[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: More about ACL performance



I note that your analysis appears to be against OPENLDAP_REL_ENG_2
as the function in question has been significant changed in HEAD.
Please forward port enhancement suggestions whenever possible.

At 07:13 AM 2002-01-07, Stephan Siano wrote:
>during my tests with ACL caching, I found out, that the ad_inlist function is 
>much slower than it could be. Each attrs value is first converted into a 
>AttributeDescription, than the AttributeType is compared and finally the 
>AttributeDescription is thrown away.

No, the AttributeDescription is compared.

>The following patch just receices the AttributeType and compares it.

That would break access controls specific to an Attribute
Description (an attribute type with zero or more options).

Kurt