Actually, C on VMS/Alpha knows very well what a long long is, and
    knows how to make use of it.  So let's stop pretending the Alpha
    doesn't know long long...

And who said we were pretending? SIXTY_FOUR_BIT defines already BN_ULONG as long long...


  --- openssl/makevms.com       11 Jul 2004 20:30:33 -0000      1.42
  +++ openssl/makevms.com       6 May 2005 13:33:16 -0000       1.43
  @@ -274,6 +274,7 @@
   $   WRITE H_FILE "#endif"
   $!
   $   WRITE H_FILE "#if defined(HEADER_BN_H)"
  +$   WRITE H_FILE "#define BN_LLONG"
   $   WRITE H_FILE "#undef SIXTY_FOUR_BIT_LONG"
   $   WRITE H_FILE "#undef SIXTY_FOUR_BIT"
   $   WRITE H_FILE "#define SIXTY_FOUR_BIT"

BN_LLONG may not be defined along with SIXTY_FOUR_BIT, as it normally has devastating effect. It works on VMS/Alpha only thanks to inline BN_UMULT_HIGH assembler and shall stop working the moment you disengage it with no-asm [or whatever is equivalent on VMS]. If this is somehow is spinning around pqueue problem, then I find it more appropriate to eliminate dependancy from BN_ macros, rather than trying to squeeze it to where it doesn't belong in. A.
______________________________________________________________________
OpenSSL Project http://www.openssl.org
Development Mailing List [email protected]
Automated List Manager [EMAIL PROTECTED]

Reply via email to