Re: [platform-testers] Release candidate mpc-1.3.0rc1

[Paul Zimmermann]

thank you all for your reports.

We are able to reproduce the tagm issue (even on 64-bit machines with larger
precision). This issue is under investigation.

For the teta issue, we are unable to reproduce it under 32-bit Linux.

Please can you investigate at which line of teta.c the variable eta
becomes 0?

   ok = mpcb_can_round (j, 291, 234, MPC_RNDNN);
   ok &= (re == -262537412640768000L && im == 1);
   ok &= (re == 1728 && im == 1);

I suspect the 2nd line. Maybe we should write -262537412640768000LL?

Best regards,
Paul

> From: sisyphus <sisyphus359@gmail.com>
> Date: Wed, 30 Nov 2022 23:18:10 +1100
> 
> 
> [1:multipart/alternative Hide]
> 
> 
> [1/1:text/plain Show]
> 
> 
> [1/2:text/html Hide Save:noname (1kB)]
> 
> Hi Andreas,
> On Windows 11, using 32-bit versions of gcc-11.3.0, gcc-12.1.0 and
> gcc-12.2.0, I'm seeing the same 2 failing test programs as already reported
> by Trizen.
> The test suite log is attached.
> 
> With 64-bit versions of the same compilers on the same system, it's only the
> teta.exe program that fails.
> For those 64-bit failures of teta.exe, the test-suite log provides the same
> (minimal) information as provided for the 32-bit builds.
> 
> Cheers,
> Rob
> 
> [2:application/octet-stream Show Save:test-suite-1130-stat32.log (5kB)]
>