of array types, this patch now assigns >> an 'info' to array types at their point of first instantiation, >> rather than leaving it None. > > I'm afraid this flips the value of .is_implicit() to False. Currently > harmless, but let's keep it correct anyway. > > The obvious fix is to define the trivial override method: > > def is_implicit(self): > return True > > But I'd rather do *all* the "give implicit types info" work in "qapi: > Track location that created an implicit type", i.e. move the plumbing of > info there, add the override method there, drop the "As part of" > paragraph from the commit message here. I append what's left of this > patch then. I like it, because the patch that actually changes > generated code (this one) becomes really simple, and the lengthened > patch remains mere info-plumbing that doesn't affect the generated code.
Makes sense, so I agree with how you've redone the current state of qapi-next. -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
