estools
Duplicated comments in output from Esprima
When given the Esprima output from program:
var obj =
{
// Pre-a comment
a: function()
{
var foo = 1; // hello
}
,b: function()
{
}
};
// comment above
function bar()
{
// comment below
var i = 1;
}
/**
* This is some documentation for this function
*/
function foo()
{
}
(available here)
escodegen.generate will produce the program:
var obj = {
a: function () {
var foo = 1; // hello
},
b: function () {
}
}; // comment above
// comment above
function bar() {
// comment below
var i = 1;
} /**
* This is some documentation for this function
*/
/**
* This is some documentation for this function
*/
function foo() {
}
The comments seem duplicated - is this the expected behavior for the output from Esprima? I saw that @Constellation did the work in Esprima for attachComments so wondering if this was an intentional duplication in the Esprima AST output.
Could you show me how to create this AST? It seems that AST already has duplicated comments. https://gist.github.com/mduvall/8018844 In this case, this output is expected.
I'm also being affected by this. Is this a bug on escodegen or esprima? Is a PR welcome?
Any word on this @Constellation ? I still see this issue today. The AST is created from a file that has single copies of each comment, but some of these comments are attached to both the trailing comments of a node and the leading comments of the following node, causing escodegen to output duplicates.
I think this is an issue to Esprima because the parsed AST contains duplicate comments.
/cc @ariya @michaelficarra What do you think of?
Ah, @Constellation sorry that's likely the case. Esprima should consider a comment as leading or trailing, but not both. It looks like a bug there because this doesn't happen for all comments that could be interpreted as leading or trailing.
