Auto merge of #13484 - lnicola:position-encoding, r=lnicola

Switch to upstream `positionEncoding`

Closes #13481

This drops support for the custom extension, but that's probably fine.

Draft because it's not tested yet.

Cargo.lock

@@ -872,9 +872,9 @@ dependencies = [
 
 [[package]]
 name = "lsp-types"
-version = "0.93.1"
+version = "0.93.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "a3bcfee315dde785ba887edb540b08765fd7df75a7d948844be6bf5712246734"
+checksum = "9be6e9c7e2d18f651974370d7aff703f9513e0df6e464fd795660edc77e6ca51"
 dependencies = [
  "bitflags",
  "serde",

crates/rust-analyzer/Cargo.toml

@@ -23,7 +23,7 @@ crossbeam-channel = "0.5.5"
 dissimilar = "1.0.4"
 itertools = "0.10.5"
 scip = "0.1.1"
-lsp-types = { version = "0.93.1", features = ["proposed"] }
+lsp-types = { version = "=0.93.2", features = ["proposed"] }
 parking_lot = "0.12.1"
 xflags = "0.3.0"
 oorandom = "11.1.3"

crates/rust-analyzer/src/bin/main.rs

@@ -11,7 +11,7 @@ use std::{env, fs, path::Path, process};
 
 use lsp_server::Connection;
 use project_model::ProjectManifest;
-use rust_analyzer::{cli::flags, config::Config, from_json, lsp_ext::supports_utf8, Result};
+use rust_analyzer::{cli::flags, config::Config, from_json, Result};
 use vfs::AbsPathBuf;
 
 #[cfg(all(feature = "mimalloc"))]
@@ -191,11 +191,7 @@ fn run_server() -> Result<()> {
             name: String::from("rust-analyzer"),
             version: Some(rust_analyzer::version().to_string()),
         }),
-        offset_encoding: if supports_utf8(config.caps()) {
-            Some("utf-8".to_string())
-        } else {
-            None
-        },
+        offset_encoding: None,
     };
 
     let initialize_result = serde_json::to_value(initialize_result).unwrap();

crates/rust-analyzer/src/caps.rs

@@ -6,19 +6,25 @@ use lsp_types::{
     FileOperationFilter, FileOperationPattern, FileOperationPatternKind,
     FileOperationRegistrationOptions, FoldingRangeProviderCapability, HoverProviderCapability,
     ImplementationProviderCapability, InlayHintOptions, InlayHintServerCapabilities, OneOf,
-    RenameOptions, SaveOptions, SelectionRangeProviderCapability, SemanticTokensFullOptions,
-    SemanticTokensLegend, SemanticTokensOptions, ServerCapabilities, SignatureHelpOptions,
-    TextDocumentSyncCapability, TextDocumentSyncKind, TextDocumentSyncOptions,
-    TypeDefinitionProviderCapability, WorkDoneProgressOptions,
+    PositionEncodingKind, RenameOptions, SaveOptions, SelectionRangeProviderCapability,
+    SemanticTokensFullOptions, SemanticTokensLegend, SemanticTokensOptions, ServerCapabilities,
+    SignatureHelpOptions, TextDocumentSyncCapability, TextDocumentSyncKind,
+    TextDocumentSyncOptions, TypeDefinitionProviderCapability, WorkDoneProgressOptions,
     WorkspaceFileOperationsServerCapabilities, WorkspaceServerCapabilities,
 };
 use serde_json::json;
 
 use crate::config::{Config, RustfmtConfig};
+use crate::lsp_ext::supports_utf8;
 use crate::semantic_tokens;
 
 pub fn server_capabilities(config: &Config) -> ServerCapabilities {
     ServerCapabilities {
+        position_encoding: if supports_utf8(config.caps()) {
+            Some(PositionEncodingKind::UTF8)
+        } else {
+            None
+        },
         text_document_sync: Some(TextDocumentSyncCapability::Options(TextDocumentSyncOptions {
             open_close: Some(true),
             change: Some(TextDocumentSyncKind::INCREMENTAL),

crates/rust-analyzer/src/cli/lsif.rs

@@ -20,7 +20,7 @@ use crate::cli::{
     load_cargo::{load_workspace, LoadCargoConfig},
     Result,
 };
-use crate::line_index::{LineEndings, LineIndex, OffsetEncoding};
+use crate::line_index::{LineEndings, LineIndex, PositionEncoding};
 use crate::to_proto;
 use crate::version::version;
 
