On Mon, 2010-10-04 at 18:12 +1300, Amos Jeffries wrote: > On 20/09/10 00:41, Andrew Beverley wrote: > >>>>> I've moved it next to the headers check. I have also removed the error > >>>>> message that was generated if they don't exist. However, this means that > >>>>> if somebody explicitly sets --with-netfilter-conntrack and the libraries > >>>>> don't exist, then it will silently fail. Is this the behaviour we want? > >>>> > >>>> Hmm, we at least want to MSG_NOTICE for both cases, with preferrably a > >>>> hard error if its explicitly stated. > >>> > >>> There'll be the default AC_SEARCH_LIBS notice in any case, and then it > >>> will also be shown later assuming --enable-zph-qos is set. I've just > >>> realised though that by default the QOS functions are disabled. I > >>> thought the new concept was that everything was enabled by default? > >>> Should I change the default to enabled for --enable-zph-qos? > >> > >> Yes please. > > > > Now enabled by default. > > > > I have also decided that (as previously suggested) a miss option would > > be useful in addition to the value preservation for a miss. This allows > > a miss value to be set when Squid has been compiled without > > libnetfilter-conntrack, and also makes it easier to set a miss value if > > you're happy with a consistent value. I have therefore added this into > > the latest patch (attached). The parameter is called 'miss' and it takes > > precedence over the preserve-miss feature. > > > > I've not heard back from my 2 testers yet, and I am yet to use it in > > anger myself, but I will give feedback on all these once I have it. > > > > Andy > > > > Any news? >
One of the testers has just now compiled a build with the patch in, so I should hear back from him shortly; the other I have not heard from. I am yet to use it in a production environment myself, but hope to do so soon. > Unless someone has a reason not to I'm going to commit this when the > current trunk build issues are resolved. I certainly wouldn't have any objections :)
