On my machines the test suite complains that a couple of the restraints
defined in python are missing the virtual functions for version and
last_modified_by. This patch adds those.
r345. But I'm curious to know what warnings/errors you're seeing. The
test suite runs on all of our build machines without complaining
(although I guess it shouldn't, without these implementations). So I
wonder what's different about your machines.
On a separate note, it is not clear to me that anyone will ever bother
to actually update these fields. Especially since they are defined
primarily in the headers and code changes are most likely to occur in
the .cpp files. Is it possible to use a SVN hook to update them
automatically?
For stuff in the kernel, it's probably not worth ever updating them -
except perhaps to bump the version number every time we cut an IMP
release. I think the value is only for third party extension modules -
and they, of course, can update them whenever they choose (or not all, I
guess).