@@ -126,7 +126,7 @@ impl LsifManager<'_> {
         let line_index = self.db.line_index(file_id);
         let line_index = LineIndex {
             index: line_index,
-            encoding: OffsetEncoding::Utf16,
+            encoding: PositionEncoding::Utf16,
             endings: LineEndings::Unix,
         };
         let range_id = self.add_vertex(lsif::Vertex::Range {
@@ -248,7 +248,7 @@ impl LsifManager<'_> {
         let line_index = self.db.line_index(file_id);
         let line_index = LineIndex {
             index: line_index,
-            encoding: OffsetEncoding::Utf16,
+            encoding: PositionEncoding::Utf16,
             endings: LineEndings::Unix,
         };
         let result = folds

crates/rust-analyzer/src/cli/scip.rs

@@ -5,7 +5,7 @@ use std::{
     time::Instant,
 };
 
-use crate::line_index::{LineEndings, LineIndex, OffsetEncoding};
+use crate::line_index::{LineEndings, LineIndex, PositionEncoding};
 use hir::Name;
 use ide::{
     LineCol, MonikerDescriptorKind, StaticIndex, StaticIndexedFile, TextRange, TokenId,
@@ -91,7 +91,7 @@ impl flags::Scip {
 
             let line_index = LineIndex {
                 index: db.line_index(file_id),
-                encoding: OffsetEncoding::Utf8,
+                encoding: PositionEncoding::Utf8,
                 endings: LineEndings::Unix,
             };
 

crates/rust-analyzer/src/config.rs

@@ -32,7 +32,7 @@ use vfs::AbsPathBuf;
 use crate::{
     caps::completion_item_edit_resolve,
     diagnostics::DiagnosticsMapConfig,
-    line_index::OffsetEncoding,
+    line_index::PositionEncoding,
     lsp_ext::{self, supports_utf8, WorkspaceSymbolSearchKind, WorkspaceSymbolSearchScope},
 };
 
@@ -948,11 +948,11 @@ impl Config {
         .is_some()
     }
 
-    pub fn offset_encoding(&self) -> OffsetEncoding {
+    pub fn position_encoding(&self) -> PositionEncoding {
         if supports_utf8(&self.caps) {
-            OffsetEncoding::Utf8
+            PositionEncoding::Utf8
         } else {
-            OffsetEncoding::Utf16
+            PositionEncoding::Utf16
         }
     }
 

crates/rust-analyzer/src/diagnostics/to_proto.rs

@@ -8,7 +8,7 @@ use stdx::format_to;
 use vfs::{AbsPath, AbsPathBuf};
 
 use crate::{
-    global_state::GlobalStateSnapshot, line_index::OffsetEncoding, lsp_ext,
+    global_state::GlobalStateSnapshot, line_index::PositionEncoding, lsp_ext,
     to_proto::url_from_abs_path,
 };
 
