Class: MilkTea::LSP::Diagnostics

Inherits:
Object
  • Object
show all
Defined in:
lib/milk_tea/lsp/diagnostics.rb

Overview

Collects parse and semantic errors and formats them as LSP Diagnostics

Class Attribute Summary collapse

Class Method Summary collapse

Class Attribute Details

.position_encoding ⇒ Object

Returns the value of attribute position_encoding.



11
12
13
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 11

def position_encoding
  @position_encoding
end

.protocol ⇒ Object

Returns the value of attribute protocol.



11
12
13
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 11

def protocol
  @protocol
end

Class Method Details

.client_char_to_internal(line_text, client_char) ⇒ Object 



615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 615

def self.client_char_to_internal(line_text, client_char)
  return client_char if @position_encoding == 'utf-16'

  case @position_encoding
  when 'utf-8'
    bytes_seen = 0
    utf16_count = 0
    line_text.each_char do |ch|
      break if bytes_seen >= client_char
      bytes_seen += ch.bytesize
      utf16_count += ch.ord > 0xFFFF ? 2 : 1
    end
    utf16_count
  when 'utf-32'
    utf16_count = 0
    line_text.each_char.with_index do |ch, idx|
      break if idx >= client_char
      utf16_count += ch.ord > 0xFFFF ? 2 : 1
    end
    utf16_count
  else
    client_char
  end
end

.collect(uri, content, shared_module_cache: nil, source_overrides: nil, workspace_root_path: nil, dependency_resolution_mode: :auto, platform_override: nil, sema_snapshot: nil, strict_current_root_diagnostics: false, lint_tier: :full) ⇒ Object 



15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 15

def self.collect(uri, content, shared_module_cache: nil, source_overrides: nil, workspace_root_path: nil, dependency_resolution_mode: :auto, platform_override: nil, sema_snapshot: nil, strict_current_root_diagnostics: false, lint_tier: :full)
  total_start = perf_logging? ? monotonic_time : nil
  diagnostics = []
  sema_facts = nil
  unresolved_import_paths = []
  normalized_lint_tier = Linter.normalize_lint_tier(lint_tier)
  parse_ms = 0.0
  imports_ms = 0.0
  sema_ms = 0.0
  strict_root_ms = 0.0
  lint_ms = 0.0
  lint_profile = total_start ? Linter::Profile.new : nil
  path = uri_to_path(uri)
  resolution = DependencyResolution.resolve(path, mode: dependency_resolution_mode)
  effective_platform = effective_platform_for_path(path, platform_override:)

  # Parse
  begin
    parse_start = total_start ? monotonic_time : nil
    ast = if path && File.file?(path)
            parse_result = Parser.parse_collecting_errors(content, path: uri)
            parse_result.errors.each { |error| diagnostics << format_error(error) }
            parse_result.ast
          else
            Parser.parse(content, path: uri)
          end
    parse_ms = elapsed_ms(parse_start) if parse_start
    return { diagnostics: diagnostics, facts: sema_facts, sema_snapshot: sema_snapshot } unless ast

    if resolution.error_message
      diagnostics << format_error(SemanticError.new(resolution.error_message, line: 1, column: 1))
      return { diagnostics: diagnostics, facts: sema_facts, sema_snapshot: sema_snapshot }
    end

    conflict_error = root_platform_conflict_error(path, effective_platform)
    if conflict_error
      diagnostics << format_error(SemanticError.new(conflict_error.message, line: 1, column: 1))
      return { diagnostics: diagnostics, facts: sema_facts, sema_snapshot: sema_snapshot }
    end

    imports_start = total_start ? monotonic_time : nil
    imported_modules = resolve_imported_modules(
      uri,
      ast,
      diagnostics,
      resolution:,
      effective_platform:,
      shared_module_cache: shared_module_cache,
      source_overrides: source_overrides,
      workspace_root_path: workspace_root_path,
      content: content,
    )
    imports_ms = elapsed_ms(imports_start) if imports_start
    unresolved_import_paths = imported_modules.fetch(:unresolved_import_paths)
    inferred_module_name = imported_modules[:module_name]

