keyring extension on Windows

[Steve Borho]

On Mon, Jan 11, 2010 at 8:27 AM, Marcin Kasperski
<Marcin.Kasperski at mekk.waw.pl> wrote:
>> So it seems my only options are to hack mercurial_keyring to always
>> use keys that are ConfigParser safe (seems not very robust) or
>> implement my own keyring backend in TortoiseHg, perhaps storing
>> encrypted keys in the registry instead of a text file as the
>> Win32CryptoKeyring class does.
>
> Steve, I don't have any Windows machine handy, but as far as
> mercurial_keyring is considered, I am open to patches (and ready to test
> whether they spoil something on Linux, or not). Preferably just
> fork http://bitbucket.org/Mekk/mercurial_keyring/ and let me know
> that there is a clone worth pulling.
>
> In particular, the key format was picked completely arbitrarily and I
> have no problems with changing it to something more suitable { of course
> I could change it myself, but it would be nice to ensure that the
> updated version really works well and resolves the problem }

I think it's more robust to use the registry rather than ConfigParser,
so I implemented a new backend for keyring itself.

http://bitbucket.org/sborho/python-keyring-lib/

> PS there is new keyring backend for Win7 available using win32 crypto
> api available at
> http://bitbucket.org/michaelgruenewald/python-keyring-lib/
> (either it was already pulled to main keyiring repo, or will be soon)

I'll take a look.  Thanks.

--
Steve Borho