[Biopython-dev] Biopython 1.66 release?

[Peter Cock]

On Wed, Oct 21, 2015 at 6:14 PM, Peter Cock <p.j.a.cock at googlemail.com> wrote:
> Hi all,
>
> Sanity testing of these provisional source bundles and binaries already on
> http://biopython.org/DIST/ would be appreciated, especially during the next
> few hours (during which time I will draft the release announcement and
> have dinner), and prior to the push to PyPI:
>
> 2f09afb0a7740a70f64b6d10dfc2582b  biopython-1.66.tar.gz
> 4344c2e7b91deb47f9ae2ebf7ad1e375  biopython-1.66.zip
> 192bc8a70fe098e5eae71fc3ad38e3fd  biopython-1.66.win32-py2.6.exe
> 6e4bf2b9eb9533ebd6930df6b506db51  biopython-1.66.win32-py2.7.exe
> fa7a6010b7ecce7fdd8fa614a3f5845e  biopython-1.66.win32-py3.3.exe
> c985a3477032314c1ec78e8ada916371  biopython-1.66.win32-py3.4.exe
>
> These should match this commit (not tagged yet):
>
> https://github.com/biopython/biopython/commit/c06a3066d858c4818cc6b4bf7e8d776bb9473a63

Update to PyPI done:

https://pypi.python.org/pypi/biopython/1.66

Interestingly this seem to insist on rebuilding the tar-ball, which
means the tar ball now up on PyPI has a different md5sum:

d3f4c14e5cba3dfd354d0e0013bc350f  dist/biopython-1.66.tar.gz

I presume this is down to containing different date stamp(s).

If we start to advertise the expected md5 checksums in future, this may
mean revising our release process slightly... not that I have followed
http://biopython.org/wiki/Building_a_release exactly by delaying the
PyPI release until I'd had a third party confirm the downloads looked
fine.

Peter