Re: CVS update: subversion/subversion/libsvn_subr svn_string.c

> (The main casting I see going on is when we cast all those (void *)
> batons and userData arguments back to their rightful types at the
> top of parser and editor callback functions. Those casts are fairly
> numerous, and as far as I understand, they are necessary too.)

Just so that your understanding is correct: they are not necessary for
any reason in C. (Except on very old compilers which wouldn't
understand our function declarations anyway.) They will generate a
warning in C++. We have no particular design goal of making our code
compile under a C++ compiler, but it might be worthwhile to go to
minimal effort not to make such a thing unnecessarily difficult.

I don't have a particular opinion on the subject. I think the clutter
and the explanatory value of such casts are roughly equal, so I'm
happy to have them or not to have them.
Received on Sat Oct 21 14:36:11 2006