On Thu, Jun 25, 2009 at 02:55:15PM +0200, Greg Stein wrote:
> Shouldn't those arguments be 'const' ?
Yeah they should be.
Don't blame me, I didn't write that code, I just moved it :P
I'll consider changing the function's parameter anyway as
outlined in the log message (to const char *buf and apr_size_t len).
Stefan
> On Thu, Jun 25, 2009 at 13:52, Stefan Sperling <stsp_at_elego.de> wrote:
> >
> > Author: stsp
> > Date: Thu Jun 25 04:52:42 2009
> > New Revision: 38193
> >
> > Log:
> > Add an assertion to guard against invalid use of svn_subst_detect_eol().
> >
> > * subversion/libsvn_subr/subst.c
> > (svn_subst_detect_eol): We should consider converting this function to
> > take a pointer and a buffer size argument, instead of two pointers which
> > define the buffer's boundaries. For now, just add an assert which if
> > triggered indicates a bug in the caller.
> >
> > Modified:
> > trunk/subversion/libsvn_subr/subst.c
> >
> > Modified: trunk/subversion/libsvn_subr/subst.c
> > URL: http://svn.collab.net/viewvc/svn/trunk/subversion/libsvn_subr/subst.c?pathrev=38193&r1=38192&r2=38193
> > ==============================================================================
> > --- trunk/subversion/libsvn_subr/subst.c Wed Jun 24 14:43:19 2009 (r38192)
> > +++ trunk/subversion/libsvn_subr/subst.c Thu Jun 25 04:52:42 2009 (r38193)
> > @@ -1657,7 +1657,10 @@ svn_subst_find_eol_start(char *buf, apr_
> > const char *
> > svn_subst_detect_eol(char *buf, char *endp)
> > {
> > - const char *eol = svn_subst_find_eol_start(buf, endp - buf);
> > + const char *eol;
> > +
> > + SVN_ERR_ASSERT_NO_RETURN(buf <= endp);
> > + eol = svn_subst_find_eol_start(buf, endp - buf);
> > if (eol)
> > {
> > if (*eol == '\n')
> >
> > ------------------------------------------------------
> > http://subversion.tigris.org/ds/viewMessage.do?dsForumId=495&dsMessageId=2365242
>
> ------------------------------------------------------
> http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2365258
Received on 2009-06-25 15:16:00 CEST