Re: svn commit: r1242397 - in /subversion/trunk/subversion/libsvn_subr: skel.c stream.c

Greg Stein <gstein_at_gmail.com> writes:

> I don't think this is the correct approach. svn_stringbuf_t is
> *designed* to put a NUL at the end of the public length. Thus, it is
> supposed to properly manage the +1 inside its functions.
>
> The correct fix is to put a ++minimum_size into svn_stringbuf_ensure()
> rather than make callers worry about space for the private NUL
> character.
>
> Please revert this commit. Code should not have to compensate for
> stringbuf's internal concept.

We would be changing long standing behaviour if we did this. I'd like it
to be possible as it would remove the inconsistency between the current
svn_stringbuf_ensure and svn_stringbuf_create_ensure. I wonder if there
are is any code that relies on the existing behaviour? It seems
unlikely that there is any real world code that would care.

-- 
uberSVN: Apache Subversion Made Easy
http://www.uberSVN.com