Hi everyone, I'm going to spend the next few weeks slowly bringing over the patches I've written in my parserwork SF branch. I'll start with patches which reduce pressure on the memory allocator and then move onto incorporating the request line pand client-side parser changes I made to Squid-3 a few weeks ago.
Let me know if you're able to review my work and possibly do some tests. The individual patches won't themselves show any appreciable CPU drops but testing I've done with Mark Nottingham shows quite an increase in request rate for small cache hits - 1k responses went up from 6k req/sec to ~9.3kreq/sec. Adrian -- - Xenion - http://www.xenion.com.au/ - Hosting and Commercial Squid Support -
