This is the mail archive of the
mailing list for the glibc project.
Re: Ping^2 Re: Use gen-libm-test.py to generate ulps table for manual
- From: Zack Weinberg <zackw at panix dot com>
- To: "Carlos O'Donell" <carlos at redhat dot com>
- Cc: Joseph Myers <joseph at codesourcery dot com>, "Dmitry V. Levin" <ldv at altlinux dot org>, Andreas Schwab <schwab at suse dot de>, GNU C Library <libc-alpha at sourceware dot org>, Paul Eggert <eggert at cs dot ucla dot edu>
- Date: Mon, 1 Oct 2018 11:34:26 -0400
- Subject: Re: Ping^2 Re: Use gen-libm-test.py to generate ulps table for manual
- References: <alpine.DEB.firstname.lastname@example.org> <alpine.DEB.email@example.com> <firstname.lastname@example.org> <CAKCAbMgo=+vv_cX5edqEHAf-Uo41zzaJZJpPZFwb9MuGcNbZrw@mail.gmail.com> <email@example.com> <alpine.DEB.firstname.lastname@example.org> <alpine.DEB.email@example.com> <20180910181224.GC8807@altlinux.org> <alpine.DEB.firstname.lastname@example.org> <alpine.DEB.email@example.com> <alpine.DEB.firstname.lastname@example.org> <email@example.com>
On Mon, Oct 1, 2018 at 11:21 AM Carlos O'Donell <firstname.lastname@example.org> wrote:
> On 10/1/18 11:06 AM, Joseph Myers wrote:
> > On Tue, 25 Sep 2018, Joseph Myers wrote:
> >> Ping. Any conclusion on this question about indirect dependencies
> >> introduced by use of Python?
> > Another ping on this question....
> Could you help me to gather consensus?
> If you could write up the following:
> * Exact version of python to use.
> * Exact set of imports we need.
> I'll take that and try to get an acknowledgment from each of the
> distributions. I'll email all the maintainers and follow up with
> them (all on libc-alpha) in a new thread starting with the information
> you will put together for me.
> I'll also include some of the resolution about the concerns regarding
> building python on an old distribution.
For the record, although I offered some details earlier about how
Debian packages Python, I have no official relationship to Debian and
cannot speak for any of their teams that might care about this
As a glibc developer who might well take on some of the work of
rewriting existing build logic in Python, I would like to put in a
word for requiring Python 3.(>=4) rather than allowing 2.7 as well.
The 2.x series is close to end-of-life, writing scripts that work with
both major versions of the language is significantly harder than
writing 3.x-only code, and I fear 2.x build compatibility will not be
adequately tested. 3.4 is the oldest version of the 3.x series still
in wide use AFAIK.