Some issues with OS X 10.4.8 are being reported, and I'm wondering if these affect FileMaker?

1. 4D has found a serious calculation bug:

http://fmforums.com/forum/showtopic.php?fid/144/tid/181094

2. Folks are reporting other calculation issues. Since 10.4.8 supposedly "Improves the accuracy of Rosetta numerics and addresses Altivec translation issues on Intel-based Macs." it's plausible that something has indeed changed. http://docs.info.apple.com/article.html?artnum=304200

Is FileMaker affected? Anyone out there have a self-auditing solution that would catch any math changes?

AFAIK, FM 6 solutions are affected, but not FM 8.5 for which an Universal Binary is available.

Do you have a link to any info about this?

If non-UB versions are affected on intel (e.g. Rosetta), then this could presumably affect everyone running server 8v3 on intel, as well as FM client 8v3 or 8v4 on intel...

FileMaker Server 8 is not supported on Rosetta. Also tehre are reports of FMP 8.5 being affected by this issue.

Steven

You are correct of course that FM Server 8v3 won't run under Rosetta.

However, FM Client 8.5, being a universal binary, should not be running under Rosetta either.

So I'm a bit confused -- Apple's release notes seem to suggest that the changes in math calcs. were for Rosetta apps, in which case one would expect that both Server 8v4 and Client 8.5v1 would be immune?

Anyone have any actual info they can share?

Can't comment on Apple's Release Notes. Howver the various FMP lists are reporting instances of issues with FMP 8.0 and 8.5 with this problem.

Steven

FYI, 4D now has a patch that lets you use the 10.4.7 math libraries under 10.4.8 http://www.4d.com/1048fix.html

Does anyone know if 10.4.9 will fix the Intel Mac 10.4.8 math bug? Or Leopard, for that matter? This is a huge problem in FMP6. We would like to upgrade to FMP 8.5, but it will take time. We have a bound solution in FM6 and this affects the serialization of the program and some plugins.