@@ -66,17 +66,17 @@ fn location(
     let uri = url_from_abs_path(&file_name);
 
     let range = {
-        let offset_encoding = snap.config.offset_encoding();
+        let position_encoding = snap.config.position_encoding();
         lsp_types::Range::new(
-            position(&offset_encoding, span, span.line_start, span.column_start),
-            position(&offset_encoding, span, span.line_end, span.column_end),
+            position(&position_encoding, span, span.line_start, span.column_start),
+            position(&position_encoding, span, span.line_end, span.column_end),
         )
     };
     lsp_types::Location::new(uri, range)
 }
 
 fn position(
-    offset_encoding: &OffsetEncoding,
+    position_encoding: &PositionEncoding,
     span: &DiagnosticSpan,
     line_offset: usize,
     column_offset: usize,
@@ -93,9 +93,9 @@ fn position(
             };
         }
         let mut char_offset = 0;
-        let len_func = match offset_encoding {
-            OffsetEncoding::Utf8 => char::len_utf8,
-            OffsetEncoding::Utf16 => char::len_utf16,
+        let len_func = match position_encoding {
+            PositionEncoding::Utf8 => char::len_utf8,
+            PositionEncoding::Utf16 => char::len_utf16,
         };
         for c in line.text.chars() {
             char_offset += 1;

crates/rust-analyzer/src/from_proto.rs

@@ -8,7 +8,7 @@ use vfs::AbsPathBuf;
 use crate::{
     from_json,
     global_state::GlobalStateSnapshot,
-    line_index::{LineIndex, OffsetEncoding},
+    line_index::{LineIndex, PositionEncoding},
     lsp_ext,
     lsp_utils::invalid_params_error,
     Result,
@@ -25,10 +25,10 @@ pub(crate) fn vfs_path(url: &lsp_types::Url) -> Result<vfs::VfsPath> {
 
 pub(crate) fn offset(line_index: &LineIndex, position: lsp_types::Position) -> Result<TextSize> {
     let line_col = match line_index.encoding {
-        OffsetEncoding::Utf8 => {
+        PositionEncoding::Utf8 => {
             LineCol { line: position.line as u32, col: position.character as u32 }
         }
-        OffsetEncoding::Utf16 => {
+        PositionEncoding::Utf16 => {
             let line_col =
                 LineColUtf16 { line: position.line as u32, col: position.character as u32 };
             line_index.index.to_utf8(line_col)

crates/rust-analyzer/src/global_state.rs

@@ -383,7 +383,7 @@ impl GlobalStateSnapshot {
     pub(crate) fn file_line_index(&self, file_id: FileId) -> Cancellable<LineIndex> {
         let endings = self.vfs.read().1[&file_id];
         let index = self.analysis.file_line_index(file_id)?;
-        let res = LineIndex { index, endings, encoding: self.config.offset_encoding() };
+        let res = LineIndex { index, endings, encoding: self.config.position_encoding() };
         Ok(res)
     }
 

crates/rust-analyzer/src/line_index.rs

@@ -7,7 +7,7 @@
 
 use std::sync::Arc;
 
-pub enum OffsetEncoding {
+pub enum PositionEncoding {
     Utf8,
     Utf16,
 }
@@ -15,7 +15,7 @@ pub enum OffsetEncoding {
 pub(crate) struct LineIndex {
     pub(crate) index: Arc<ide::LineIndex>,
     pub(crate) endings: LineEndings,
-    pub(crate) encoding: OffsetEncoding,
+    pub(crate) encoding: PositionEncoding,
 }
 
 #[derive(Debug, Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Hash)]

crates/rust-analyzer/src/lsp_ext.rs

@@ -3,6 +3,7 @@
 use std::{collections::HashMap, path::PathBuf};
 
 use lsp_types::request::Request;
+use lsp_types::PositionEncodingKind;
 use lsp_types::{
     notification::Notification, CodeActionKind, DocumentOnTypeFormattingParams,
     PartialResultParams, Position, Range, TextDocumentIdentifier, WorkDoneProgressParams,
@@ -455,7 +456,15 @@ pub(crate) enum CodeLensResolveData {
 }
 
 pub fn supports_utf8(caps: &lsp_types::ClientCapabilities) -> bool {
-    caps.offset_encoding.as_deref().unwrap_or_default().iter().any(|it| it == "utf-8")
+    match &caps.general {
+        Some(general) => general
+            .position_encodings
+            .as_deref()
+            .unwrap_or_default()
+            .iter()
+            .any(|it| it == &PositionEncodingKind::UTF8),
+        _ => false,
+    }
 }
 
