Re: NA and NaN representation on mips64el

Hi,

On 2025-05-13 09:24, Rafael Laboissi=C3=A8re wrote:
> * YunQiang Su <[email protected]> [2025-05-13 08:49]:
>=20
> > Rafael Laboissi=C3=A8re <[email protected]> =E4=BA=8E2025=E5=B9=B45=E6=
=9C=8813=E6=97=A5=E5=91=A8=E4=BA=8C 00:30=E5=86=99=E9=81=93=EF=BC=9A
> > >=20
> > > [=E2=80=A6]
> > >=20
> > > It seems to be a problem related to how floating point NA and NaN
> > > are represented.
> > >=20
> > > Strangely, I cannot reproduce this problem. I built the package
> > > manually on a mips64el porterbox using schroot and everything worked
> > > fine.
> >=20
> > I see. The problem is that the machine (mipsel-osuosl-03.debian.org) is
> > Loongson 3A4000. The problem is that MIPS swap the encodings of sNaN and
> > qNaN since MIPSr5.
> >=20
> > I guess that octave uses sNaN as NA? We have no good solution for it. We
> > have meet some other packages with this case. Let's just ask the help
> > from DSA to pin this package on non-3A4000 machines.
>=20
> Could you please tell me how to make do such a request?

You should contact the wanna-build team [1], but the possibility to=20
block a package on a specific machine is being phased-out with the=20
switch to pybuildd.

Ideally you should disabled the test on nan 2008 machine, for instance by=
=20
looking at "nan_2008" in /proc/cpuinfo (as opposed to "nan_legacy").

Regards
Aurelien

[1] [email protected]

--=20
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
[email protected]                     http://aurel32.net