Jakub Zawadzki skrev 2010-04-12 13:38:
> On Mon, Apr 12, 2010 at 01:20:46PM +0300, Kaul wrote:
>    
>> In some cases, wouldn't it make more sense to put the more
>> commonly used string near the top and perform a normal search?
>>      
> Maybe (wireshark is used in so many different environments and smth which is 
> common in network A, might not be in network B)
> I think also it makes small sense to use value_string_fast with less than 16 
> or 32 (or even 64) entries.
>
> Anyway I would keep linear search API and make dissector developer decide.
>
> I thought about binary search when I tried to fix bug #594 [1]
> with value_string array [2]. There're about 5k entries for tcp services.
>
> [1] https://bugs.wireshark.org/bugzilla/show_bug.cgi?id=594
> [2] https://bugs.wireshark.org/bugzilla/attachment.cgi?id=4324
>    
Would it be faster to use hash tables instead of very large 
value_strings? In diameter i think
3gpp AVP codes are in a value_string > 2k
Anders
> ___________________________________________________________________________
> Sent via:    Wireshark-dev mailing list<[email protected]>
> Archives:    http://www.wireshark.org/lists/wireshark-dev
> Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
>               mailto:[email protected]?subject=unsubscribe
>
>    

___________________________________________________________________________
Sent via:    Wireshark-dev mailing list <[email protected]>
Archives:    http://www.wireshark.org/lists/wireshark-dev
Unsubscribe: https://wireshark.org/mailman/options/wireshark-dev
             mailto:[email protected]?subject=unsubscribe

Reply via email to