    # Wrap AST with module name before sema so facts.module_name is non-nil.
    unless ast.respond_to?(:module_name) && ast.module_name
      mn_parts = inferred_module_name ? inferred_module_name.split('.') : []
      ast = AST::SourceFile.new(
        module_name: mn_parts.any? ? AST::QualifiedName.new(parts: mn_parts) : nil,
        module_kind: ast.respond_to?(:module_kind) ? ast.module_kind : nil,
        imports: ast.respond_to?(:imports) ? (ast.imports || []) : [],
        directives: ast.respond_to?(:directives) ? (ast.directives || []) : [],
        declarations: ast.respond_to?(:declarations) ? (ast.declarations || []) : [],
        line: ast.line || 1,
        node_ids: ast.respond_to?(:node_ids) ? ast.node_ids : {},
      )
    end

    # Semantic analysis — collect errors from all functions, not just first.
    begin
      sema_start = total_start ? monotonic_time : nil
      sema_snapshot ||= SemanticAnalyzer.tooling_snapshot(ast, imported_modules: imported_modules.fetch(:modules), path: path, allow_missing_imports: true)
      sema_facts = sema_snapshot.facts
      cross_by_uri = {}
      target_path = path.is_a?(String) ? File.expand_path(path) : nil
      sema_snapshot.diagnostics.reject { |diagnostic| redundant_unknown_import_diagnostic?(diagnostic, unresolved_import_paths) }
        .each do |diagnostic|
          if target_path && diagnostic.path && File.expand_path(diagnostic.path) != target_path
            target_uri = path_to_uri(diagnostic.path)
            (cross_by_uri[target_uri] ||= []) << format_tooling_diagnostic(diagnostic, stage: 'sema')
          else
            diagnostics << format_tooling_diagnostic(diagnostic, stage: 'sema')
          end
        end
      sema_ms = elapsed_ms(sema_start) if sema_start
    rescue StandardError => e
      log_diagnostics_warning("Error collecting diagnostics: #{e.message}")
    end

    begin
      if strict_current_root_diagnostics && normalized_lint_tier == :full
        strict_root_start = total_start ? monotonic_time : nil
        collect_strict_root_diagnostics(
          path,
          ast,
          sema_facts,
          diagnostics,
          resolution:,
          effective_platform:,
          shared_module_cache:,
          source_overrides:,
          workspace_root_path:,
        ).each do |diagnostic|
          diagnostics << diagnostic
        end
        strict_root_ms = elapsed_ms(strict_root_start) if strict_root_start
      end
    rescue StandardError => e
      log_diagnostics_warning("Error collecting strict root diagnostics: #{e.message}")
    end
  rescue MilkTea::LexError => e
    diagnostics << format_error(e)
  rescue StandardError => e
    log_diagnostics_warning("Error collecting diagnostics: #{e.message}")
  end

  # Lint warnings (severity: 2 = Warning)
  begin
    lint_start = total_start ? monotonic_time : nil
    Linter.lint_source(
      content,
      path: uri,
      sema_facts: sema_facts,
      unresolved_import_paths: unresolved_import_paths,
      profile: lint_profile,
      lint_tier: normalized_lint_tier,
    ).each do |w|
      diagnostics << format_warning(w, content: content)
    end
    lint_ms = elapsed_ms(lint_start) if lint_start
  rescue MilkTea::LexError, MilkTea::ParseError
    # Best-effort only while the user is mid-edit; lex/parse failures are
    # already reported by the main diagnostics path.
  rescue StandardError => e
    log_diagnostics_warning("Error collecting lint diagnostics: #{e.message}")
  end

  { diagnostics: diagnostics, cross_file_diagnostics: cross_by_uri, facts: sema_facts, sema_snapshot: sema_snapshot }
ensure
  if total_start
    lint_breakdown = lint_profile&.summary(limit: 12)
    detail = "uri=#{uri} lint_tier=#{normalized_lint_tier} diagnostics=#{diagnostics.length} facts=#{sema_facts ? 'ok' : 'nil'} stages_ms=parse:#{parse_ms},imports:#{imports_ms},sema:#{sema_ms},strict_root:#{strict_root_ms},lint:#{lint_ms}"
    detail += " lint_passes_ms=#{lint_breakdown}" unless lint_breakdown.nil? || lint_breakdown.empty?
    log_perf_breakdown(
      'diagnostics.collect',
      elapsed_ms(total_start),
      detail,
    )
  end
end

