Add JSX text nodes into skip list of walker (fixes #2007)

[IllusionMH]

PR checklist

• Addresses an existing issue: comment-format TSX false positive #2007
• New feature, bugfix, or enhancement
  • Includes tests
• Documentation update
• Enable CircleCI for your fork (https://circleci.com/add-projects)

What changes did you make?

This PR adds JsxText nodes into skip map in SkippableTokenAwareRuleWalker.

Is there anything you'd like reviewers to focus on?

1. This change should definitely skip JsxText nodes in EnableDisableRulesWalker, CommentWalker and JsdocWalker, but I'm unsure about WhitespaceWalker and NoTrailingWhitespaceWalker.

2. If JsxText node consist of whitespace and "comment" it isn't added to skip map since getStart() and getEnd() return same value. See test case

[ajafff]

I'm unsure about WhitespaceWalker

WhitespaceWalker skips Jsx(Selfclosing)Element anyways, so skipping JsxText may not be a big deal here. I don't know why they are skipped though. Maybe to solve a similar problem like this PR?

[IllusionMH]

WhitespaceWalker most likely won't be affected by this change, but NoTrailingWhitespaceWalker might have some logic here and will require separate implementation.
Looks like NoTrailingWhitespaceWalker already reported trailing whitespaces on JSX text nodes so I restored its functionality. But implementation isn't very nice.

Still there is a question: if there is need to support text nodes that consist of whitespaces and text that matches comment format?

[IllusionMH]

Is there other way to restore super.super.visitJsxText functionality?

[nchen63]

how about adding a constructor to SkippableTokenAwareRuleWalker with a parameter skipJsx?

[ajafff]

Maybe it would be sufficient to just pass the correct ts.LanguageVariant to ts.createScanner. Currently on LanguageVariant.Standard is used, so the scanner doesn't know how to parse jsx. In this case LanguageVariant.JSX is needed.
The easiest way would be to use sourceFile.languageVariant.

[IllusionMH]

@ajafff Thanks for suggestion.
I tried to use sourceFile.languageVariant as well as LanguageVariant.JSX in scanner initialization, and haven't found any difference at first glance.

I'll try to dive deep into TS scanner difference between language variants and other ways to handle text nodes that has only comment text.

[IllusionMH]

Fixed in #2036.

[nchen63]

superseded by #2036