Re: glibc 2.2: math test failures

[Michael Deutschmann]

I have a new followup.

That I was (apparently) first to notice the math problems on the 486DX4,
indicated to me that glibc was not being tested much on lower end
machines.  So I decided to try running make check with the oldest and
slowest computer I own -- a 16MHz 80386 with a Weitek coprocessor.
Needless to say, this took awhile...

This also had lots more math errors, and not the same errors as on the
486.  Also, this time more errors occur in inline mode.  Quick summary:

basic-test:             2 errors
test-float:             1 error
test-ifloat:            5 errors
test-doubl:             3 errors
test-idoubl:            7 errors
test-ldoubl:            82 errors
test-ildoubl:           86 errors

The complete .out files are available in:
ftp://ftp.ocis.net/pub/ldeutsch/transient/math-errors.tar.gz

There were also some other new test failures outside of math, but I'm
not quite sure they are real problems -- I'll raise them once I've
confirmed them.

---- Michael Deutschmann <address@hidden>