Merge #3159

3159: Server-side Semantic Tokens r=matklad a=kjeremy

Takes the output of `syntax_highlighting` and converts it to the proposed semantic tokens API (so far only `textDocument/semanticTokens`. There's a lot of cool stuff we could do with this and the "Inspect Editor Tokens and Scopes" vscode command (pic below) is a cool way to see what has tokens and what doesn't.

Incredibly hacky and could panic due to unwraps, `panic!` etc. To use: run with `code-insiders --enable-proposed-api matklad.rust-analyzer`. If you try to run this without `--enable-proposed-api` it will crash.

![image](https://user-images.githubusercontent.com/4325700/74595603-7c66cf00-5011-11ea-9593-312663f04fc1.png)


@matklad I'm mostly looking for design feedback.

Co-authored-by: kjeremy <[email protected]>
Co-authored-by: Jeremy Kolb <[email protected]>

9 files changed, 247 insertions, 40 deletions

diff --git a/crates/ra_ide/src/lib.rs b/crates/ra_ide/src/lib.rs
index f86f98be7..82e10bc7e 100644
--- a/crates/ra_ide/src/lib.rs
+++ b/crates/ra_ide/src/lib.rs
@@ -75,7 +75,7 @@ pub use crate::{
     runnables::{Runnable, RunnableKind, TestId},
     source_change::{FileSystemEdit, SourceChange, SourceFileEdit},
     ssr::SsrError,
-    syntax_highlighting::HighlightedRange,
+    syntax_highlighting::{tags, HighlightedRange},
 };
 
 pub use hir::Documentation;
diff --git a/crates/ra_ide/src/syntax_highlighting.rs b/crates/ra_ide/src/syntax_highlighting.rs
index d873f153e..812229b4e 100644
--- a/crates/ra_ide/src/syntax_highlighting.rs
+++ b/crates/ra_ide/src/syntax_highlighting.rs
@@ -17,32 +17,32 @@ use crate::{
 };
 
 pub mod tags {
-    pub(crate) const FIELD: &str = "field";
-    pub(crate) const FUNCTION: &str = "function";
-    pub(crate) const MODULE: &str = "module";
-    pub(crate) const CONSTANT: &str = "constant";
-    pub(crate) const MACRO: &str = "macro";
-
-    pub(crate) const VARIABLE: &str = "variable";
-    pub(crate) const VARIABLE_MUT: &str = "variable.mut";
-
-    pub(crate) const TYPE: &str = "type";
-    pub(crate) const TYPE_BUILTIN: &str = "type.builtin";
-    pub(crate) const TYPE_SELF: &str = "type.self";
-    pub(crate) const TYPE_PARAM: &str = "type.param";
-    pub(crate) const TYPE_LIFETIME: &str = "type.lifetime";
-
-    pub(crate) const LITERAL_BYTE: &str = "literal.byte";
-    pub(crate) const LITERAL_NUMERIC: &str = "literal.numeric";
-    pub(crate) const LITERAL_CHAR: &str = "literal.char";
-
-    pub(crate) const LITERAL_COMMENT: &str = "comment";
-    pub(crate) const LITERAL_STRING: &str = "string";
-    pub(crate) const LITERAL_ATTRIBUTE: &str = "attribute";
-
-    pub(crate) const KEYWORD: &str = "keyword";
-    pub(crate) const KEYWORD_UNSAFE: &str = "keyword.unsafe";
-    pub(crate) const KEYWORD_CONTROL: &str = "keyword.control";
+    pub const FIELD: &str = "field";
+    pub const FUNCTION: &str = "function";
+    pub const MODULE: &str = "module";
+    pub const CONSTANT: &str = "constant";
+    pub const MACRO: &str = "macro";
+
+    pub const VARIABLE: &str = "variable";
+    pub const VARIABLE_MUT: &str = "variable.mut";
+
+    pub const TYPE: &str = "type";
+    pub const TYPE_BUILTIN: &str = "type.builtin";
+    pub const TYPE_SELF: &str = "type.self";
+    pub const TYPE_PARAM: &str = "type.param";
+    pub const TYPE_LIFETIME: &str = "type.lifetime";
+
+    pub const LITERAL_BYTE: &str = "literal.byte";
+    pub const LITERAL_NUMERIC: &str = "literal.numeric";
+    pub const LITERAL_CHAR: &str = "literal.char";
+
+    pub const LITERAL_COMMENT: &str = "comment";
+    pub const LITERAL_STRING: &str = "string";
+    pub const LITERAL_ATTRIBUTE: &str = "attribute";
+
+    pub const KEYWORD: &str = "keyword";
+    pub const KEYWORD_UNSAFE: &str = "keyword.unsafe";
+    pub const KEYWORD_CONTROL: &str = "keyword.control";
 }
 
