[Namazu-users-en] Namazu encoding Unix vs win32

[starr at houston.rr.com]

 I'm trying to generate the Namazu index on Unix (FreeBSD) and search it
on win32. 
 When the win32 search runs, a considerable amount of garbage is
produced, along
 with some echoing of templates.  Unix search works fine.
 Looks to me like it may be a coding problem.
 mknmz scripts are identical on the two ports.

I'm scanning English, normal HTML.

Any Idea where the difference is ?