Wireshark-bugs: [Wireshark-bugs] [Bug 7149] fast conversation lookup

Date: Thu, 19 Apr 2012 12:40:51 -0700 (PDT)
https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=7149

--- Comment #13 from Cristian Constantin <const.crist@xxxxxxxxxxxxxx> 2012-04-19 12:40:51 PDT ---
(In reply to comment #10)
> It's really O(n log n) you forgot about hash tree :-)
> 
> But if it's *really* so expensive (from 20min down to 6seconds), 
> maybe it'd be better to generate/find conversation once, and store it in packet
> structure. mhm.

cristian: jakub, which hash tree exactly?

anyway, the capture I am having is handcrafted to produce one _large_ chain
list behind the _same_ hash bucket. so everything reduces to operations on that
list.
I was curious about the performance in this peculiar case.
what led me to this? some capture with sip packets (generated by a _real_ sip
testing tool) which had exactly this problem as well.

pretty much offtopic: does any of you know why the hash tables in glib do NOT
provide support for handling the chain lists (for collisions)?

-- 
Configure bugmail: https://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are watching all bug changes.