Re: apr_off_t is of an ambiguous size.

On Tue, Jan 13, 2004 at 04:31:19AM -0500, Russell Yanofsky wrote:
> I agree that svn_filesize_t should be used instead of apr_off_t in the diff
> and blame apis.
>
> I the functions that wrap around apr calls for error handling would be
> better off using APR types. But if you're only going to change
> svn_io_file_seek, then I guess it's not really a big deal. That's as a good
> a place as any to do bounds checking on platforms where APR doesn't do 64
> bit i/o.

We're not even wrapping svn_io_file_seek. As far as I'm concerned it
could stay as apr_off_t but then we're just pushing the typecasting up
one level.

As far as I can see the only interface we export that takes apr_off_t
and passes it on to an actual APR function is svn_io_file_seek. So it
would be the only place where we'd need bounds checking if we moved to
an always 64-bit int for our offsets.

-- 
Ben Reser <ben@reser.org>
http://ben.reser.org
"Conscience is the inner voice which warns us somebody may be looking."
- H.L. Mencken
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org