Fixed bugs in Spec parser for terminals (#145)

[jhnaldo]

I fixed a bug in the Spec parser for terminals.
The PR tc39/ecma262#2418 in ecma262 defines a new terminal: `` as follows:

      <emu-grammar type="definition">
        ClassSetReservedDoublePunctuator :: one of
          `&amp;&amp;` `!!` `##`
          `$$` `%%` `**`
          `++` `,,` `..`
          `::` `;;` `&lt;&lt;`
          `==` `&gt;&gt;` `??`
          `@@` `^^` `&grave;&grave;`
          `~~`
      </emu-grammar>

However, the current Spec parser for terminals only supports a single quote `. So, I fixed this problem by extending the Spec parser to support one or more repetitions of `.

[michaelficarra]

Shouldn't 1 be replaced with the length of the delimiter here?

[jmdyck]

When would the length of the delimiter not be 1?

[michaelficarra]

I'm unsure what 4 or more backticks in a row is supposed to mean. It doesn't appear to be legal. So I don't know.

[jmdyck]

ecmarkup accepts 4 backticks and renders it as a literal terminal consisting of two backticks: 2418 preview

[michaelficarra]

I'm reading the code for grammarkdown and it doesn't seem it would support 4 backticks in a row, though I could be misreading. Does it have something to do with tc39/ecma262#2418 using ` and the entities being processed by esmeta but not grammarkdown?

[jhnaldo]

ESMeta unescapes the HTML entities before parsing the spec.html. Thus, the text ```` is unescaped as: ````, and ESMeta parses the four backticks (````) as a literal terminal consists of two backticks (``) because the first and last backtick is used to represent literal terminals. I supported this feature because tc39/ecma262#2418 defines the literal terminal consisting of two backticks. Is there any problem with this extension?

[michaelficarra]

Okay, that should be fine then. I didn't know about the pre-processing of HTML entities by esmeta.

[michaelficarra]

The docs for grammarkdown don't mention this: https://github.com/rbuckton/grammarkdown#terminals

[jmdyck]

There are a few things (used by the ES spec) that the grammarkdown docs don't mention or aren't clear on. I think that's more of an issue for grammarkdown than for esmeta.