John Peacock <jpeacock@rowman.com> writes:
> It _does_ alters the calling signature of three functions (two to just
> append a pool) and adds a new function. Because of this, it will
> require stub wrappers before it could be considered for 1.1.0. It is
> seems especially clumsy to have to add wrappers to provide a pool
> entry version; would it have been such a bad thing to provide an
> optional pool entry on all/most functions' call stack?
These don't sound like very difficult steps. Stub wrappers are pretty
easy... Is the API-compatibility stuff really causing a problem here,
or just adding a tiny administrative burden to certain coding tasks?
I'm not familiar with the exact technical issue you're discussing,
btw. I guess I'm just asking: what would your ideal compat guidelines
be? The ones we have now seem to me like that address the issues
Subversion needs to address, especially w.r.t. people writing against
the core APIs.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Fri Feb 20 22:01:26 2004