How are the startCharacter and endCharacter properties on folding ranges defined?

[msujew]

We are currently in the process of implementing folding support for Xtext. (eclipse/xtext-core#1672)

There has been some discussion regarding the use of the startCharacter and endCharacter properties of the FoldingRange. It's not quite clear how exactly these are defined. Especially since I couldn't find any reference implementation and vscode only uses line based folding, so I couldn't test it either.

Could someone please give some clarification on how these are defined? An actual example would be highly appreciated.

[dbaeumer]

@aeschli can you please clarify. Thanks!

[aeschli]

startLine/startCharacter - endLine/endCharacter define a Range. That's the range that is hidden when that folding range is collapsed.

[msujew]

@aeschli Right, I think my question wasn't clear enough: Are the startCharacter and endCharacter properties indices of the full document string or are they defined as the column of the startLine/endLine?

[KamasamaK]

Based on them defaulting to the length of the relevant line, it should be the column of the line.

[aeschli]

Yes, character is the offset in the line. In LSP all positions are line/character (like line/column except that columns typically start with 1, while characters start with 0)
In editors that only support folding full lines (such as VSCode), the line at startLine gets the folding marker, but when folded, is still fully visible.

You can say it folds after the last character of the first line. That's why the comment says: defaults to the length of the line

[msujew]

I see, thank you for your explanation. I will close this issue.