1.1.42 / Win32 MinGW : random number generation is very slow

[Ludovic LANGE]

Hello Low Halo,

>>(More precisely, I'm trying to add support for an encrypted
>>protocol in GAIM, an Instant Messenging application)
> 
> 
>     I started a project that does just this back in November 2002.
> Gaim + libgcrypt = Ultramagnetic (http://ultramagnetic.sourceforge.net/)
> 
>     Please have a look at what I've created already.  Its currently
> in beta, but is rapidly maturing.  Essentially, the only missing
> features that prevent this project from declaring a stable release
> is more testing and an ECB mode-to-CTR/CBC mode patch.
> 
>     Win32 support works, kind of.  Ultramagnetic currently uses
> libgcrypt v1.1.12 because v1.1.42 deadlocks due to some kind of
> thread library conflict with GTK/GDK.  And from experience, Win32
> doesn't like v1.1.12 too much so Windows support is a little
> shaky at the moment.  I hope to fix this soon.
> 
>     I've been working on this project entirely by myself for the
> last ten months.  I'd be more than happy to welcome another
> developer aboard!


Thanks for your answer. I discovered your project a few days ago, and 
that confirmed me into chosing libgcrypt as a solution for my needs. 
(thanks by the way)

Our goals seem similar (implementing an encrypted protocol in Gaim) 
indeed. What I'm trying to do, is to add support, in Gaim, for the 
Trillian 'SecureIM' encrypted protocol. I  unfortunately don't have to 
time to work full time on such a project, and my goal was to publish, in 
the form of a patch, the first bits allowing to receive such encrypted 
messages.

I did succeed, and posted a first version ( 
http://sourceforge.net/tracker/index.php?func=detail&aid=777300&group_id=235&atid=300235 
) of the patch to Gaim. However this was refused, because I used OpenSSL 
as a library, and there are license incompatibilites between OpenSSL and 
Gaim.

So I rewrote my patch to use libgcrypt library.

Now my work is almost done, my Gaim version is working (win32 + 
libgcrypt 1.1.42) enough for me to update my patch.

As I said, I don't have nor time to work (full-time or part-time) on 
such a big project for the time being, nor the competence (in encryption 
field). However, I'll have a closer look on Ultramagnetic.

By the way, you are of course free to use my 'Trillian' patch in your 
work if you want to support this protocol also (which should be regarded 
as a compatibility protocol, because I read somewhere that the 
Diffie-Hellman key exchange they use could be subject to a 
man-in-the-middle attack, which may be a problem.)

Best regards,

Ludovic LANGE