Negotiate LSP Position Encoding (#5894)

So far LSP always required that `PositionEncoding.characters` is an
UTF-16 offset. Now that LSP 3.17 is available in `lsp-types` request
the server to send char offsets (UTF-32) or byte offsets (UTF-8)
instead. For compatability with old servers, UTF-16 remains as the
fallback as required by the standard.

1 files changed, 17 insertions, 5 deletions

diff --git a/helix-lsp/src/lib.rs b/helix-lsp/src/lib.rs
index bcaff10a..72456b37 100644
--- a/helix-lsp/src/lib.rs
+++ b/helix-lsp/src/lib.rs
@@ -20,7 +20,6 @@ use std::{
     },
 };
 
-use serde::{Deserialize, Serialize};
 use thiserror::Error;
 use tokio_stream::wrappers::UnboundedReceiverStream;
 
@@ -45,13 +44,14 @@ pub enum Error {
     Other(#[from] anyhow::Error),
 }
 
-#[derive(Clone, Copy, Debug, Serialize, Deserialize)]
+#[derive(Clone, Copy, Debug, Default)]
 pub enum OffsetEncoding {
     /// UTF-8 code units aka bytes
-    #[serde(rename = "utf-8")]
     Utf8,
+    /// UTF-32 code units aka chars
+    Utf32,
     /// UTF-16 code units
-    #[serde(rename = "utf-16")]
+    #[default]
     Utf16,
 }
 
@@ -168,6 +168,11 @@ pub mod util {
                 let line_end = line_end_char_index(&doc.slice(..), pos_line);
                 doc.char_to_utf16_cu(line_start)..doc.char_to_utf16_cu(line_end)
             }
+            OffsetEncoding::Utf32 => {
+                let line_start = doc.line_to_char(pos_line);
+                let line_end = line_end_char_index(&doc.slice(..), pos_line);
+                line_start..line_end
+            }
         };
 
         // The LSP spec demands that the offset is capped to the end of the line
@@ -177,10 +182,10 @@ pub mod util {
             .unwrap_or(line.end)
             .min(line.end);
 
-        // TODO prefer UTF32/char indices to avoid this step
         match offset_encoding {
             OffsetEncoding::Utf8 => doc.try_byte_to_char(pos).ok(),
             OffsetEncoding::Utf16 => doc.try_utf16_cu_to_char(pos).ok(),
+            OffsetEncoding::Utf32 => Some(pos),
         }
     }
 
@@ -207,6 +212,13 @@ pub mod util {
 
                 lsp::Position::new(line as u32, col as u32)
             }
+            OffsetEncoding::Utf32 => {
+                let line = doc.char_to_line(pos);
+                let line_start = doc.line_to_char(line);
+                let col = pos - line_start;
+
+                lsp::Position::new(line as u32, col as u32)
+            }
         }
     }