hokein added inline comments.

Comment at: clangd/index/Index.h:32
+    // Character offset on a line in a document (zero-based).
+    int Character = 0;
+  };
sammccall wrote:
> sammccall wrote:
> > Column?
> > 
> > LSP calls this "character" but this is nonstandard and I find it very 
> > confusing with offset. 
> We should document what this is an offset into: bytes, utf-16 code units, or 
> unicode codepoints. (Or even grid offsets - glyphs and doublewidth are a 
> thing)
> Given that we intend to send it over LSP without reading the source, only 
> utf-16 code units is really correct. Unicode codepoints is "nicer" and will 
> give correct results in the BMP, while bytes will be correct for ASCII only.
> I'd vote for making this utf-16 code units.
> It's OK if the code populating it doesn't get this right (confuses bytes and 
> code units) but add a fixme.
Done. Added FIXME.

  rCTE Clang Tools Extra


cfe-commits mailing list

Reply via email to