On 10/05/2016 08:33 AM, Hardik Dangar wrote: > One more thing can you tell me if we are already violating http via > options like nocache, ignore-no-store ignore-private ignore-reload, why > can't we do the same for Vary header ? We can, but ignoring Vary requires more/different work than adding another refresh_pattern option. Vary is not a refresh mechanism so different code areas need to be modified to ignore (but still forward!) Vary. Also, a good implementation would not just ignore Vary (because it is likely to break things on the client side) but support _replacing_ it with a less restrictive one (as far as Squid internal interpretation is concerned) while still forwarding the original value. One could argue that only such reinterpreting implementation should be officially accepted. If you want to work on that, search Squid sources for httpMakeVaryMark to find starting points. The reply_header_replace code may also be useful as far as reinterpret_vary configuration is concerned. Posting a detailed proposal and securing "acceptance in principle" on squid-dev before starting this potentially controversial work is a good idea if you care about official acceptance. Alex. _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
