User`s guide

Microsoft Windows on PCs

2-15

Known Compiler Limitations. There are several known restrictions regarding the

use of supported compilers:

• Some compilers, e.g., Watcom, do not handle denormalized floating-point

values correctly. Denormalized floating-point numbers are numbers that are

greater than 0 and less than the value of

DBL_MIN in your compiler’s float.h

file.

• The MATLAB Compiler sometimes will generate

goto statements for

complicated

if conditions. The Borland C++ Compiler prohibits the goto

statement within a

try catch block. This error can occur if you use the

-A debugline:on option, because its implementation uses try catch. To

work around this limitation, simplify the

if conditions.

• There is a limitation with the Borland C++ Compiler. In your M-code, if you

use a constant number that includes a leading zero and contains the digit ‘8’

or ‘9’ before the decimal point, the Borland compiler will display the error

message

Error <file>.c <line>: Illegal octal digit in function

For example, the Borland compiler considers 009.0 an illegal octal integer as

opposed to a legal floating-point constant, which is how it is defined in the

ANSI C standard.

As an aside, if all the digits are in the legal range for octal numbers (0-7),

then the compiler will incorrectly treat the number as a floating-point value.

So, if you have code such as

x = [007 06 10];

and want to use the Borland compiler, you should edit the M-code to remove

the leading zeros and write it as

x = [7 6 10];