Re: ezmlm-idx and charsets

On 27-09-2008 at 4:36, Bruce Guenter wrote:

 >The problem is that the author and thread hashes are based on the
 >decoded headers.  Any non-US-ASCII index created after the change will
 >be different than the same one created before.

I was missing that point.  Still, I wouldn't require users to reindex
their archives unless they want to.  I can imagine situations where
having a slightly inconsistent index with duplicate authors and threads
across the switch would be better than having to reindex everything just
to upgrade their version of ezmlm-idx.

--=20
alessandro ranellucci.
http://ezmlm-www.sourceforge.net/