PRB: C2118: Negative Subscript May Be Caused by int Expression (68475)

The information in this article applies to:

Microsoft C for MS-DOS
Microsoft C for OS/2
Microsoft Visual C++ for Windows, 16-bit edition 1.0
Microsoft Visual C++ for Windows, 16-bit edition 1.5
Microsoft Visual C++ for Windows, 16-bit edition 1.51
Microsoft Visual C++, 32-bit Editions 1.0
Microsoft Visual C++, 32-bit Editions 2.0
Microsoft Visual C++, 32-bit Editions 2.1
Microsoft Visual C++, 32-bit Editions 4.0
Microsoft Visual C++, 32-bit Editions 5.0

This article was previously published under Q68475

SYMPTOMS

The compiler error:

C2118: negative subscript

-or-

C2118: negative subscript or subscript is too large

may be caused by an integer expression in the subscript of an array. If the expression evaluates to greater than INT_MAX, as defined in limits.h, its value becomes negative, thus causing the aforementioned error. For the 16- bit compilers, INT_MAX equals 32767; for 32-bit compilers, INT_MAX equals 2,147,483,647.

CAUSE

This is correct and expected behavior for the Microsoft C Compiler because the evaluation of integer expressions is done using integer math. In some cases, integer math produces an incorrect result because the value of the expression is larger than an integer.

RESOLUTION

16-bit: To work around this situation, add an uppercase (or lowercase) "L" to one of the terms of the expression. This will force the compiler to evaluate the expression using long math which is less likely to overflow.

32-bit (or 16-bit expressions greater than LONG_MAX): Unfortunately, the above resolution will not work as INT_MAX equals LONG_MAX. In these cases, we recommend that the array be broken into several smaller arrays, each of whose size will not exceed INT_MAX.

MORE INFORMATION

If the sample code below is compiled with C/C++ version 7.0 or Visual C++ for Windows, the following error message will be generated:

error C2118: negative subscript

Moreover, if the warning level is set to 2 or above, the following warning message will also be generated:

warning C4307: '*' : integral constant overflow; result truncated

To eliminate the error, add "L" to one of the terms to indicate it is a long constant. For example:

   char mystring[1000L * 33];

Sample Code

/* Compile options needed: <none>; for warning, /W2 or greater
*/ 

char mystring[1000 * 33];

Visual C++ 32-bit edition will not generate an error with the above Sample Code because integers are 4 bytes instead of 2 bytes and the limit for the subscript is 2,147,483,647. However, if we include LIMITS.H and use INT_MAX + 2 as the size of the array, we can generate the desired error.

Modification Type:	Minor	Last Reviewed:	7/5/2005
Keywords:	kbCompiler kbprb KB68475