It's also bundled in my Apple osX where my iPhone development environment 
lives, not to stray TOO far afield. *g*

Stev3
Lyrasis

-----Original Message-----
From: [email protected] 
[mailto:[email protected]] On Behalf Of Thomas 
Berezansky
Sent: Tuesday, December 22, 2009 9:25 AM
To: [email protected]
Subject: Re: [OPEN-ILS-DEV] ***SPAM*** Re: ***SPAM*** Re: SHA1 code in sha.c

Additional note:

Several things, such as postgresql (client and server) and apache with
SSL, already require that the openssl implementation be installed on
the system. The SSH implementation installed on the systems will
likely also require it. Adding another makes no sense, especially when
we already have openssl's sitting there from everything else. Given
that we can SSH across endianess and int sizes I assume that all of
that is covered properly in the openssl implementation.

Thomas Berezansky
Merrimack Valley Library Consortium


Quoting Jason Stephenson <[email protected]>:

> Quoting Scott McKellar <[email protected]>:
>
>>
>>
>> --- On Mon, 12/21/09, Jason Stephenson <[email protected]> wrote:
>>> Why reinvent the wheel? Why not just
>>> farm the work out to libcrypto? Last I checked OpenSSL
>>> typically has a SHA1 implementation that you don't have to
>>> fiddle with.
>>>
>>> Jason
>>
>> I have no intention of reinventing any wheels.  I did a quick Google and
>> saw the OpenSSL project among others.  What I don't know is how portable
>> they are.  The ones I looked at so far don't say very much on their
>> websites about architectures, int sizes, endianness, or other aspects of
>> portability.  I'd rather not have to decipher a dozen different
>> implementations of an algorithm I don't understand in order to figure
>> out which ones are better than what we've got.
>
> OpenSSL is about as portable as you can get, and is pretty much
> guaranteed to be installed on any machine that can host OpenILS.
>
>
>>
>> If anybody already knows of a suitable implementation, or knows of one
>> that we should avoid, you may be able to save me some time.
>>
>> Scott McKellar
>>
>>
>
>
>


Reply via email to