On Thu, Mar 28, 2019 at 11:40 PM Yann Ylavic <[email protected]> wrote:
>
> On Thu, Mar 28, 2019 at 10:30 PM Ruediger Pluem <[email protected]> wrote:
> >
> >
> >
> > On 03/28/2019 10:27 PM, Yann Ylavic wrote:
> > > On Thu, Mar 28, 2019 at 10:24 PM Ruediger Pluem <[email protected]> wrote:
> > >>
> > >> On 03/28/2019 05:39 PM, [email protected] wrote:
> > >>>
> > >>> +    for (wpos = str; *str; ++str) {
> > >>>          if (!quoted) {
> > >>> -            if (**last == '\"' && !ap_strchr_c(sep, '\"')) {
> > >>> +            if (*str == '"') {
> > >>
> > >> Question: Is the token allowed to the quoted?
> > >
> > > Hmm no, I asked and was told to RTFM, and then forgot :)
> > > Will fix, thanks!
> > >
> >
> > All good. It is not allowed to be quoted and then the code behaves 
> > correctly and returns with APR_EINVAL.
>
> Not really, 'Cache-Control: "private"' for instance is not allowed but
> accepted here.
> Hopefully fixed in r1856507.

Oh, you were right actually, the quoted token was already caught by
the first "skip separators" checks.
But I find the new code from r1856507 clearer anyway, so I'll leave it
(unless I'm the only one)..

Regards,
Yann.

Reply via email to