I've got some Javascript code that I need to parse with Perl:
var materials ={
foo: "bar",
bar: "baz",
baz: "foo"
},
I've got this Javascript variable as a string, and I want to match the associative array's body so that I can parse it as JSON with Perl using parse_json(). I cannot figure out what I'm doing wrong with my regular expression:
my ($json_str) = $js_code =~ m/var\smaterials\s=\s+({.+}),/i;
$json_str ends up being uninitialized.
my ($json_str) = $js_code =~ m/var\smaterials\s=\s*({[\s\S]+?}),/i;
^^^^
The problem is . does not match \n by default.So either use [\s\S] or use (?s) the DOTALL flag.
See demo.
https://regex101.com/r/cJ6zQ3/7
or
[\s\S]+ doesn't match '}' also? (and all the rest till the end of the string) Doesn't it need the lazy modifier '?' to work? Like this [\s\S]+?
} but then it will backtrack as re has to match a } as well.To [\s\S] will backtrack leaving } for } to match.Same with , as well.? is needed only if there are multiple } as then it will match till last }If you can quote the JSON object's keys (like the example below), you could try JSON::Decode::Regexp, the module includes a regex you can use to match a JSON. As a bonus, you get the JSON object loaded as Perl hash. Example code:
use Data::Dump;
use JSON::Decode::Regexp;
my $json_code = <<'_';
var materials ={
"foo": "bar",
"bar": "baz",
"baz": "foo"
},
_
if ($json_code =~ /(\{.+)/s) {
local $_ = $1;
local $^R;
eval { /\A$JSON::Decode::Regexp::FROM_JSON/ } or die "No match";
die "No match: $@" if $@;
print "Match: "; dd $_;
}
will print:
Match: { bar => "baz", baz => "foo", foo => "bar" }
Using an exclude group like [^}]+ works as expected also if there are multiple elements:
#!/usr/bin/env perl
my $js_code = <<'__END__';
var previousOne = {
pFoo: "pBar",
pBar: "pBaz",
pBaz: "pFoo"
},
var materials ={
foo: "bar",
bar: "baz",
baz: "foo"
},
var anotherOne = {
aFoo: "aBar",
aBar: "aBaz",
aBaz: "aFoo"
}
__END__
my ($json_str) = $js_code =~ m/\s*var\s+materials\s*=\s*({[^}]+}),?/;
print "json_str = ${json_str}\n";
I've relaxed some whitespace constraints. You can test it and edit it online here
[\s\S]++) that not backtrack at all. I've removed the wrong assumption but leave the code cause it works.There's no spaces between the equal sign and the curly bracket, but the pattern requires at least one. Remove the \s+ or change it to \s*.
\s*instead of requiring\s+). However, JSON can be more complex than this example. It might be easier to just drop the portion you do not want:s/^[^{]+//. This will strip everything that comes before the first{.JSON.stringify(materials)will give you real JSON to start with, which makes everything a whole lot easier.