[PATCH] largefiles: Use sha1sum.py since OS X has generic shasum utility

[Na'Tosha Bard]

2011/10/26 Na'Tosha Bard <natosha at unity3d.com>

> 2011/10/25 Matt Mackall <mpm at selenic.com>
>
>> On Mon, 2011-10-24 at 20:35 -0400, Greg Ward wrote:
>> > On Mon, Oct 24, 2011 at 2:15 PM, Lee Cantey <lcantey at gmail.com> wrote:
>> > > # HG changeset patch
>> > > # User Lee Cantey <lcantey at gmail.com>
>> > > # Date 1319479837 25200
>> > > # Branch stable
>> > > # Node ID 17323a5ee76870b1b48e7308511c34c98825f35a
>> > > # Parent  8b8dd13295dbd733cc03ebb3c2af2f33d24d5428
>> > > largefiles: Use sha1sum.py since OS X has generic shasum utility.
>> > >
>> > > diff -r 8b8dd13295db -r 17323a5ee768 tests/sha1sum.py
>> > > --- /dev/null   Thu Jan 01 00:00:00 1970 +0000
>> > > +++ b/tests/sha1sum.py  Mon Oct 24 11:10:37 2011 -0700
>> > > @@ -0,0 +1,42 @@
>> > > +#!/usr/bin/env python
>> > > +#
>> > > +# Based on python's Tools/scripts/md5sum.py
>> > > +#
>> > > +# This software may be used and distributed according to the terms
>> > > +# of the PYTHON SOFTWARE FOUNDATION LICENSE VERSION 2, which is
>> > > +# GPL-compatible.
>> > > +
>> > > +import sys, os
>> > > +
>> > > +try:
>> > > +    from hashlib import sha1
>> > > +except ImportError:
>> > > +    from sha import sha as sha1
>> >
>> > It looks like you just photocopied tests/md5sum.py and called it
>> > tests/sha1sum.py. It would be ever so much nicer to refactor, e.g.
>> >
>> >   * rename md5sum.py to hash.py
>> >   * make it take two args, hash algorithm and filename
>> >   * update existing callers in tests/
>> >   * make test-largefiles.t use it
>>
>> As far as I'm aware, we only need one hash function for tests and it
>> needn't be cryptographically strong. I'll change the largefiles test to
>> use the existing md5sum.py.
>>
>
> Is there really a good reason to do this?  This means (as far as I can
> tell) that suddenly kbfiles users are now going to have to convert their
> entire repositories to be able to use largefiles.
>

Oops, nevermind -- apparently I'm not awake enough yet.  I thought you were
going to change largefiles to fundamentally use md5sums instead of SHA1
hashes for the standins.

I still think it's proper to use SHA1 in the tests described here as well,
but it's less serious.

Na'Tosha



-- 
*Na'Tosha Bard*
Build & Infrastructure Developer | Unity Technologies

*E-Mail:* natosha at unity3d.com
*Skype:* natosha.bard
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://selenic.com/pipermail/mercurial-devel/attachments/20111026/ffed59c8/attachment.html>