.diagnostic_code(error) ⇒ Object 



507
508
509
510
511
512
513
514
515
516
517
518
519
520
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 507

def self.diagnostic_code(error)
  case error
  when MilkTea::LexError
    'lex/error'
  when MilkTea::ParseError
    'parse/error'
  when MilkTea::ModuleLoadError
    'import/load-error'
  when MilkTea::SemanticError
    'sema/error'
  else
    'tooling/error'
  end
end

.diagnostic_message_to_client(message, code) ⇒ Object 



600
601
602
603
604
605
606
607
608
609
610
611
612
613
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 600

def self.diagnostic_message_to_client(message, code)
  return message unless message.is_a?(String)

  code_spans = []
  if code.to_s.start_with?('sema')
    message.scan(/`([^`]+)`/) { |m| code_spans << m[0] }
    message.scan(/'([A-Za-z_]\w*(?:\.[A-Za-z_]\w*)*)'/) { |m| code_spans << m[0] unless code_spans.include?(m[0]) }
  end

  return message if code_spans.empty?

  markdown = message.gsub(/`([^`]+)`/, '`\1`')
  { kind: 'markdown', value: markdown }
end

.diagnostic_stage(error) ⇒ Object 



522
523
524
525
526
527
528
529
530
531
532
533
534
535
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 522

def self.diagnostic_stage(error)
  case error
  when MilkTea::LexError
    'lex'
  when MilkTea::ParseError
    'parse'
  when MilkTea::ModuleLoadError
    'import'
  when MilkTea::SemanticError
    'sema'
  else
    'tooling'
  end
end

.elapsed_ms(start_time) ⇒ Object 



257
258
259
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 257

def self.elapsed_ms(start_time)
  ((monotonic_time - start_time) * 1000).round(1)
end

.extract_column(error) ⇒ Object 



565
566
567
568
569
570
571
572
573
574
575
576
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 565

def self.extract_column(error)
  case error
  when MilkTea::LexError
    error.column || 1
  when MilkTea::ParseError
    error.token&.column || 1
  when MilkTea::SemanticError
    error.column || 1
  else
    1
  end
end

.extract_length(error) ⇒ Object 



578
579
580
581
582
583
584
585
586
587
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 578

def self.extract_length(error)
  case error
  when MilkTea::ParseError
    error.token&.lexeme.to_s.length
  when MilkTea::SemanticError
    error.length || 1
  else
    1
  end
end

.extract_line(error) ⇒ Object 



552
553
554
555
556
557
558
559
560
561
562
563
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 552

def self.extract_line(error)
  case error
  when MilkTea::LexError
    error.line || 1
  when MilkTea::ParseError
    error.token&.line || 1
  when MilkTea::SemanticError
    error.line || 1
  else
    1
  end
end

.extract_quoted_name(message) ⇒ Object 



439
440
441
442
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 439

def self.extract_quoted_name(message)
  match = message.match(/'([^']+)'/)
  match && match[1]
end

.extract_warning_range(warning, content) ⇒ Object

Returns [start_char, end_char] in 0-based LSP coordinates. Falls back to [0, 1] when a precise token span cannot be inferred.



408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 408

def self.extract_warning_range(warning, content)
  if warning.column
    start_char = [warning.column.to_i - 1, 0].max
    len = warning.respond_to?(:length) ? warning.length.to_i : 0
    len = 1 if len <= 0
    return [start_char, start_char + len]
  end

  return [0, 1] unless content && warning.line

  lines = content.split("\n", -1)
  line_text = lines[warning.line - 1]
  return [0, 1] unless line_text

  token_name = warning.respond_to?(:symbol_name) ? warning.symbol_name : nil
  token_name ||= extract_quoted_name(warning.message.to_s)
  return [0, 1] unless token_name && !token_name.empty?

