After a bit of digging I found that my header files define it as a
struct. I think the best option of all would be to change the declared
type from fpos_t to int since that is the type of variable that the
bochs code is expecting.
I did disable the sb16 to get it to compile actually, and I was pleased
that it ran faster than plex86 (a whopping 1.32 MIPS!) It didn't seem to
complain when a used the BIOS files from plex86 so I gues that doing
that is ok?
Oh, and what would happen if did not configure to build the FPU
emulator?

--John

"Parker, Ron" wrote:
> 
> The fpos_t type should be defined in stdio.h.   It is permissable for fpos_t
> to be defined as any type except an array.  Most compilers define it as some
> form of integer and math operations can be performed upon it.  However, this
> is not part of the standard.
> 
> The C standard only mentions using fpos_t as a parameter to the fgetpos and
> fsetpos functions.  It also permits using a struct for this type.  This
> would mean that no mathmatical operations would work.  I believe that this
> may be the case that is causing your problem.
> 
> As I see it you have two solutions.  The first is to disable sb16 support
> when you configure bochs.  The second is to look at you compiler's
> definition of fpos_t and replace the error-causing lines with code
> appropriate to your machine.
> 
> -----Original Message-----
> From: Johnathan Hicks [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, March 22, 2001 12:24 PM
> To: plex86
> Subject: [plex86] problem building bochs
> 
> I'm guessing that this is the place to take bochs issues as well as
> plex86? Presuming that it is, I seem to be haveing problems building
> bochs as this shows:
> 
> sb16.cc: In function `static void bx_sb16_c::finishmidifile()':
> sb16.cc:2788: no match for `fpos_t & -= int'
> sb16.cc:2794: no match for `fpos_t & << int'
> sb16.cc:2794: no match for `fpos_t & >> int'
> sb16.cc:2795: no match for `fpos_t & & int'
> sb16.cc:2796: no match for `fpos_t & & int'
> 
> I hope this can be fixed.
> Thanks.
> 
> --John


Reply via email to