 pub enum MoveItem {}

crates/rust-analyzer/src/lsp_utils.rs

@@ -6,7 +6,7 @@ use lsp_server::Notification;
 use crate::{
     from_proto,
     global_state::GlobalState,
-    line_index::{LineEndings, LineIndex, OffsetEncoding},
+    line_index::{LineEndings, LineIndex, PositionEncoding},
     LspError,
 };
 
@@ -140,7 +140,7 @@ pub(crate) fn apply_document_changes(
         index: Arc::new(ide::LineIndex::new(old_text)),
         // We don't care about line endings or offset encoding here.
         endings: LineEndings::Unix,
-        encoding: OffsetEncoding::Utf16,
+        encoding: PositionEncoding::Utf16,
     };
 
     // The changes we got must be applied sequentially, but can cross lines so we

crates/rust-analyzer/src/to_proto.rs

@@ -21,7 +21,7 @@ use crate::{
     cargo_target_spec::CargoTargetSpec,
     config::{CallInfoConfig, Config},
     global_state::GlobalStateSnapshot,
-    line_index::{LineEndings, LineIndex, OffsetEncoding},
+    line_index::{LineEndings, LineIndex, PositionEncoding},
     lsp_ext,
     lsp_utils::invalid_params_error,
     semantic_tokens, Result,
@@ -30,8 +30,8 @@ use crate::{
 pub(crate) fn position(line_index: &LineIndex, offset: TextSize) -> lsp_types::Position {
     let line_col = line_index.index.line_col(offset);
     match line_index.encoding {
-        OffsetEncoding::Utf8 => lsp_types::Position::new(line_col.line, line_col.col),
-        OffsetEncoding::Utf16 => {
+        PositionEncoding::Utf8 => lsp_types::Position::new(line_col.line, line_col.col),
+        PositionEncoding::Utf16 => {
             let line_col = line_index.index.to_utf16(line_col);
             lsp_types::Position::new(line_col.line, line_col.col)
         }
@@ -1394,7 +1394,7 @@ fn main() {
         let line_index = LineIndex {
             index: Arc::new(ide::LineIndex::new(text)),
             endings: LineEndings::Unix,
-            encoding: OffsetEncoding::Utf16,
+            encoding: PositionEncoding::Utf16,
         };
         let converted: Vec<lsp_types::FoldingRange> =
             folds.into_iter().map(|it| folding_range(text, &line_index, true, it)).collect();

docs/dev/lsp-extensions.md

@@ -1,5 +1,5 @@
 <!---
-lsp_ext.rs hash: 7b710095d773b978
+lsp_ext.rs hash: 62068e53ac202dc8
 
 If you need to change the above hash to make the test pass, please check if you
 need to adjust this doc as well and ping this issue:
@@ -19,12 +19,6 @@ Requests, which are likely to always remain specific to `rust-analyzer` are unde
 
 If you want to be notified about the changes to this document, subscribe to [#4604](https://github.com/rust-lang/rust-analyzer/issues/4604).
 
-## UTF-8 offsets
-
-rust-analyzer supports clangd's extension for opting into UTF-8 as the coordinate space for offsets (by default, LSP uses UTF-16 offsets).
-
-https://clangd.llvm.org/extensions.html#utf-8-offsets
-
 ## Configuration in `initializationOptions`
 
 **Upstream Issue:** https://github.com/microsoft/language-server-protocol/issues/567

lib/lsp-server/Cargo.toml

@@ -13,4 +13,4 @@ serde = { version = "1.0.144", features = ["derive"] }
 crossbeam-channel = "0.5.6"
 
 [dev-dependencies]
-lsp-types = "0.93.1"
+lsp-types = "=0.93.2"