[comp.os.os2.misc] bug in MSC 6.00A

oleg@ibm.com (11/28/90)

I just upgraded to 6.00A (maintenance release). Still have the same problem
as with 6.00. I get compiler internal error:


  cl /DKERNEL /DOS2 /c /Gs /Os /Asnw /Zpe /Fc /G2 uipc_so1.c
uipc_so1.c
uipc_so1.c(79) : warning C4017: cast of int expression to far pointer
uipc_so1.c(142) : warning C4033: function must return a value
uipc_so1.c(517) : warning C4017: cast of int expression to far pointer
uipc_so1.c(847) : fatal error C1001: Internal Compiler Error
                (compiler file '@(#)regMD.c:1.110', line 4688)
                Contact Microsoft Product Support Services

As far as I know this bug is not on the known bugs for 6.00A

Oleg Vishnepolsky
IBM T. J. Watson Research Center

oleg@ibm.com (11/28/90)

I just upgraded to 6.00A (maintenance release). Still have the same problem
as with 6.00. I get compiler internal error:


  cl /DKERNEL /DOS2 /c /Gs /Os /Asnw /Zpe /Fc /G2 uipc_so1.c
uipc_so1.c
uipc_so1.c(79) : warning C4017: cast of int expression to far pointer
uipc_so1.c(142) : warning C4033: function must return a value
uipc_so1.c(517) : warning C4017: cast of int expression to far pointer
uipc_so1.c(847) : fatal error C1001: Internal Compiler Error
                (compiler file '@(#)regMD.c:1.110', line 4688)
                Contact Microsoft Product Support Services

As far as I know this bug is not on the list of known bugs for 6.00A

Oleg Vishnepolsky
IBM T. J. Watson Research Center

jack@cscdec.cs.com (Jack Hudler) (11/29/90)

In article <1990Nov27.214743.22138@arnor.uucp> oleg@ibm.com writes:
>I just upgraded to 6.00A (maintenance release). Still have the same problem
>as with 6.00. I get compiler internal error:
>
I just updated also, I guess those of us that have monochrome screens
are left in the dark. Setup doesn't show the highlighted selections!

-- 
Jack           Computer Support Corporation             Dallas,Texas
Hudler         Internet: jack@cscdec.cs.com

ddl@husc6.harvard.edu (Dan Lanciani) (11/30/90)

In article <1990Nov27.214743.22138@arnor.uucp>, oleg@ibm.com writes:
| I just upgraded to 6.00A (maintenance release). Still have the same problem
| as with 6.00. I get compiler internal error:
| 
| 
|   cl /DKERNEL /DOS2 /c /Gs /Os /Asnw /Zpe /Fc /G2 uipc_so1.c
| uipc_so1.c
| uipc_so1.c(79) : warning C4017: cast of int expression to far pointer
| uipc_so1.c(142) : warning C4033: function must return a value
| uipc_so1.c(517) : warning C4017: cast of int expression to far pointer
| uipc_so1.c(847) : fatal error C1001: Internal Compiler Error
|                 (compiler file '@(#)regMD.c:1.110', line 4688)
|                 Contact Microsoft Product Support Services

	Yes, while compiling code also derived from the Berkeley
networking distribution I found the following:

The hundreds (i.e., every mtod construct) of internal compiler errors
from 6.00 were reduced to dozens in 6.00A.

All but one (!) of the remaining internal errors went away with
default optimization (== no -O switch at all).

The remaining error could be eliminated with -Od.

However, the resulting driver crashed immediately suggesting that
there might be some code generation errors as well...  Perhaps
6.1 will improve the situation.

				Dan Lanciani
				ddl@harvard.*

billf@progress.COM (Bill Ferro) (11/30/90)

oleg@ibm.com writes:

>I just upgraded to 6.00A (maintenance release). Still have the same problem
>as with 6.00. I get compiler internal error:


>  cl /DKERNEL /DOS2 /c /Gs /Os /Asnw /Zpe /Fc /G2 uipc_so1.c
>uipc_so1.c
>uipc_so1.c(79) : warning C4017: cast of int expression to far pointer
>uipc_so1.c(142) : warning C4033: function must return a value
>uipc_so1.c(517) : warning C4017: cast of int expression to far pointer
>uipc_so1.c(847) : fatal error C1001: Internal Compiler Error
>                (compiler file '@(#)regMD.c:1.110', line 4688)
>                Contact Microsoft Product Support Services

>As far as I know this bug is not on the known bugs for 6.00A

Try adding "/B1 C1L.EXE" to your parameter list.  This invokes the
large model of the compiler.  I saw similar fatal errors that went
away when switched....


-bf