Non matching end tag in example xml parser not taken into account

[medihack]

It seems that the example XML parser does not take into account if a start and end tag have matching tag names.

When running the spec with the below input text the parser treats it as syntactically correct.

const inputText = 
  '<?xml version="1.0"?>\n' +
  "<foo>\n" +
  "</bar>"

I want to fix this (check that the tag names match). I am a complete newbie using chevrotain, but from looking at the docs it seems look ahead parsing is not enough here as the content of that element is unknown. So I guess I have to use backtracking somehow, maybe together with a parameterized rule. But I don't see how to check this in the subrule then. Is there an example of how to accomplish something like that?

[NaridaL]

@medihack The XML Parser extends CSTParser, so it includes all the parsed tokens in the output and a check whether start and end tags match could be implemented as a second pass.

If you want to check immediately, you could assign the two consumed Name tokens to variables and then check if they match.

if (openName.image !== closeName.image) throw new Error("unexpected closing tag")

[medihack]

@NaridaL Thanks a lot, makes total sense. It seems I thought way too complicated. Greetings from Heidelberg :-)

[bd82]

Hi @medihack

As a generalization of what @NaridaL wrote above:
Chevrotain in a library for syntactic analysis, it could be used to implement part of some language spec
but is not meant to implement all of the spec.

For example have a look at an XML Language Service:

• Parser
• AST Building
• "framework" to perform validations
• Open/Close tags name validation
  • Note this validation works on the AST, not the CST.

[medihack]

@bd82 Thanks so much. These are great resources, especially as my plan is to build a whole language service for my custom language. I learn a lot from your links.