patch with larger errors for "float"

[paul zimmermann]

       Dear Carlos,

thank you for your detailed review of my patch. Attached is a v2.
I removed test cases for atan2, cosh, erf and sin which did not
give larger errors, as noticed by Joseph (probably because I did
not use the testrun.sh wrapper, thus I used a previous glibc version).

> Just to confirm, does it pass make check without regression?

well, not really. I get *before* and after my patch:

Summary of test results:
    141 FAIL
   5025 PASS
     34 UNSUPPORTED
     17 XFAIL
      2 XPASS

Best regards,
Paul
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 0001-math-Add-inputs-that-yield-larger-errors-for-float-t.patch
Type: application/octet-stream
Size: 29680 bytes
Desc: not available
URL: <http://sourceware.org/pipermail/libc-alpha/attachments/20200121/711a0636/attachment.obj>