Re: SV: Potential bug in win32 binary version?

[Simon Josefsson]

"Daniel Armyr" <address@hidden> writes:

>>Fwiw, libgcrypt is packaged for Windows via GnuTLS4Win see
> <http://josefsson.org/gnutls4win/>.
> Does this mean that if I simply install libcrypt, things should start
> working?

No, unfortunately.  GNU SASL detects whether libgcrypt is installed
during build-time, but the Vortex build wasn't built against libgcrypt.

> Presently I have implemented a simplified version of CRAM-MD5 by myself
> using a pesudo-random number challenge and it seems to work. Unfortunately,
> the legal considderations are slightly complex and the implementation
> probably suffers unnecessary weakness. I would therefore like to go back to
> gsasl authentication as soon as possible.

Right.

Building for Windows is easy if you have MinGW installed, maybe you can
try it.  Otherwise you'll have to wait for Francis to update the Vortex
build, or me to produce a GSASL4Win package.  I'll see if a simple hack
for gnutls4win leads to a usable gsasl build as well..

/Simon