 #[derive(Debug)]
diff --git a/crates/rust-analyzer/src/caps.rs b/crates/rust-analyzer/src/caps.rs
index c9fd645f1..638987ee8 100644
--- a/crates/rust-analyzer/src/caps.rs
+++ b/crates/rust-analyzer/src/caps.rs
@@ -1,12 +1,15 @@
 //! Advertizes the capabilities of the LSP Server.
 
+use crate::semantic_tokens;
+
 use lsp_types::{
     CallHierarchyServerCapability, CodeActionProviderCapability, CodeLensOptions,
     CompletionOptions, DocumentOnTypeFormattingOptions, FoldingRangeProviderCapability,
     ImplementationProviderCapability, RenameOptions, RenameProviderCapability, SaveOptions,
-    SelectionRangeProviderCapability, ServerCapabilities, SignatureHelpOptions,
-    TextDocumentSyncCapability, TextDocumentSyncKind, TextDocumentSyncOptions,
-    TypeDefinitionProviderCapability, WorkDoneProgressOptions,
+    SelectionRangeProviderCapability, SemanticTokensDocumentProvider, SemanticTokensLegend,
+    SemanticTokensOptions, SemanticTokensServerCapabilities, ServerCapabilities,
+    SignatureHelpOptions, TextDocumentSyncCapability, TextDocumentSyncKind,
+    TextDocumentSyncOptions, TypeDefinitionProviderCapability, WorkDoneProgressOptions,
 };
 
 pub fn server_capabilities() -> ServerCapabilities {
@@ -57,7 +60,20 @@ pub fn server_capabilities() -> ServerCapabilities {
         execute_command_provider: None,
         workspace: None,
         call_hierarchy_provider: Some(CallHierarchyServerCapability::Simple(true)),
-        semantic_tokens_provider: None,
+        semantic_tokens_provider: Some(SemanticTokensServerCapabilities::SemanticTokensOptions(
+            SemanticTokensOptions {
+                legend: SemanticTokensLegend {
+                    token_types: semantic_tokens::supported_token_types().iter().cloned().collect(),
+                    token_modifiers: semantic_tokens::supported_token_modifiers()
+                        .iter()
+                        .cloned()
+                        .collect(),
+                },
+
+                document_provider: Some(SemanticTokensDocumentProvider::Bool(true)),
+                ..SemanticTokensOptions::default()
+            },
+        )),
         experimental: Default::default(),
     }
 }
diff --git a/crates/rust-analyzer/src/conv.rs b/crates/rust-analyzer/src/conv.rs
index 90ef74056..5fcb46b61 100644
--- a/crates/rust-analyzer/src/conv.rs
+++ b/crates/rust-analyzer/src/conv.rs
@@ -4,11 +4,12 @@
 use lsp_types::{
     self, CreateFile, DiagnosticSeverity, DocumentChangeOperation, DocumentChanges, Documentation,
     Location, LocationLink, MarkupContent, MarkupKind, Position, Range, RenameFile, ResourceOp,
-    SymbolKind, TextDocumentEdit, TextDocumentIdentifier, TextDocumentItem,
-    TextDocumentPositionParams, Url, VersionedTextDocumentIdentifier, WorkspaceEdit,
+    SemanticTokenModifier, SemanticTokenType, SymbolKind, TextDocumentEdit, TextDocumentIdentifier,
+    TextDocumentItem, TextDocumentPositionParams, Url, VersionedTextDocumentIdentifier,
+    WorkspaceEdit,
 };
 use ra_ide::{
-    translate_offset_with_edit, CompletionItem, CompletionItemKind, FileId, FilePosition,
+    tags, translate_offset_with_edit, CompletionItem, CompletionItemKind, FileId, FilePosition,
     FileRange, FileSystemEdit, Fold, FoldKind, InsertTextFormat, LineCol, LineIndex,
     NavigationTarget, RangeInfo, ReferenceAccess, Severity, SourceChange, SourceFileEdit,
 };
@@ -16,7 +17,7 @@ use ra_syntax::{SyntaxKind, TextRange, TextUnit};
 use ra_text_edit::{AtomTextEdit, TextEdit};
 use ra_vfs::LineEndings;
 
