>
>I did find that (just after posting here, natch) but although it explains the issue, I couldn't find any pointers to solutions or workarounds.
>
>For instance, I thought, OK, if I test CPCURRENT and find it doesn't match CPDBF, then why not simply issue a "CODEPAGE=CPDBF()" and run the query as usual. Tried it, got a marginal improvement, (knocked the 17 down to 10 seconds) but still not a patch on the original vfp8 speed.
>
>Presumably the answer to the absence of the problem with ref to date fields is because both vfp8 and 9 deal with dates the same way regardless of codepages.

Only character/memo fields are affected by codepage translation.

>There also seems to be some dispute as to whether it is a bug or the result of a bug fix! One way or another, we obviously need a fix for it or a lot of our clients and users are going to be very peed off!

The confusion comes from the fact that this change wasn't documented in VFP9 help.
You have to change table codepage to the current codepage if you want Rushmore to work. If you want to convert data to the current codepage, use COPY TO command. If you want to change codepage only, use cpzero.prg from VFP Tools\Cpzero subdirectory.