使用 Perl 解析 JavaScript：从 JE->parse 获取“undef”

Question

我有 Javascript 代码，我需要将 javascript 变量解析为 perl 哈希。有没有现成的模块？我尝试了JE::parse() 和JavaScript::HashRef::Decode，但都没有成功。

预期行为：

use Data::Dumper;
use SomeModule::ParseJSVariables qw/decode_js/;

my $str = qq/
var data = {
    'abc': 1,
    'def' : 2
    'xyz' : { 'foo' : 'bar' }
}
/;

my $res = decode_js($str);
warn Dumper $res; #

# expected result: 
# { 
#   name => 'data', 
#   value => {
#     'abc' => 1,
#     'def' => 2
#     'xyz' => { 'foo' => 'bar' }
#     }
# }


use JE;
my $j = new JE;
my $parsed = $j->parse($str);
warn Dumper $parsed; # undef :(

如果没有现成的模块，如果有人建议正确的正则表达式或解析方法，我会很高兴。

UPD，澄清。我有大约数千行 javascript 代码，我只需要获取在全局范围内明确给出的变量内容，例如 var x = { 'foo' : 'bar' }。其他代码可以跳过解析。

我的环境：

$ perl --version

This is perl 5, version 22, subversion 1 (v5.22.1) built for x86_64-linux-gnu-thread-multi

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description:    Ubuntu 16.04.5 LTS
Release:    16.04
Codename:   xenial

$ uname -r
4.19.24-041924-generic

Answer 1

嗯，我有好消息和坏消息，它们是同一个消息 :) 。您的 JS 存在语法错误，因此 JE 根据 the JE docs 返回 undef。具体来说，def 行上缺少 ,。以下测试适用于我：

代码：

use Data::Dumper;

# Note: use q[ ] instead of qq/ /.  q instead of qq so Perl doesn't interpolate
# into the contents, and [ ] instead of / / so that JS comments can appear
# in the block.
my $str = q[
var data = {
    'abc': 1,
    'def' : 2,   // <==== There was a comma missing here!
    'xyz' : { 'foo' : 'bar' }
}
];

use JE;
my $j = new JE;
my $parsed = $j->parse($str);
warn Dumper $parsed;

输出：太大，无法在此处显示 :)。但它确实包括你想要的！

提取输出

这将是一个挑战。希望以下内容可以帮助您入门。

代码：

use Data::Dumper::Compact 'ddc';   # <== for briefer output
use JE;

# Note: use q[ ] instead of qq/ /.  q instead of qq so Perl doesn't interpolate
# into the contents, and [ ] instead of / / so that JS comments can appear
# in the block.
my $str = q[
var data = {
    'abc': 1,
    'def' : 2,   // <==== There was a comma missing here!
    'xyz' : { 'foo' : 'bar' }
}
];

my $j = new JE;
my $parsed = $j->parse($str);
print ddc $parsed->{tree};     # <== {tree} holds the parsed source

输出（带注释）：

bless( [
  [
    0,
    118,
  ],
  "statements",
  bless( [
    [
      1,
      118,
    ],
    "var",
    [
      "data",
      bless( [
        [
          12,
          117,
        ],
        "hash",    <== here's where your hash starts
        "abc",     <== 'abc': 1
        1,
        "def",     <== 'def': 2
        2,
        "xyz",     <== 'xyz': nested hash
        bless( [
          [
            98,
            115,
          ],
          "hash",
          "foo",
          "sbar",
        ], 'JE::Code::Expression' ),
      ], 'JE::Code::Expression' ),
    ],
  ], 'JE::Code::Statement' ),
], 'JE::Code::Statement' )

Answer 2

我找到了最简单的解决方案:)

关键思想是使用JavaScript::V8 或JavaScript::Any 和模拟console.log 函数在新上下文中执行javascript 代码。

my $str = qq/
var data = {
    'abc': 1,
    'def' : 2,
    'xyz' : { 'foo' : 'bar' }
};
/;

use Data::Dumper;
use JavaScript::V8;

sub extract_js_glob_var {
    my ( $code, $var_name ) = @_;
    my $res;
    my $context = JavaScript::V8::Context->new();
    $context->eval($str);
    $context->bind( console_log => sub { $res = @_[0] } );
    $context->eval('console_log('.$var_name.')');
    undef $context;
    return $res;
}

warn Dumper extract_js_glob_var($str, 'data');  # 'data.xyz' is also supported ;)