-use crate::{req, world::WorldSnapshot, Result};
+use crate::{req, semantic_tokens, world::WorldSnapshot, Result};
 
 pub trait Conv {
     type Output;
@@ -302,6 +303,76 @@ impl ConvWith<&FoldConvCtx<'_>> for Fold {
     }
 }
 
+impl Conv for &'static str {
+    type Output = (SemanticTokenType, Vec<SemanticTokenModifier>);
+
+    fn conv(self) -> (SemanticTokenType, Vec<SemanticTokenModifier>) {
+        let token_type: SemanticTokenType = match self {
+            tags::FIELD => SemanticTokenType::MEMBER,
+            tags::FUNCTION => SemanticTokenType::FUNCTION,
+            tags::MODULE => SemanticTokenType::NAMESPACE,
+            tags::CONSTANT => {
+                return (
+                    SemanticTokenType::VARIABLE,
+                    vec![SemanticTokenModifier::STATIC, SemanticTokenModifier::READONLY],
+                )
+            }
+            tags::MACRO => SemanticTokenType::MACRO,
+
+            tags::VARIABLE => {
+                return (SemanticTokenType::VARIABLE, vec![SemanticTokenModifier::READONLY])
+            }
+            tags::VARIABLE_MUT => SemanticTokenType::VARIABLE,
+
+            tags::TYPE => SemanticTokenType::TYPE,
+            tags::TYPE_BUILTIN => SemanticTokenType::TYPE,
+            tags::TYPE_SELF => {
+                return (SemanticTokenType::TYPE, vec![SemanticTokenModifier::REFERENCE])
+            }
+            tags::TYPE_PARAM => SemanticTokenType::TYPE_PARAMETER,
+            tags::TYPE_LIFETIME => {
+                return (SemanticTokenType::LABEL, vec![SemanticTokenModifier::REFERENCE])
+            }
+
+            tags::LITERAL_BYTE => SemanticTokenType::NUMBER,
+            tags::LITERAL_NUMERIC => SemanticTokenType::NUMBER,
+            tags::LITERAL_CHAR => SemanticTokenType::NUMBER,
+
+            tags::LITERAL_COMMENT => {
+                return (SemanticTokenType::COMMENT, vec![SemanticTokenModifier::DOCUMENTATION])
+            }
+
+            tags::LITERAL_STRING => SemanticTokenType::STRING,
+            tags::LITERAL_ATTRIBUTE => SemanticTokenType::KEYWORD,
+
+            tags::KEYWORD => SemanticTokenType::KEYWORD,
+            tags::KEYWORD_UNSAFE => SemanticTokenType::KEYWORD,
+            tags::KEYWORD_CONTROL => SemanticTokenType::KEYWORD,
+            unknown => panic!("Unknown semantic token: {}", unknown),
+        };
+
+        (token_type, vec![])
+    }
+}
+
+impl Conv for (SemanticTokenType, Vec<SemanticTokenModifier>) {
+    type Output = (u32, u32);
+
+    fn conv(self) -> Self::Output {
+        let token_index =
+            semantic_tokens::supported_token_types().iter().position(|it| *it == self.0).unwrap();
+        let mut token_modifier_bitset = 0;
+        for modifier in self.1.iter() {
+            token_modifier_bitset |= semantic_tokens::supported_token_modifiers()
+                .iter()
+                .position(|it| it == modifier)
+                .unwrap();
+        }
+
+        (token_index as u32, token_modifier_bitset as u32)
+    }
+}
+
 impl<T: ConvWith<CTX>, CTX> ConvWith<CTX> for Option<T> {
     type Output = Option<T::Output>;
 
diff --git a/crates/rust-analyzer/src/lib.rs b/crates/rust-analyzer/src/lib.rs
index 0dae30e46..a0f968823 100644
--- a/crates/rust-analyzer/src/lib.rs
+++ b/crates/rust-analyzer/src/lib.rs
@@ -36,6 +36,7 @@ pub mod req;
 mod config;
 mod world;
 mod diagnostics;
+mod semantic_tokens;
 
 use serde::de::DeserializeOwned;
 
