On Sep 18, 2007, at 4:35 PM, Howard Chu wrote:

donn@u.washington.edu wrote:

...

On Aug 23, 2007, at 2:56 PM, Howard Chu wrote: [ ... re attribute value matching ... ]

...

The only way to do this right, without dying on an O(n^2) algorithm, is to sort both arrays first and then walk through them in order as the original code did. If you want to submit such a patch, please do. I don't have a lot of time to focus on this at the moment.

I did look at this, and the attached patch is what I came up with, relative to HEAD 1.356. I have tested it, but unfortunately the

Thanks but I already went ahead with a different fix for this since I didn't hear back from you.

Yes, I know, I understand that to be the O(n^2) option you mention above. It's certainly the simplest way.

Donn Cave, donn@u.washington.edu