Keith Scroggins wrote:
First, GMP has lots of issues being built for 68000. Building for the 68020-60 target is good, passes ALL tests.
Excellent for 68020-60 !But I don't understand why there are so much issues with 68000, it should be simpler.
Error in mpfr_sprintf (s, "%'Rf", x); expected: "100000000000000000.000000" got: "100,000,000,000,000,000.000000" FAIL: tsprintf
The "'" character asks to separate groups of thousands with the separator specified in the locale. So if the current locale is set to US, that looks good.
The following message seems to indicate this test is wrong: http://comments.gmane.org/gmane.comp.lib.mpfr.general/359 So I believe the MPFR test results are 100% correct :-)
Will also see if there is any hope with PPL/CLooG as well (since I don't have time to actively code, but I can just let my Falcon compile).
As I told you, PPL/CLooG are enabled in my cross binaries, I tried it once and the result was worse. While no one can't prove that stuff is useful for m68k, we should not waste time compiling that (since it is optional).
Very good work, Keith ! -- Vincent Rivière