Re: svn commit: r12074 - branches/svnserve-ssl/subversion/libsvn_ra_svn

On Mon, 2004-11-29 at 14:49, Mark Benedetto King wrote:
> On Mon, Nov 29, 2004 at 02:08:35PM -0500, Greg Hudson wrote:
> > I'm not terribly happy with the idea of adding timeout and I/O-pending
> > operations on all streams, when they're only going to be used in very
> > specific contexts. Specific functionality should not be driven deeply
> > into commonly-used layers of a system.

> I agree, but I didn't feel that this case was clear-cut. The run-time
> cost of the additional function pointers is virtually nil, and the
> implementation does not complicate the existing code.

Oh, the resource cost is entirely secondary. The real cost is that
someone reading about svn_stream_t is presented with grot which is only
relevant to a particular type of stream, and which is only used by a
very specific piece of code.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Mon Nov 29 21:25:28 2004