I am working on building a calculator compiler using Lex and Yacc. The idea is based upon the following resource: http://epaperpress.com/lexandyacc/index.html.
For the given input file I need to identify all of the comments:
//.TEST -- JWJ
//.Step final -- testing all requirements
//.source: test-1m.cal
//.expected output: test-1m_expected.out
/**
* This program will use Newton's method to estimate the roots of
This should be a comment as well, but does not get picked up
* f(x) = x^3 - 3*x
*/
float xn;
float xo;
// int num_iterations;
xo = 3.0;
xn = 3.0;
num_iterations = 1;
/* A do-while loop */
do {
print xo;
xo = xn;
xn = xo - ( xo * xo * xo - 3.0 * xo ) / ( 3.0 * xo * xo - 3.0);
num_iterations = num_iterations + 1;
} while ( num_iterations <= 6 )
print xn; // The root found using Newton's method.
print (xo * xo * xo - 3.0 * xo ); // Print f(xn), which should be 0.
I am using the following regular expressions in my lex file:
"//"[^\n]*|"\/\*".*"\*\/"
"\/\*"([^\n])*
(.)*"\*\/"
I do not understand why multiple line comments are not being matched? Could someone please offer some insight?
The . character in flex matches any character EXCEPT a newline (so its the same as [^\n]). As a result, none of your regexes match any comment containing a newline.
The usual regex for a C-style comment is:
"/*"([^*]|\*+[^*/])*\*+"/"
This matches 0 or more "anything except *" OR "1 or more *s not followed by a * or /" inside the comment markers.