On Thu, 20 Feb 2003 14:29:34 -0800 (Pacific Standard Time) Mark Crispin <[EMAIL PROTECTED]> wrote:
MC> > Is there any reason to not do the same thing you did in SSL code for MC> > Kerberos code as well? My program doesn't currently support it anyhow but MC> > I'd like to add support in the future if possible -- but without making it MC> > impossible to run it under the Kerberos-less versions of Windows. MC> MC> Considering how many variants of Windows there are, and the resulting DLL MC> hell, I don't want to even pretend. Either it's the old way or the new MC> way. Again, I perfectly understand that you don't want to support Kerberos on old systems, but all I want is to allow the program to run there _without_ Kerberos support -- but still have it under W2K. Would you know how Pine deals with this problem BTW? MC> > I know that it's currently impossible, but it doesn't seem to be that hard MC> > to make it possible: simply rename all driver functions ssl_xxx() to MC> > ssl_nt_xxx() and ssl_w2k_xxx() and provide ssl_xxx() wrappers which would MC> > route to either NT or W2K version depending on which one is available. Is MC> > there any obvious reason for which this couldn't work? MC> MC> It won't start on the old systems if the W2K version is in the binary. Well, if the code was written in the same way as ssl_nt.c, i.e. loaded the libraries dynamically, it would start -- or is there something I'm missing again? Thank you for your explanations, VZ -- ------------------------------------------------------------------ For information about this mailing list, and its archives, see: http://www.washington.edu/imap/c-client-list.html ------------------------------------------------------------------
