Re: [Bug-wget] [PATCH] Allow to redefine ciphers list for OpenSSL

[Tomas Hozza]

----- Original Message -----
> I'm not sure, but looking at the patch, it /does/ seem like it tries
> to override the user settings, which IMO should not happen. If that is
> indeed the case, I do not support this patch either.

Please describe why do you think it "seems like it tries" to override the
user settings? As all it overrides are values hardcoded in the source.
 
> @Giuseppe: About the failing test, that particular test seems to have
> some weird timing problems. I'm assuming it is probably a fault in the
> perl server. It randomly fails for me too. However, if you run make
> check again, chances are the test will pass. :)
> 
> 
> On Sat, Jul 19, 2014 at 3:12 AM, Ángel González <address@hidden> wrote:
> > On 17/07/14 13:49, Tomas Hozza wrote:
> >>
> >> I agree. The patch didn't take any configuration possibility from the
> >> user.
> >> The users would be able to configure whatever in the same way they were
> >> before.
> >>
> >> Please really see some of those patches I sent. The discussion was little
> >> bit confusing at some points ~ like the intentions were interpreted
> >> differently.
> >>
> >> Regards,
> >
> >
> > I still strongly oppose to the patch. If the user configures wget to only
> > use Perfect
> > Forward Security, and your patch makes wget connect to a server not using
> > it
> > you
> > are overriding user configuration (in the weakening direction).
> > See my last email for details.
> >
> > Patch v3 also seem to coalesce the different options of --secure-protocol
> > if
> > using
> > GnuTLS, which IMHO doesn't make sense either.
> >
> > PS: s/cipers/ciphers/ in v3
> >

Regards,
-- 
Tomas Hozza
Software Engineer - EMEA ENG Developer Experience

PGP: 1D9F3C2D
Red Hat Inc.                               http://cz.redhat.com