More user-controllable verbatim property

[azu]

Hi,

escodegen 1.2.0 generate unnecessary ( ) around verbatim property value.

var code = "fn([1,2,3])";
var ast = esprima.parse(code, {range: true});
// Actually, edit AST ...
function embedVerbatim(node) {
    // embed code
    node["x-verbatim-property"] = code.substring(node.range[0], node.range[1])
}
estraverse.traverse(ast, {
    enter: function enter(node, parent) {
        var fn = {
            "ArrayExpression": embedVerbatim
        }[node.type];
        if (fn) {
            fn(node, parent);
        }
    }
});

escodegen.generate(ast, {
    verbatim: "x-verbatim-property"
}); // => "fn(([1,2,3]));"

I expected result "fn([1,2,3])"

More detail in that project.

• https://github.com/azu/escodegen_array

I want to control the parenthesize.

Thanks.

[Constellation]

What do you think of exposing Precendence?
For example,

node['x-verbatim-property'] = {
    content: 'string',
    precedence: escodegen.Precedence.XXX
};

Pros

• Provides fine grained control over parentheses.

Cons

• Depends on unstable Precedence. Precedence contents may be changed at some time.

[azu]

@Constellation I think it is a good idea to set a option object.
(We can make each object difference Precedence. )

I'm worried about Cons, too.
However, It seems the impact is small.

[Constellation]

Agreed. So I'll expose Precedence.

[Constellation]

I've published version 1.3.0 with this feature.

[azu]

@Constellation Thnaks! great job!

Test passed on example project 💚

azu/escodegen_array@66ccc61

--- a/index.js
+++ b/index.js
@@ -23,7 +23,10 @@ function main(code) {
     function embedVerbatim(node) {
         var embed = helper.getCodeFromRange(node.range);
         if (embed) {
-            node["x-verbatim-property"] = embed
+            node["x-verbatim-property"] = {
+                content : embed,
+                precedence : escodegen.Precedence.Primary
+            }
         }
     }