On Wed, Jun 28, 2023 at 9:08 AM Ruediger Pluem <rpl...@apache.org> wrote:
>
>
>
> On 6/28/23 2:37 PM, cove...@apache.org wrote:
> > Author: covener
> > Date: Wed Jun 28 12:37:50 2023
> > New Revision: 1910662
> >
> > URL: http://svn.apache.org/viewvc?rev=1910662&view=rev
> > Log:
> > back to original patch in PR66672
> >
> > Since QSNONE will skip splitoutqueryargs, it's where we should fixup the 
> > trailing ?
> >
> > Modified:
> >     httpd/httpd/trunk/modules/mappers/mod_rewrite.c
> >
> > Modified: httpd/httpd/trunk/modules/mappers/mod_rewrite.c
> > URL: 
> > http://svn.apache.org/viewvc/httpd/httpd/trunk/modules/mappers/mod_rewrite.c?rev=1910662&r1=1910661&r2=1910662&view=diff
> > ==============================================================================
> > --- httpd/httpd/trunk/modules/mappers/mod_rewrite.c (original)
> > +++ httpd/httpd/trunk/modules/mappers/mod_rewrite.c Wed Jun 28 12:37:50 2023
> > @@ -3909,10 +3909,12 @@ static const char *cmd_rewriterule(cmd_p
> >      }
> >
> >      if (*(a2_end-1) == '?') {
> > -        *(a2_end-1) = '\0'; /* trailing ? has done its job */
> >          /* a literal ? at the end of the unsubstituted rewrite rule */
> > -        if (!(newrule->flags & RULEFLAG_QSAPPEND))
> > -        {
> > +        if (!(newrule->flags & RULEFLAG_QSAPPEND)) {
> > +            /* trailing ? has done its job.  with QSA, splitoutqueryargs
> > +             * will handle it
>
> I think only if QSLAST is set. Otherwise substitutions or the matching URL 
> may place a '?' left of the terminating '?' that came
> in as encoded '?' and thus would add the original querystring as "&&" + 
> r->args. Hence we would end up with
>
> <one character after some decoded '?' somewhere in the URL or after 
> substitution><further chars>+ "?&&" + r->args.
>
> Hence I think it should be:
>
>        if (!(newrule->flags & RULEFLAG_QSAPPEND)) {
>            /* trailing ? has done its job.
>             */
>            *(a2_end-1) = '\0';
>            newrule->flags |= RULEFLAG_QSNONE;
>        }
>        else {
>            /* with QSA, splitoutqueryargs  will handle it if RULEFLAG_QSLAST 
> is set.
>            newrule->flags |= RULEFLAG_QSLAST;
>        }
>

flipping it around and refreshing some comments:

    if (*(a2_end-1) == '?') {
        /* a literal ? at the end of the unsubstituted rewrite rule */
        if (newrule->flags & RULEFLAG_QSAPPEND) {
           /* with QSA, splitoutqueryargs will safely handle it if
RULEFLAG_QSLAST is set */
           newrule->flags |= RULEFLAG_QSLAST;
        }
        else {
            /* avoid getting a a query string via inadvertent capture */
            newrule->flags |= RULEFLAG_QSNONE;
            /* trailing ? has done its job, but splitoutqueryargs will
not chop it off */
            *(a2_end-1) = '\0';
       }
    }
    else if (newrule->flags & RULEFLAG_QSDISCARD) {
        if (NULL == ap_strchr(newrule->output, '?')) {
            newrule->flags |= RULEFLAG_QSNONE;
        }
    }

close?


-- 
Eric Covener
cove...@gmail.com

Reply via email to