How can I ignore C comments when I process a C source file with Perl?

I'm running a code that read files, do some parsing, but need to ignore all comments. There are good explanations how to conduct it, like the answer to How can I strip multiline C comments from a file using Perl?

$/ = undef;
$_ = <>;
s#/\*[^*]*\*+([^/*][^*]*\*+)*/|("(\\.|[^"\\])*"|'(\\.|[^'\\])*'|.[^/"'\\]*)#defined $2 ? $2 : ""#gse;
print;

My first problem is that after run this line $/ = undef; my code doesn't work properly. Actually, I don't know what it does. But if I could turn it back after ignoring all comments it will be helpful.

In general, What is the useful way to ignore all comments wi开发者_StackOverflowthout changing the rest of the code?

---

awk

$ cat file.c
one
two
three // comment at the back
// comment in front
four /* another comment */
/* comment spanning
   multiple
   lines
*/  five
six
seven

$ awk -vRS='*/' '{ gsub(/\/\*.*/,"");gsub("//.*","")}1' file.c
one
two
three


  five
six
seven

the awk command sets the record separator RS to */, which is the ending tag for the multiline style comment. so it iterates the records, checking for /*, the opening tag, and then get whatever is in front of /*. this concept is simple, and you don't have to craft out complicated regex for this. Similar, if you were to do it with Python,

>>> data=open("file").read() 
>>> for item in data.split("*/"):
...     if "//" in item: item=item.split("//")[0]
...     if "/*" in item: item=item.split("/*")[0]
...     print item
...
one
two
three


  five
six
seven

---

You want to make $/ local, as in

$_ = do { local $/; <> };

or

{
    local $/;
    $_ = <>;
    #...
}

Alternately, you could use File::Slurp

---

If you are stripping "nested" comments, i.e.:

/* This is a comment 
/* that has been re-commented */ possibly /* due to */ 
various modifications */

regexp may not be the best solution. Especially if this spans multiple lines as in the example above.

Last time I had to do something like this, I read the lines one at a time, keeping a count of how many levels of "/*" (or whatever the delimiter was for the specific language) and not printing anything unless the count was at 0.

Here is an example - I apologize in advance because it's pretty bad Perl, but this should give you an idea, at least:

use strict;

my $infile = $ARGV[0]; # File name

# Slurp up input file in an array
open (FH, "< $infile")  or die "Opening: $infile";
my @INPUT_ARRAY = <FH>;
my @ARRAY;
my ($i,$j);
my $line;


# Removes all kind of comments (single-line, multi-line, nested).
# Further parsing will be carried on the stripped lines (in @ARRAY) but
# the error messaging routine will reference the original @INPUT_ARRAY
# so line fragments may contain comments.
my $commentLevel = 0;

for ($i=0; $i < @INPUT_ARRAY; $i++)
{
    my @explodedLine = split(//,$INPUT_ARRAY[$i]);
    my $resultLine ="";

    for ($j=0; $j < @explodedLine; $j++)
    {
        if ($commentLevel > 0)
        {
            $resultLine .= " ";
        }
        if ($explodedLine[$j] eq "/" && $explodedLine[($j+1)] eq "*")
        {
                $commentLevel++;
                next;
        }           
        if ($explodedLine[$j] eq "*" && $explodedLine[($j+1)] eq "/")
        {
                $commentLevel--;
                $j++;
                next;
        }       
        if (($commentLevel == 0) || ($explodedLine[$j] eq "\n"))
        {
            $resultLine .= $explodedLine[$j];
        }
    }

 $ARRAY[$i]=join(" ",$resultLine);  
}   


close(FH)   or die "Closing: $!";