[Numpy-discussion] float128 / longdouble on PPC - is it broken?

Pauli Virtanen pav@iki...
Wed Oct 26 11:41:05 CDT 2011


26.10.2011 10:07, Nathaniel Smith kirjoitti:
> On Tue, Oct 25, 2011 at 4:49 PM, Matthew Brett <matthew.brett@gmail.com> wrote:
>> I guess from your answer that such a warning would be complicated to
>> implement, and if that's the case, I can imagine it would be low
>> priority.
> 
> I assume the problem is more that it would be a weirdo check that
> becomes a maintenance burden ("what is this doing here? Do we still
> need it? who knows?") than that it would be hard to do.

This check should be done compile-time, and set a flag like
BROKEN_LONG_DOUBLE_FORMATTING for compilation. IIRC, we already have
some related workarounds in the code (for Windows).

The main argument is, (i) it's not our bug, (ii) adding workarounds for
broken platforms should be weighed having the commonness of the issue in
mind.

But I'll admit that someone could have got into work and implemented the
workaround in the same time it has taken to write mails to this thread :)

	Pauli



More information about the NumPy-Discussion mailing list