diff --git a/crates/rust-analyzer/src/main_loop.rs b/crates/rust-analyzer/src/main_loop.rs
index 98306986b..6e9e604a6 100644
--- a/crates/rust-analyzer/src/main_loop.rs
+++ b/crates/rust-analyzer/src/main_loop.rs
@@ -528,6 +528,7 @@ fn on_request(
         .on::<req::CallHierarchyIncomingCalls>(handlers::handle_call_hierarchy_incoming)?
         .on::<req::CallHierarchyOutgoingCalls>(handlers::handle_call_hierarchy_outgoing)?
         .on::<req::Ssr>(handlers::handle_ssr)?
+        .on::<req::SemanticTokensRequest>(handlers::handle_semantic_tokens)?
         .finish();
     Ok(())
 }
diff --git a/crates/rust-analyzer/src/main_loop/handlers.rs b/crates/rust-analyzer/src/main_loop/handlers.rs
index bb7bab372..e13e7c95a 100644
--- a/crates/rust-analyzer/src/main_loop/handlers.rs
+++ b/crates/rust-analyzer/src/main_loop/handlers.rs
@@ -16,8 +16,9 @@ use lsp_types::{
     CodeAction, CodeActionOrCommand, CodeActionResponse, CodeLens, Command, CompletionItem,
     Diagnostic, DocumentFormattingParams, DocumentHighlight, DocumentSymbol, FoldingRange,
     FoldingRangeParams, Hover, HoverContents, Location, MarkupContent, MarkupKind, Position,
-    PrepareRenameResponse, Range, RenameParams, SymbolInformation, TextDocumentIdentifier,
-    TextEdit, WorkspaceEdit,
+    PrepareRenameResponse, Range, RenameParams, SemanticTokenModifier, SemanticTokenType,
+    SemanticTokens, SemanticTokensParams, SemanticTokensResult, SymbolInformation,
+    TextDocumentIdentifier, TextEdit, WorkspaceEdit,
 };
 use ra_ide::{
     AssistId, FileId, FilePosition, FileRange, Query, RangeInfo, Runnable, RunnableKind,
@@ -38,6 +39,7 @@ use crate::{
     diagnostics::DiagnosticTask,
     from_json,
     req::{self, Decoration, InlayHint, InlayHintsParams, InlayKind},
+    semantic_tokens::SemanticTokensBuilder,
     world::WorldSnapshot,
     LspError, Result,
 };
@@ -1068,3 +1070,25 @@ pub fn handle_call_hierarchy_outgoing(
 
     Ok(Some(res))
 }
+
+pub fn handle_semantic_tokens(
+    world: WorldSnapshot,
+    params: SemanticTokensParams,
+) -> Result<Option<SemanticTokensResult>> {
+    let _p = profile("handle_semantic_tokens");
+
+    let file_id = params.text_document.try_conv_with(&world)?;
+    let line_index = world.analysis().file_line_index(file_id)?;
+
+    let mut builder = SemanticTokensBuilder::default();
+
+    for h in world.analysis().highlight(file_id)?.into_iter() {
+        let type_and_modifiers: (SemanticTokenType, Vec<SemanticTokenModifier>) = h.tag.conv();
+        let (token_type, token_modifiers) = type_and_modifiers.conv();
+        builder.push(h.range.conv_with(&line_index), token_type, token_modifiers);
+    }
+
+    let tokens = SemanticTokens { data: builder.build(), ..Default::default() };
+
+    Ok(Some(tokens.into()))
+}
diff --git a/crates/rust-analyzer/src/req.rs b/crates/rust-analyzer/src/req.rs
index 7ff7f60b3..3734899bc 100644
--- a/crates/rust-analyzer/src/req.rs
+++ b/crates/rust-analyzer/src/req.rs
@@ -12,9 +12,9 @@ pub use lsp_types::{
     DocumentSymbolResponse, FileSystemWatcher, Hover, InitializeResult, MessageType,
     PartialResultParams, ProgressParams, ProgressParamsValue, ProgressToken,
     PublishDiagnosticsParams, ReferenceParams, Registration, RegistrationParams, SelectionRange,
-    SelectionRangeParams, ServerCapabilities, ShowMessageParams, SignatureHelp, SymbolKind,
-    TextDocumentEdit, TextDocumentPositionParams, TextEdit, WorkDoneProgressParams, WorkspaceEdit,
-    WorkspaceSymbolParams,
+    SelectionRangeParams, SemanticTokensParams, SemanticTokensResult, ServerCapabilities,
+    ShowMessageParams, SignatureHelp, SymbolKind, TextDocumentEdit, TextDocumentPositionParams,
+    TextEdit, WorkDoneProgressParams, WorkspaceEdit, WorkspaceSymbolParams,
 };
 
