SyntaxAwareSearch (sas) is a grep-like tool with an understanding of the syntax of C/C++. sas uses a domain specific language to allow very concise but powerful queries of code.

sas currently supports the following kinds of queries. In this document, <regex> represents some ECMAscript regulary expression pattern.

• Variable: <regex>:<regex>- The first regular expression is used to match against the type of the variable. The second matches against the name. If one of the expressions is ommited, it will default to ".*" (i.e., it will match anything).

• Function: <regex>:<regex>(<parameter list>)- The first regular expression is used to match against the return type of the function, the second matches against the name of the function. A parameter match takes the form <regex>:<regex>, where the first expression matches against the type of the parameter and the second matches against the name. An ellipses ... can also be provided to match any number of parameters of arbitrary name and type.

• Class: #<regex> - The regex is used to match against the name of the class.

• Qualifiers: <regex>:: - A qualifier may be attached to any of the above form to restric the class or namespace searched for matches. For example, boost::int:x would match a variable which has a name matching x and a type matching int in the namespace/class boost.

In some cases, a regex may need to include characters which are reserved for the search language (e.g., colon or parenthesis). In these cases, the regular expression may be written between two forward slashes. For example: .*: void (*) () will result in a parsing error, but .*:/void (*) ()/ will match variables (in C/C++) which are function pointers with the appropriate signatures.

Given the following C++ program "demo.cpp":

class DemoClass {
public:
    void member_func() {}
};

int x;

float top_level(){
    float y;
    return 0.0;
}

The following expressions would give the provided results.