clang-tools-extra/clang-doc/Representation.h

   1 ///===-- Representation.h - ClangDoc Representation -------------*- C++ -*-===//
   2 //
   3 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
   4 // See https://llvm.org/LICENSE.txt for license information.
   5 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
   6 //
   7 //===----------------------------------------------------------------------===//
   8 //
   9 // This file defines the internal representations of different declaration
  10 // types for the clang-doc tool.
  11 //
  12 //===----------------------------------------------------------------------===//
  13
  14 #ifndef LLVM_CLANG_TOOLS_EXTRA_CLANG_DOC_REPRESENTATION_H
  15 #define LLVM_CLANG_TOOLS_EXTRA_CLANG_DOC_REPRESENTATION_H
  16
  17 #include "clang/AST/Type.h"
  18 #include "clang/Basic/Specifiers.h"
  19 #include "clang/Tooling/StandaloneExecution.h"
  20 #include "llvm/ADT/APSInt.h"
  21 #include "llvm/ADT/SmallVector.h"
  22 #include "llvm/ADT/StringExtras.h"
  23 #include <array>
  24 #include <optional>
  25 #include <string>
  26
  27 namespace clang {
  28 namespace doc {
  29
  30 // SHA1'd hash of a USR.
  31 using SymbolID = std::array<uint8_t, 20>;
  32
  33 struct BaseRecordInfo;
  34 struct EnumInfo;
  35 struct FunctionInfo;
  36 struct Info;
  37 struct TypedefInfo;
  38
  39 enum class InfoType {
  40   IT_default,
  41   IT_namespace,
  42   IT_record,
  43   IT_function,
  44   IT_enum,
  45   IT_typedef
  46 };
  47
  48 // A representation of a parsed comment.
  49 struct CommentInfo {
  50   CommentInfo() = default;
  51   CommentInfo(CommentInfo &Other) = delete;
  52   CommentInfo(CommentInfo &&Other) = default;
  53   CommentInfo &operator=(CommentInfo &&Other) = default;
  54
  55   bool operator==(const CommentInfo &Other) const;
  56
  57   // This operator is used to sort a vector of CommentInfos.
  58   // No specific order (attributes more important than others) is required. Any
  59   // sort is enough, the order is only needed to call std::unique after sorting
  60   // the vector.
  61   bool operator<(const CommentInfo &Other) const;
  62
  63   SmallString<16>
  64       Kind; // Kind of comment (FullComment, ParagraphComment, TextComment,
  65             // InlineCommandComment, HTMLStartTagComment, HTMLEndTagComment,
  66             // BlockCommandComment, ParamCommandComment,
  67             // TParamCommandComment, VerbatimBlockComment,
  68             // VerbatimBlockLineComment, VerbatimLineComment).
  69   SmallString<64> Text;      // Text of the comment.
  70   SmallString<16> Name;      // Name of the comment (for Verbatim and HTML).
  71   SmallString<8> Direction;  // Parameter direction (for (T)ParamCommand).
  72   SmallString<16> ParamName; // Parameter name (for (T)ParamCommand).
  73   SmallString<16> CloseName; // Closing tag name (for VerbatimBlock).
  74   bool SelfClosing = false;  // Indicates if tag is self-closing (for HTML).
  75   bool Explicit = false; // Indicates if the direction of a param is explicit
  76                          // (for (T)ParamCommand).
  77   llvm::SmallVector<SmallString<16>, 4>
  78       AttrKeys; // List of attribute keys (for HTML).
  79   llvm::SmallVector<SmallString<16>, 4>
  80       AttrValues; // List of attribute values for each key (for HTML).
  81   llvm::SmallVector<SmallString<16>, 4>
  82       Args; // List of arguments to commands (for InlineCommand).
  83   std::vector<std::unique_ptr<CommentInfo>>
  84       Children; // List of child comments for this CommentInfo.
  85 };
  86
  87 struct Reference {
  88   // This variant (that takes no qualified name parameter) uses the Name as the
  89   // QualName (very useful in unit tests to reduce verbosity). This can't use an
  90   // empty string to indicate the default because we need to accept the empty
  91   // string as a valid input for the global namespace (it will have
  92   // "GlobalNamespace" as the name, but an empty QualName).
  93   Reference(SymbolID USR = SymbolID(), StringRef Name = StringRef(),
  94             InfoType IT = InfoType::IT_default)
  95       : USR(USR), Name(Name), QualName(Name), RefType(IT) {}
  96   Reference(SymbolID USR, StringRef Name, InfoType IT, StringRef QualName,
  97             StringRef Path = StringRef())
  98       : USR(USR), Name(Name), QualName(QualName), RefType(IT), Path(Path) {}
  99
 100   bool operator==(const Reference &Other) const {
 101     return std::tie(USR, Name, QualName, RefType) ==
 102            std::tie(Other.USR, Other.Name, QualName, Other.RefType);
 103   }
 104
 105   bool mergeable(const Reference &Other);
 106   void merge(Reference &&I);
 107
 108   /// Returns the path for this Reference relative to CurrentPath.
