fix: extract style/script tag followed by destructuring in the template

[jasonlyu123]

#2854 #2155

This builds upon the original logic of #2914. Replacing the isInsideMoustacheTag logic with a simple bracket match with a simplified string check. Using the Svelte compiler in this preprocessor isn't really viable because even the loose parse mode still throws a syntax error with an expression while editing.

This also rewrites the parseHTML to use a custom parse logic instead of preprocess. Previously, we scanned the HTML to remove problematic "<" and ">". Now, it does one scan/parse. This should improve performance and might reduce memory usage since V8 won't store the preprocessed string in memory for the sliced string. The parse code is derived from vscode-html-languageservice, andthe main difference is

1. Parse expression tag in Whitespace state
2. parse attribute with interpolation in AttributeValue state
3. Detect Svelte blocks/tags in Content state

@abdel-17 Are there other problems you encountered like this?

@paoloricciuti, since you also commented #2914, could you check if there are other situations that aren't covered by this? This function is used to blank < and > in HTML that confuses the vscode-html-languageservice's HTML parser.

[changeset-bot]

🦋 Changeset detected

Latest commit: 97908d5

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
svelte-language-server	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[abdel-17]

none that I can reproduce no

[abdel-17]

Hey I just wanted to say I really appreciate you working on this issue. Looking forward to seeing it fixed. :)

[dummdidumm]

nice!