  # Prefer whole-token matches so names like "arg" don't match inside "argv".
  token_match = line_text.match(/\b#{Regexp.escape(token_name)}\b/)
  return [token_match.begin(0), token_match.end(0)] if token_match

  # Fallback for non-word symbols if any rule introduces them.
  index = line_text.index(token_name)
  return [index, index + token_name.length] if index

  [0, 1]
rescue StandardError
  [0, 1]
end

.format_error(error) ⇒ Object 



444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 444

def self.format_error(error)
  line = extract_line(error)
  column = extract_column(error)
  length = extract_length(error)
  start_char = [column.to_i - 1, 0].max
  end_char = start_char + [length.to_i, 1].max
  code = diagnostic_code(error)

  diagnostic = {
    range: {
      start: {
        line: (line.to_i - 1),
        character: start_char
      },
      end: {
        line: (line.to_i - 1),
        character: end_char
      }
    },
    severity: 1,  # Error
    code: code,
    message: diagnostic_message_to_client(error.message.to_s, code),
    source: 'milk-tea',
    data: {
      stage: diagnostic_stage(error)
    }
  }
  if error.respond_to?(:suggestion) && error.suggestion
    diagnostic[:data][:suggestion] = error.suggestion
  end
  diagnostic
end

.format_import_error(error, import, content: nil) ⇒ Object 



477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 477

def self.format_import_error(error, import, content: nil)
  return format_error(SemanticError.new(error.message)) unless import

  column, length = import_path_span(import, content)
  line_index = [import.line.to_i - 1, 0].max
  start_char = [column.to_i - 1, 0].max
  end_char = start_char + [length.to_i, 1].max
  code = diagnostic_code(error)

  {
    range: {
      start: {
        line: line_index,
        character: start_char,
      },
      end: {
        line: line_index,
        character: end_char,
      }
    },
    severity: 1,
    code: code,
    message: diagnostic_message_to_client(error.message.to_s, code),
    source: 'milk-tea',
    data: {
      stage: diagnostic_stage(error)
    }
  }
end

.format_tooling_diagnostic(diagnostic, content: nil, stage:) ⇒ Object 



379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 379

def self.format_tooling_diagnostic(diagnostic, content: nil, stage:)
  line = diagnostic.line.to_i
  line_index = [line - 1, 0].max
  start_char, end_char = extract_warning_range(diagnostic, content)

  lsp_severity = case diagnostic.severity
  when :error   then 1
  when :warning then 2
  when :info    then 3
  when :hint    then 4
  else               2
  end
  {
    range: {
      start: { line: line_index, character: start_char },
      end:   { line: line_index, character: end_char }
    },
    severity: lsp_severity,
    code: diagnostic.code,
    message: diagnostic_message_to_client(diagnostic.message.to_s, diagnostic.code),
    source: 'milk-tea',
    data: {
      stage: stage
    }
  }
end

.format_warning(warning, content: nil) ⇒ Object 



375
376
377
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 375

def self.format_warning(warning, content: nil)
  format_tooling_diagnostic(warning.to_diagnostic, content:, stage: 'lint')
end

.import_path_span(import, content) ⇒ Object 



537
538
539
540
541
542
543
544
545
546
547
548
549
550
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 537

def self.import_path_span(import, content)
  return [import.column || 1, import.length || 1] unless content && import.line

  line_text = content.split("\n", -1)[import.line - 1]
  return [import.column || 1, import.length || 1] unless line_text

  path_text = import.path.to_s
  index = line_text.index(path_text)
  return [index + 1, path_text.length] if index

  [import.column || 1, import.length || 1]
rescue StandardError
  [import.column || 1, import.length || 1]
end

.log_diagnostics_warning(message) ⇒ Object 



589
590
591
592
593
594
595
596
597
598
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 589

def self.log_diagnostics_warning(message)
  return unless @protocol

  @protocol.write_notification('window/logMessage', {
    type: 2,
    message: message,
  })
rescue StandardError
  nil
end

.log_perf_breakdown(name, elapsed_ms_value, detail) ⇒ Object 



261
262
263
264
265
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 261

def self.log_perf_breakdown(name, elapsed_ms_value, detail)
  return unless perf_breakdown_logging?(elapsed_ms_value)

  warn "[LSP perf] breakdown #{name} #{elapsed_ms_value}ms #{detail}"
end

.monotonic_time ⇒ Object 



253
254
255
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 253

def self.monotonic_time
  Process.clock_gettime(Process::CLOCK_MONOTONIC)
end

.path_to_uri(path) ⇒ Object 



227
228
229
230
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 227

def self.path_to_uri(path)
  escaped_path = path.split('/').map { |seg| CGI.escape(seg).gsub('+', '%20') }.join('/')
  "file://#{escaped_path}"
end

