After a few investigation, I found the statement from http://www.squid-cache.org/Doc/config/ecap_service/: vectoring_point = reqmod_precache|reqmod_postcache|respmod_precache|respmod_postcache This specifies at which point of transaction processing the eCAP service should be activated. *_postcache vectoring points are not yet supported. Also in http://wiki.squid-cache.org/Features/ICAP, similiar statement was found: Pre-cache REQMOD and RESPMOD vectoring points are supported Notice 6.1 Vectoring points from rfc5703 suggests 4 classess of different adaption. I guess the above statemets is class 1, client requests "on its way into the cache", and class 3, responses "on its way into the cache"? A positive answer might be really bad news for me, since I am looking for class 4, client-specific responses coming from the surrogate.. Would anyone please make me clear? Thanks, Kaiwang 2011/10/18 Kaiwang Chen <kaiwang.chen@xxxxxxxxx>: > Great! That should be what I am looking for. > > Thanks, > Kaiwang > > 2011/10/18 zozo zozo <flam4@xxxxxxx>: >> eCAP can do the trick. http://wiki.squid-cache.org/Features/eCAP >> Implementation samples are here http://www.e-cap.org/Downloads >> >> You can cache whole response, then calculate what you need and send it. >> >> 18 октября 2011, 17:13 от Kaiwang Chen <kaiwang.chen@xxxxxxxxx>: >>> Hello, >>> >>> There is some requirement that squid replaces certain pattern in a >>> HTTP response(assign some cookie, or substitues certain pattern) with >>> something calculated from the response instead of fixed string. Is >>> there any know way to do this? If negative, please points to some >>> further reading for me to do some hack. >>> >>> Thanks, >>> Kaiwang >>> >
