Regex to get string from function param?

Question

Can anyone help me with regex problem. Im making a script to go through all my .php files and get all strings passed to certain function. I need to match this cases:

/* Double quotes */
function("some string"); // Match: some string
function("some \"string\""); // Match: some "string"
function("some 'string'"); // Match: some 'string'

/* Single quotes */
function('some string'); // Match: some string
function('some \'string\''); // Match: some 'string'
function('some "string"'); // Match: some "string"

Function can also accept parameters after string, so it also needs to match this cases:

/* Additional parameters */
function("some string", "param"); // Match: some string
function("some string", $param); // Match: some string

So essentially, param can be either a string (quoted or double quoted) or unquoted variable. I need to get string only from first parameter of function, regardless if second parameter exists or is quoted in any way...

Thanks in advance...

This might help you: http://stackoverflow.com/questions/2630812/get-a-list-of-function-names-in-a-shell-script — Markus Steindl, Dec 21 '12 at 13:54
I managed only to get string between 2 same quotes (single or double) with this: [single quote]'/\_\_\(\'(.*?)\'/i' and [double]'/\_\_\(\"(.*?)\"/i' — pajcho, Dec 21 '12 at 13:58

melpomene · Answer 1 · 2012-12-21T14:12:11.697

0

Here's a quick sketch that might help get you started:

while (readline) {
    my ($matched) = m{
        \b function \s* \( \s*
        (
            " (?: [^"\\] | \\ .)* "
        |
            ' (?: [^'\\] | \\ .)* '
        )
    }sx or next;
    my $value = php_unescape $matched; # XXX: write me
    print $value, "\n";
}

edited Dec 21 '12 at 14:12

answered Dec 21 '12 at 13:56

melpomene

84,125
8
85
148

raina77ow · Accepted Answer · 2012-12-21T14:08:54.150

0

Your particular example was successfully processed with...

preg_match_all('#\\(\\s*("((\\\\.|[^"])+)"|\'((\\\\.|[^\'])+)\'),?#s', $test, $matches);

Here's ideone demo.

Explanation: we try to match opening parenthesis (thankfully, it's PHP; it'd be far more difficult in Ruby), followed by any number of whitespace characters, followed by...

either "(\\.|[^"])+"
or '(\\.|[^'])+'

... followed by optional comma.

Each of this sequences covers both 'special characters' (escaped with slash) and 'normal ones' (that are not the same as delimiters).

edited Dec 21 '12 at 14:08

answered Dec 21 '12 at 14:02

raina77ow

103,633
15
192
229

there is no need to double all the backslashes when using single-quoted strings. only the four backslashes for the literal one need to remain in place. `\\(\\s` can become `\(\s` – Martin Ender Dec 21 '12 at 19:56

score 0 · Answer 3 · edited May 23 '17 at 12:04

Instead of using a regular expression (yourself) you could use a php parser that gives you some kind of AST, e.g. the one accepted as answer at Generate AST of a PHP source file

<?php
require 'path/to/PHP-Parser-master/lib//bootstrap.php';

class MyNodeVisitor extends PHPParser_NodeVisitorAbstract
{
    public function beforeTraverse(array $nodes) {}
    public function enterNode(PHPParser_Node $node) { }
    public function leaveNode(PHPParser_Node $node) { 
        if ($node instanceof PHPParser_Node_Expr_FuncCall) {
            if ( 'foo'===(string)$node->name ) {
                foreach( $node->args as $arg ) {
                    echo $arg->value->value, "\n";
                }
            }

        }
    }
    public function afterTraverse(array $nodes) {}
}


$parser = new PHPParser_Parser(new PHPParser_Lexer);
$nv = new MyNodeVisitor;
$traverser = new PHPParser_NodeTraverser;
$traverser->addVisitor($nv);


try {
    $stmts = $parser->parse( data() );
        $stmts = $traverser->traverse($stmts);

} catch (PHPParser_Error $e) {
    echo 'Parse Error: ', $e->getMessage();
}





function data() {
    return <<< eot
<?php   
/* Double quotes */
foo("some string"); // Match: some string
foo("some \"string\""); // Match: some "string"
foo("some 'string'"); // Match: some 'string'

/* Single quotes */
foo('some string'); // Match: some string
foo('some \'string\''); // Match: some 'string'
foo('some "string"'); // Match: some "string"   
eot;
}

prints

some string
some "string"
some 'string'
some string
some 'string'
some "string"

score 0 · Answer 4 · answered Dec 21 '12 at 14:27

Here I wrote a script in sed. Save it in a file file.sed

bs
:gf
s:,.*$::
s:^.::
s;.$;;
s:[\]\(["']\):\1:g
p;d
:s
/.*(\([^)]*\).*/ s::\1:
tgf
d

Next, run it sed -f file.sed FILE.py:

for x in `find -name \*.py`; do sed -f file.sed $x; done

edit:

one can replace the script with oneliner sed command, but calling it so it is much clear to debug what it does.

Regex to get string from function param?

4 Answers4