Problem Regular Expression to analyse text file
3 views (last 30 days)
Show older comments
Sagar Talaviya
on 20 Jan 2018
Commented: Walter Roberson
on 25 Jan 2018
Hello, I want to create given structure from any text file,
Which shows Function argument type, data type, list of variables used and defined, variable data type and whether they are local, global.
here is one .txt file, which contains C code into the file.
#include "add_lib11.h"
double PI_Regler_a(double a,double b, double c, double d, double e, double f)
{
double x;
x = (d*a) + (e*b) + (f*c);
return(x);
}
So, How can I use textscan and regexp to get structure like
functioncall.name= 'PI_Regler_a'
functioncall.type= 'double'
functioncall.returnvalueexists= '1'
variables(1).name = 'a'
variables(1).dtype = 'double'
variable(1).type = 'local'
like that,.....
please help me... I need to know that how I can separate and get what I need by using regexp and inputparsing class.
1 Comment
Accepted Answer
Walter Roberson
on 23 Jan 2018
You need to build a C parser first. regexp() is not powerful enough for that task. You can find a formal grammar up to C11 at http://www.quut.com/c/ANSI-C-grammar-y.html
You would need to enhance this to support #include and all other preprocessing . It might be easier to invoke a C preprocessor to have it emit the expanded code, if doing so would not be considered cheating. Supporting #include and #define and so on is absolutely crucial in C since just about every "word" can be #define to be something other than it appears. And you cannot possibly get types correctly without processing typedef, which are often in #include files.
You might want to look at https://www.mathworks.com/matlabcentral/fileexchange/32769-mparser which is a parser for MATLAB written partly in MATLAB that uses the ANTLR parser (calling that with a mex interface.)
Getting all of this right is a non-trivial project.
Or you could do something half-cooked that would have trouble getting types correct and would have trouble determining whether a variable is local or global.
... Did you forget comments can have things that look like code?
3 Comments
Walter Roberson
on 25 Jan 2018
You need to build a parser routine, such as in yacc and lex, or in ANTLR, or re-invent such routines in MATLAB.
More Answers (0)
See Also
Categories
Find more on Logical in Help Center and File Exchange
Products
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!