Remove unnecessary catch which caused confusing error when iteration scopes were missing

[AndreasArvidsson]

Today's implementation when you say something like take every item will show an error message Couldn't find containing collectionItem. even if it's the iteration scope that is missing.

This change makes it so that we get the error message Couldn't find containing iteration scope for collectionItem. instead. This greatly helps us with diagnosing what actually went wrong.

For program languages were a scope is defined in Tree sitter the current error message makes little sense. You could use take scope and it works fine, but when you say take every scope you get an error message that there is no scope. Much better to actually show that this language is missing an iteration scope.

Checklist

• [/] I have added tests
• [/] I have updated the docs and cheatsheet
• [/] I have not broken the cheatsheet

[phillco]

This basically reverts the logic added in https://github.com/cursorless-dev/cursorless/pull/2081/files#diff-a47f14c27291690e54ca5b2e034c62193e9da0615bdb7b3be4153ede7668914f.

It was necessary when there was this fallback code after it:

    if (scopes.length === 0) {
      if (scopeType.type === "collectionItem") {
        // For `collectionItem`, fall back to generic implementation
        return this.modifierStageFactory
          .getLegacyScopeStage(this.modifier)
          .run(target);
      }

      throw new NoContainingScopeError(scopeType.type);
    }

We have since removed the legacy code but didn't know that this code also needed to be removed