On 01/10/2006 02:11 PM, William A. Rowe, Jr. wrote: > Plüm wrote: > >>> William A. Rowe, Jr. wrote: >>> >>>> Ruediger Pluem wrote:
[..cut..] > > > The straightforward thing is to close the client socket. Obviously it's > not that trivial; unix can reuse the same fd almost immediately. Perhaps > close the write side? In any case, the connection should be marked > aborted. > > Number 2 follows from 1, of course we don't 'finish' the response. > > Remember, the back end is *busted* midstream. We have to convey that, but > we don't have to maintain the integrety of every byte sent by the (errant) > server, IMHO. In the case of the broken backend the patch already does this. Do I understand you correctly that the behaviour currently done for a broken backend should be applied to *any* type of error bucket? [..cut..] > as you might have guessed, my comments were aimed at those 'interesting' > applications that were otherwise cacheable - e.g. those folks who keep > arguing that things like mod_authnz_hostname should interact with the cache > (which yes I disagree with, but this would provide basic mechanisms to > handle > this case.) I think we disagree on this as I also belong to those folks, but thats another story and thread :-). Furthermore I do not see right now how this helps in the 'mod_authnz_hostname case' as the response should still be *cached*. It shouldn't be delivered from the cache without checking. But as said this point does not belong to this thread. [..cut..] Apart from getting very welcome improvement feedback I also wanted to check if someone is -1 on the patch. As I regard the "partial pages problem" as something that should be fixed in 2.2.1 I will propose it for backport once Brian did a successful test with the patch and I do not see any -1 on it. Regards Rüdiger