 109   llvm::SmallString<64> getRelativeFilePath(const StringRef &CurrentPath) const;
 110
 111   /// Returns the basename that should be used for this Reference.
 112   llvm::SmallString<16> getFileBaseName() const;
 113
 114   SymbolID USR = SymbolID(); // Unique identifier for referenced decl
 115
 116   // Name of type (possibly unresolved). Not including namespaces or template
 117   // parameters (so for a std::vector<int> this would be "vector"). See also
 118   // QualName.
 119   SmallString<16> Name;
 120
 121   // Full qualified name of this type, including namespaces and template
 122   // parameter (for example this could be "std::vector<int>"). Contrast to
 123   // Name.
 124   SmallString<16> QualName;
 125
 126   InfoType RefType = InfoType::IT_default; // Indicates the type of this
 127                                            // Reference (namespace, record,
 128                                            // function, enum, default).
 129   // Path of directory where the clang-doc generated file will be saved
 130   // (possibly unresolved)
 131   llvm::SmallString<128> Path;
 132 };
 133
 134 // Holds the children of a record or namespace.
 135 struct ScopeChildren {
 136   // Namespaces and Records are references because they will be properly
 137   // documented in their own info, while the entirety of Functions and Enums are
 138   // included here because they should not have separate documentation from
 139   // their scope.
 140   //
 141   // Namespaces are not syntactically valid as children of records, but making
 142   // this general for all possible container types reduces code complexity.
 143   std::vector<Reference> Namespaces;
 144   std::vector<Reference> Records;
 145   std::vector<FunctionInfo> Functions;
 146   std::vector<EnumInfo> Enums;
 147   std::vector<TypedefInfo> Typedefs;
 148 };
 149
 150 // A base struct for TypeInfos
 151 struct TypeInfo {
 152   TypeInfo() = default;
 153   TypeInfo(const Reference &R) : Type(R) {}
 154
 155   // Convenience constructor for when there is no symbol ID or info type
 156   // (normally used for built-in types in tests).
 157   TypeInfo(StringRef Name, StringRef Path = StringRef())
 158       : Type(SymbolID(), Name, InfoType::IT_default, Name, Path) {}
 159
 160   bool operator==(const TypeInfo &Other) const { return Type == Other.Type; }
 161
 162   Reference Type; // Referenced type in this info.
 163 };
 164
 165 // Represents one template parameter.
 166 //
 167 // This is a very simple serialization of the text of the source code of the
 168 // template parameter. It is saved in a struct so there is a place to add the
 169 // name and default values in the future if needed.
 170 struct TemplateParamInfo {
 171   TemplateParamInfo() = default;
 172   explicit TemplateParamInfo(StringRef Contents) : Contents(Contents) {}
 173
 174   // The literal contents of the code for that specifies this template parameter
 175   // for this declaration. Typical values will be "class T" and
 176   // "typename T = int".
 177   SmallString<16> Contents;
 178 };
 179
 180 struct TemplateSpecializationInfo {
 181   // Indicates the declaration that this specializes.
 182   SymbolID SpecializationOf;
 183
 184   // Template parameters applying to the specialized record/function.
 185   std::vector<TemplateParamInfo> Params;
 186 };
 187
 188 // Records the template information for a struct or function that is a template
 189 // or an explicit template specialization.
 190 struct TemplateInfo {
 191   // May be empty for non-partial specializations.
 192   std::vector<TemplateParamInfo> Params;
 193
 194   // Set when this is a specialization of another record/function.
 195   std::optional<TemplateSpecializationInfo> Specialization;
 196 };
 197
 198 // Info for field types.
 199 struct FieldTypeInfo : public TypeInfo {
 200   FieldTypeInfo() = default;
 201   FieldTypeInfo(const TypeInfo &TI, StringRef Name = StringRef(),
 202                 StringRef DefaultValue = StringRef())
 203       : TypeInfo(TI), Name(Name), DefaultValue(DefaultValue) {}
 204
 205   bool operator==(const FieldTypeInfo &Other) const {
 206     return std::tie(Type, Name, DefaultValue) ==
 207            std::tie(Other.Type, Other.Name, Other.DefaultValue);
 208   }
 209
 210   SmallString<16> Name; // Name associated with this info.
 211
 212   // When used for function parameters, contains the string representing the
 213   // expression of the default value, if any.
 214   SmallString<16> DefaultValue;
 215 };
 216
 217 // Info for member types.
 218 struct MemberTypeInfo : public FieldTypeInfo {
 219   MemberTypeInfo() = default;
 220   MemberTypeInfo(const TypeInfo &TI, StringRef Name, AccessSpecifier Access)
 221       : FieldTypeInfo(TI, Name), Access(Access) {}
 222
 223   bool operator==(const MemberTypeInfo &Other) const {
 224     return std::tie(Type, Name, Access, Description) ==
 225            std::tie(Other.Type, Other.Name, Other.Access, Other.Description);
 226   }
 227
 228   // Access level associated with this info (public, protected, private, none).
