...
Noncompliant code examples are typically followed by compliant solutions, which show how the noncompliant code example can be recoded in a secure, compliant manner. Except where noted, noncompliant code examples should contain violations only of the guideline under discussion. Compliant solutions should comply with all of the secure coding rules but may on occasion fail to comply with a recommendation.
Coding Conventions
Unless otherwise specified, all code should compile on a reasonably modern compiler, when it is following compliance with the standard. For example, you can require GCC to conform to the C11 standard with the parameter --std=c11.
Code that is only expected to run on a particular subset of platforms should have those platforms mentioned in the code's section header, e.g.: Compliant Solution (POSIX). Likewise, code that is only expected to run on more modern versions of C should indicate the oldest standard that supports them, e.g.: Compliant Solution (C99).
In order to compile the code, you will need to include appropriate header files. For example, if the code invokes malloc(), you may need to include the stdlib.h header.
Many coding examples must lead with some variables initialized to valid values. If any code example (compliant or non-compliant) requires a local variable to be initialized with an unspecified but valid value, then that code should be wrapped inside a function that takes the variable as a function argument. Consequently, unless otherwise noted, function arguments should be assumed to point to valid values. For example, character pointers should not be NULL, but rather point to valid null-terminated byte strings. Checking that function arguments are valid is important, and is described in detail in recommendation API00-C. Functions should validate their parameters. However, the code involved with checking argument validity would be redundant across all code examples, and so function arguments in these code examples are not validated.
Many code examples will contain ellipsis in comments. This indicates that the comment may be replaced by arbitrary code that satisfies the comment. A comment with only ellipsis suggests that the code may do anything.
Proper error handling is a controversial subject, and many applications and libraries provide their own idiosyncratic error handling mechanisms. See Rule 12. Error Handling (ERR) and Rec. 12. Error Handling (ERR) for our guidelines on handling errors. When our code detects that an error condition might have occurred, and handling that error condition is not endemic to the guideline itself, we will use the comment: /* Handle Error */. This comment implies that the error is somehow addressed, so that the code does not fall through. The code may abort, or fix the error somehow. For example:
| Code Block | ||||
|---|---|---|---|---|
| ||||
char *str = malloc(10);
if (str == NULL) {
/* Handle Error */
}
/* ... str can not be NULL here. Work with str... */ |
Exceptions
Any rule or recommendation may specify a small set of exceptions detailing the circumstances under which the guideline is not necessary to ensure the safety, reliability, or security of software. Exceptions are informative only and are not required to be followed.Exceptions
Risk Assessment
Each guideline in the CERT C Coding Standard contains a risk assessment section that attempts to provide software developers with an indication of the potential consequences of not addressing a particular rule or recommendation in their code (along with some indication of expected remediation costs). This information may be used to prioritize the repair of rule violations by a development team. The metric is designed primarily for remediation projects. It is generally assumed that new code will be developed to be compliant with the entire coding standard and applicable recommendations.
...
Regarding partial overlap, we try to find segments of code as examples that are inseparable and exhibit both code flaws. An example of separable code:
This The following line violates rules about integer overflow and floating-point overflow, but that does not mean that the rules about integer overflow and fp-overflow overlap:
...
Most guidelines have a small bibliography section that lists documents and sections in those documents that provide information relevant to the guideline.