pandas -- I think we should drop BE platforms

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

pandas -- I think we should drop BE platforms

Yaroslav Halchenko
Hi The Team,

"Maintaining" pandas builds on big endians (and some others, but let's
concentrate on BE for now) was always problematic.  Upstream does not
support them and I am somewhat tired of pestering them with failures on
them. Only once or twice test failures on those platforms pointed
to genuine bugs (iirc in numpy or scipy) which just were not revealed on
x86.

Selectively skipping the tests on those platforms is just hiding the
bugs away and pretending that it is all working correctly AFAIK.  The
worst thing which I think we could do to the users on those platforms
(if there are any) is to make their results knowingly incorrect,
while no alarm is raised.

As to me, the most logical step would be to stop pretending to support
BE builds for pandas and drop support of all big endian builds
altogether.  

1. remove all patches which disable tests on big endians (I would have
even patched to remove upstream skips on those archs, so we do not fall
into the trap of missing smth)

2. Please correct me if I am wrong, but I do not think that there
is some flag to say smth like

    Architecture: any [!big-endian]

so we would need to unlist them explicitly

    Architecture: any [!s390x, !powerpc, !mips]

(or powerpc is not BE?)

Let me know what you think and  either I have missed anything.  If
someone wants to step up and jump into maintaining on BEs, please do so
now.  If not, I will proceed as planned above.

--
Yaroslav O. Halchenko
Center for Open Neuroscience     http://centerforopenneuroscience.org
Dartmouth College, 419 Moore Hall, Hinman Box 6207, Hanover, NH 03755
Phone: +1 (603) 646-9834                       Fax: +1 (603) 646-1419
WWW:   http://www.linkedin.com/in/yarik       

Reply | Threaded
Open this post in threaded view
|

Re: pandas -- I think we should drop BE platforms

Sébastien Villemot-2
On Sat, Feb 24, 2018 at 02:47:19PM -0500, Yaroslav Halchenko wrote:

> 1. remove all patches which disable tests on big endians (I would have
> even patched to remove upstream skips on those archs, so we do not fall
> into the trap of missing smth)

I agree with that.

> 2. Please correct me if I am wrong, but I do not think that there
> is some flag to say smth like
>
>     Architecture: any [!big-endian]
>
> so we would need to unlist them explicitly
>
>     Architecture: any [!s390x, !powerpc, !mips]
>
> (or powerpc is not BE?)
No, leave "Architecture: any". Then the builds will fail on those arches. This
is what porters generally prefer, because they have a clear information about
what is wrong with the package on their arch (and possibly they can act and
send patches).

--
⢀⣴⠾⠻⢶⣦⠀  Sébastien Villemot
⣾⠁⢠⠒⠀⣿⡁  Debian Developer
⢿⡄⠘⠷⠚⠋⠀  http://sebastien.villemot.name
⠈⠳⣄⠀⠀⠀⠀  http://www.debian.org

signature.asc (849 bytes) Download Attachment