.perf_breakdown_logging?(elapsed_ms_value) ⇒ Boolean

Returns:

  • (Boolean) 


249
250
251
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 249

def self.perf_breakdown_logging?(elapsed_ms_value)
  perf_logging? && (perf_verbose? || elapsed_ms_value > Workspace::PERF_LOG_THRESHOLD_MS)
end

.perf_logging? ⇒ Boolean

Returns:

  • (Boolean) 


241
242
243
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 241

def self.perf_logging?
  @perf_logging ||= !ENV.fetch('MILK_TEA_LSP_PERF', nil).to_s.empty?
end

.perf_verbose? ⇒ Boolean

Returns:

  • (Boolean) 


245
246
247
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 245

def self.perf_verbose?
  @perf_verbose ||= ENV.fetch('MILK_TEA_LSP_PERF', nil).to_s == 'verbose'
end

.redundant_unknown_import_diagnostic?(diagnostic, unresolved_import_paths) ⇒ Boolean

Returns:

  • (Boolean) 


218
219
220
221
222
223
224
225
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 218

def self.redundant_unknown_import_diagnostic?(diagnostic, unresolved_import_paths)
  return false unless diagnostic&.message.to_s.start_with?("unknown import ")

  import_path = diagnostic.message.to_s.match(/unknown import (.+)$/)&.captures&.first
  return false unless import_path

  unresolved_import_paths.include?(import_path)
end

.resolve_imported_modules(uri, ast, diagnostics, resolution:, effective_platform:, shared_module_cache: nil, source_overrides: nil, workspace_root_path: nil, content: nil) ⇒ Object 



170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 170

def self.resolve_imported_modules(uri, ast, diagnostics, resolution:, effective_platform:, shared_module_cache: nil, source_overrides: nil, workspace_root_path: nil, content: nil)
  path = uri_to_path(uri)
   return { modules: {}, unresolved_import_paths: [], module_name: nil } unless path && File.file?(path)

  loader = ModuleLoader.new(
    module_roots: module_roots_for_path(path, locked: resolution.locked, workspace_root_path: workspace_root_path),
    package_graph: load_package_graph(path, locked: resolution.locked),
    shared_cache: shared_module_cache,
    source_overrides: source_overrides,
    platform: effective_platform,
  )
  module_name = loader.send(:inferred_module_name_for_path, path) rescue nil

  # Run the full two-pass program check first so that @analysis_cache
  # has fully-checked analyses for all transitive modules. When the
  # import resolution below encounters a cycle member, it finds the
  # cached analysis and resolves correctly. If the program check fails
  # (e.g. missing module), fall through to the standard resolution
  # path so that prelude modules and regular errors are still reported.
  begin
    loader.check_program_collecting(path)
  rescue ModuleLoadError, PackageLockError
    # best-effort — standard path below handles diagnostics
  end

  resolution_result = loader.imported_modules_for_ast_collecting_errors(ast, importer_path: path)
  unresolved_import_paths = []

  resolution_result.errors.each do |entry|
    if entry.error.is_a?(ModuleLoadError)
      diagnostics << format_import_error(entry.error, entry.import, content: content)
      unresolved_import_paths << entry.import.path.to_s if entry.import
    else
      diagnostics << format_error(SemanticError.new(entry.error.message))
    end
  end

  { modules: resolution_result.modules, unresolved_import_paths: unresolved_import_paths.uniq, module_name: module_name }
rescue ModuleLoadError, PackageLockError => e
  if e.is_a?(ModuleLoadError)
    import = ast.imports.find { |candidate| candidate.path.to_s == e.path }
    diagnostics << format_import_error(e, import, content: content)
  else
    diagnostics << format_error(SemanticError.new(e.message))
  end
  { modules: {}, unresolved_import_paths: [], module_name: nil }
end

.uri_to_path(uri) ⇒ Object 



232
233
234
235
236
237
238
239
 
# File 'lib/milk_tea/lsp/diagnostics.rb', line 232

def self.uri_to_path(uri)
  parsed = URI.parse(uri)
  return nil unless parsed.scheme == "file"

  CGI.unescape(parsed.path)
rescue URI::InvalidURIError
  nil
end