(5) This single-use function doesn't need to exist. If you initialize your variables in the constructor then it's more obvious locking isn't required either...
ms::BasicSurface::initialize_attributes()
(6) I think the TODO comment is still valid and should not be deleted:
395 - /*
396 - * TODO: In future, query the shell implementation for the subset of
397 - * attributes/types it implements.
398 - */
Needs^H^H^H^H^HWants fixing, somewhat. I haven't gone back through in enough detail to approve yet but won't block any more.
(4) Hmm, the protocol change uses "repeated" which actually appears to be backward compatible, so no ABI bump for libmirprotobuf is required after all; /developers. google. com/protocol- buffers/ docs/proto# updating
https:/
(5) This single-use function doesn't need to exist. If you initialize your variables in the constructor then it's more obvious locking isn't required either... e::initialize_ attributes( )
ms::BasicSurfac
(6) I think the TODO comment is still valid and should not be deleted:
395 - /*
396 - * TODO: In future, query the shell implementation for the subset of
397 - * attributes/types it implements.
398 - */
Needs^H^ H^H^H^HWants fixing, somewhat. I haven't gone back through in enough detail to approve yet but won't block any more.