 pub enum AnalyzerStatus {}
diff --git a/crates/rust-analyzer/src/semantic_tokens.rs b/crates/rust-analyzer/src/semantic_tokens.rs
new file mode 100644
index 000000000..ad000a3ce
--- /dev/null
+++ b/crates/rust-analyzer/src/semantic_tokens.rs
@@ -0,0 +1,94 @@
+//! Semantic Tokens helpers
+
+use lsp_types::{Range, SemanticToken, SemanticTokenModifier, SemanticTokenType};
+
+const SUPPORTED_TYPES: &[SemanticTokenType] = &[
+    SemanticTokenType::COMMENT,
+    SemanticTokenType::KEYWORD,
+    SemanticTokenType::STRING,
+    SemanticTokenType::NUMBER,
+    SemanticTokenType::REGEXP,
+    SemanticTokenType::OPERATOR,
+    SemanticTokenType::NAMESPACE,
+    SemanticTokenType::TYPE,
+    SemanticTokenType::STRUCT,
+    SemanticTokenType::CLASS,
+    SemanticTokenType::INTERFACE,
+    SemanticTokenType::ENUM,
+    SemanticTokenType::TYPE_PARAMETER,
+    SemanticTokenType::FUNCTION,
+    SemanticTokenType::MEMBER,
+    SemanticTokenType::PROPERTY,
+    SemanticTokenType::MACRO,
+    SemanticTokenType::VARIABLE,
+    SemanticTokenType::PARAMETER,
+    SemanticTokenType::LABEL,
+];
+
+const SUPPORTED_MODIFIERS: &[SemanticTokenModifier] = &[
+    SemanticTokenModifier::DOCUMENTATION,
+    SemanticTokenModifier::DECLARATION,
+    SemanticTokenModifier::DEFINITION,
+    SemanticTokenModifier::REFERENCE,
+    SemanticTokenModifier::STATIC,
+    SemanticTokenModifier::ABSTRACT,
+    SemanticTokenModifier::DEPRECATED,
+    SemanticTokenModifier::ASYNC,
+    SemanticTokenModifier::VOLATILE,
+    SemanticTokenModifier::READONLY,
+];
+
+/// Token types that the server supports
+pub(crate) fn supported_token_types() -> &'static [SemanticTokenType] {
+    SUPPORTED_TYPES
+}
+
+/// Token modifiers that the server supports
+pub(crate) fn supported_token_modifiers() -> &'static [SemanticTokenModifier] {
+    SUPPORTED_MODIFIERS
+}
+
+/// Tokens are encoded relative to each other.
+/// 
+/// This is a direct port of https://github.com/microsoft/vscode-languageserver-node/blob/f425af9de46a0187adb78ec8a46b9b2ce80c5412/server/src/sematicTokens.proposed.ts#L45
+#[derive(Default)]
+pub(crate) struct SemanticTokensBuilder {
+    prev_line: u32,
+    prev_char: u32,
+    data: Vec<SemanticToken>,
+}
+
+impl SemanticTokensBuilder {
+    /// Push a new token onto the builder
+    pub fn push(&mut self, range: Range, token_index: u32, modifier_bitset: u32) {
+        let mut push_line = range.start.line as u32;
+        let mut push_char = range.start.character as u32;
+
+        if !self.data.is_empty() {
+            push_line -= self.prev_line;
+            if push_line == 0 {
+                push_char -= self.prev_char;
+            }
+        }
+
+        // A token cannot be multiline
+        let token_len = range.end.character - range.start.character;
+
+        let token = SemanticToken {
+            delta_line: push_line,
+            delta_start: push_char,
+            length: token_len as u32,
+            token_type: token_index,
+            token_modifiers_bitset: modifier_bitset,
+        };
+
+        self.data.push(token);
+
+        self.prev_line = range.start.line as u32;
+        self.prev_char = range.start.character as u32;
+    }
+
+    pub fn build(self) -> Vec<SemanticToken> {
+        self.data
+    }
+}