Re: [RFC] Replacement for "assert" in the libraries

I think the most common crashes come from the path functions, some of
which don't have an svn_error_t * return.

Our philosophy has been that passing invalid arguments to a function is
the same type of error as passing "23" as a pointer value. This
philosophy breaks down a bit for the extra-language bindings (where it's
possible to pass a non-normalized path but impossible to pass an invalid
pointer), so there's a legitimate argument for changing it, but no
simple way to do so for functions which don't return errors.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-06-13 18:47:01 CEST