E.g., if I read you correctly, it's not necessary to unlink the request here:
case PROTO_INTERNAL: cbdataFree(fwdState); internalStart(r, e); return; correct? See modified patch below.
miss_access_slow.patch
Description: Binary data
Thanks again, On 29/06/2009, at 11:12 AM, Alex Rousskov wrote:
On 06/29/2009 12:30 AM, Mark Nottingham wrote:OK.WRT requestlink, I was unsure what would happen if the request didn't goforward; will moving requestLink at the bottom up to here: fwdState->request = r; /* requestLink? */ do it?Yes, provided you adjust the code to call "unlink" if the request did go forward but not all the way through to where fwdState currently unlinks.Alex.On 27/06/2009, at 3:05 AM, Alex Rousskov wrote:On 06/16/2009 09:57 PM, Mark Nottingham wrote:Thanks. Anybody else have a second to look?Please s/fwdStartFoo/fwdContinue/ and document what it is. Since this isSquid2 you do not have to do it, of course.Your cbdata and request manipulations appear technically correct to me.IMHO, the temporary lack of requestLink is poor style that will be dangerous for future code modifications. Cheers, Alex.On 11/06/2009, at 11:28 PM, Amos Jeffries wrote:Mark Nottingham wrote:Would someone mind taking a quick look at this patch: http://www.squid-cache.org/bugs/attachment.cgi?id=1989and telling me if I've royally stuffed up with managing fwdState andrequest linking? It's to make miss_access a slow lookup...Looks okay to these uneducated eyes. Probably best to wait forsomeone else to double-check before a HEAD commit, but IMO it looksgood enough for a patching. This one is long-awaited by many. Thanks. Amos -- Please be using Current Stable Squid 2.7.STABLE6 or 3.0.STABLE15 Current Beta Squid 3.1.0.8 or 3.0.STABLE16-RC1-- Mark Nottingham m...@yahoo-inc.com-- Mark Nottingham m...@yahoo-inc.com
-- Mark Nottingham m...@yahoo-inc.com