KEIICHI NAKASATO le escribe a STEPHEN LINDHOLM:

KN> I know that Borland's DOS version was about double in speed
KN> compared  to QC even before a good pre-compiled header were
KN> added with 3.1.  I  doubt QC can compile in near 1 second

Alas, Borland is too different to move my code. In any event, I expect the
big B. to die soon, or at least be reborn under a very different policy.
Furthermore, my 286 is incapable of running Borland, and I refuse on the
basis of sanity to use two different compilers on two different machines
and expect to swap code.

KN> range if header files were  100k or larger. 

I'll have to try this. You DO use a ramdisk, don't you?

KN> IMO, I think all compiler should have options to disable
KN> optimization.   

I think that they should make a separate version that compiles quickly
without a care for size nor execution speed, or at least a switch for this.

KN> BTW, I am curious if your compiler support "rt" for fopen().

  Syntax:    FILE *fopen( char *filename, char *mode );
             mode: r, w, a, r+, w+, a+
                   (t or b appended to  indicates type)

Why?

loSmaH cha'

--- timEd-B9