[PATCH 1 of 2] sha1sum.py: sha1sum for platforms which do not provide it

[Lee Cantey]

On Mon, Oct 31, 2011 at 8:11 AM, Christian Ebert <blacktrash at gmx.net> wrote:
> * Greg Ward on Monday, October 31, 2011 at 09:02:36 -0400
>> On Sun, Oct 30, 2011 at 3:08 PM, Christian Ebert <blacktrash at gmx.net> wrote:
>>> # HG changeset patch
>>> # User Christian Ebert <blacktrash at gmx.net>
>>> # Date 1320001536 -3600
>>> # Branch stable
>>> # Node ID 96211cffd4d151f9c1fb34d38bb67eba5b92b944
>>> # Parent  87248de091352f746e8cd1cee01af29138c8e443
>>> sha1sum.py: sha1sum for platforms which do not provide it
>>
>> I give a reluctant +1 on this.
>>
>> Reluctant because I think we should refactor md5sum.py and sha1sum.py
>> and have a single script for hashing in the test suite.
>
> I completely agree.
>
>> +1 because now is not the time for refactoring. Now is the time for fixing.

Agreed.  I thought that was the decision last week.

> Well, it hasn't gone in yet. I also was insecure how much we can
> change because of the copyright.
>
> I'm completely uneducated wrt the differences between md5 and sha
> - could we just use the existing md5sum.py for test-lfconvert.t?
> Or is md5 not secure enough in this case?

The use of SHA1 is to match the extension which uses SHA1 hashes to
test identity.

Regards,
Lee