fatal error 26 (Xenix 386)

[Jonathan Bayer]

In article <10 at wells.UUCP> edw at wells.UUCP (Ed Wells) writes:
>
>  Several times while compiling large 'C' programs I ran into
>'fatal error 26' errors.  This is from a 'parser stack overflow'.
>Does anyone know how to get around this one.  I would think that
>if the 386 if virtual, that the stack should never overflow unless
>you run out of real memory.
>
>  What's strange about this is that even though it crashes compiling
>in 386 mode, it works like in 286 large mode (-M2l flag).


I assume you are running SCO Xenix with the 2.2 development system. 
First thing is to know that the 86 and 286 mode use a different compiler
than the 386 mode.  That will explain why  it works in 286 large mode,
and not in 386 mode.

The stack overflow is harder.  I would assume that there are some
internal constants in the compiler which are being exceeded.  The parser
probably has it's own internal stack stored somewhere in an array.  If
the array is exceeded it would generate this error.  As to why the 386
compiler would have smaller constants than the 286, your guess is as
good as mine.


JB
-- 
Jonathan Bayer			      Beware: The light at the end of the
Intelligent Software Products, Inc.	      tunnel may be an oncoming dragon
19 Virginia Ave.				...uunet!ispi!jbayer
Rockville Centre, NY 11570  (516) 766-2867    jbayer at ispi.UUCP