 229   // AS_public is set as default because the bitcode writer requires the enum
 230   // with value 0 to be used as the default.
 231   // (AS_public = 0, AS_protected = 1, AS_private = 2, AS_none = 3)
 232   AccessSpecifier Access = AccessSpecifier::AS_public;
 233
 234   std::vector<CommentInfo> Description; // Comment description of this field.
 235 };
 236
 237 struct Location {
 238   Location(int LineNumber = 0, StringRef Filename = StringRef(),
 239            bool IsFileInRootDir = false)
 240       : LineNumber(LineNumber), Filename(Filename),
 241         IsFileInRootDir(IsFileInRootDir) {}
 242
 243   bool operator==(const Location &Other) const {
 244     return std::tie(LineNumber, Filename) ==
 245            std::tie(Other.LineNumber, Other.Filename);
 246   }
 247
 248   // This operator is used to sort a vector of Locations.
 249   // No specific order (attributes more important than others) is required. Any
 250   // sort is enough, the order is only needed to call std::unique after sorting
 251   // the vector.
 252   bool operator<(const Location &Other) const {
 253     return std::tie(LineNumber, Filename) <
 254            std::tie(Other.LineNumber, Other.Filename);
 255   }
 256
 257   int LineNumber = 0;           // Line number of this Location.
 258   SmallString<32> Filename;     // File for this Location.
 259   bool IsFileInRootDir = false; // Indicates if file is inside root directory
 260 };
 261
 262 /// A base struct for Infos.
 263 struct Info {
 264   Info(InfoType IT = InfoType::IT_default, SymbolID USR = SymbolID(),
 265        StringRef Name = StringRef(), StringRef Path = StringRef())
 266       : USR(USR), IT(IT), Name(Name), Path(Path) {}
 267
 268   Info(const Info &Other) = delete;
 269   Info(Info &&Other) = default;
 270
 271   virtual ~Info() = default;
 272
 273   SymbolID USR =
 274       SymbolID(); // Unique identifier for the decl described by this Info.
 275   const InfoType IT = InfoType::IT_default; // InfoType of this particular Info.
 276   SmallString<16> Name;                     // Unqualified name of the decl.
 277   llvm::SmallVector<Reference, 4>
 278       Namespace; // List of parent namespaces for this decl.
 279   std::vector<CommentInfo> Description; // Comment description of this decl.
 280   llvm::SmallString<128> Path;          // Path of directory where the clang-doc
 281                                         // generated file will be saved
 282
 283   void mergeBase(Info &&I);
 284   bool mergeable(const Info &Other);
 285
 286   llvm::SmallString<16> extractName() const;
 287
 288   /// Returns the file path for this Info relative to CurrentPath.
 289   llvm::SmallString<64> getRelativeFilePath(const StringRef &CurrentPath) const;
 290
 291   /// Returns the basename that should be used for this Info.
 292   llvm::SmallString<16> getFileBaseName() const;
 293 };
 294
 295 // Info for namespaces.
 296 struct NamespaceInfo : public Info {
 297   NamespaceInfo(SymbolID USR = SymbolID(), StringRef Name = StringRef(),
 298                 StringRef Path = StringRef());
 299
 300   void merge(NamespaceInfo &&I);
 301
 302   ScopeChildren Children;
 303 };
 304
 305 // Info for symbols.
 306 struct SymbolInfo : public Info {
 307   SymbolInfo(InfoType IT, SymbolID USR = SymbolID(),
 308              StringRef Name = StringRef(), StringRef Path = StringRef())
 309       : Info(IT, USR, Name, Path) {}
 310
 311   void merge(SymbolInfo &&I);
 312
 313   std::optional<Location> DefLoc;     // Location where this decl is defined.
 314   llvm::SmallVector<Location, 2> Loc; // Locations where this decl is declared.
 315 };
 316
 317 // TODO: Expand to allow for documenting templating and default args.
 318 // Info for functions.
 319 struct FunctionInfo : public SymbolInfo {
 320   FunctionInfo(SymbolID USR = SymbolID())
 321       : SymbolInfo(InfoType::IT_function, USR) {}
 322
 323   void merge(FunctionInfo &&I);
 324
 325   bool IsMethod = false; // Indicates whether this function is a class method.
 326   Reference Parent;      // Reference to the parent class decl for this method.
 327   TypeInfo ReturnType;   // Info about the return type of this function.
 328   llvm::SmallVector<FieldTypeInfo, 4> Params; // List of parameters.
 329   // Access level for this method (public, private, protected, none).
 330   // AS_public is set as default because the bitcode writer requires the enum
 331   // with value 0 to be used as the default.
 332   // (AS_public = 0, AS_protected = 1, AS_private = 2, AS_none = 3)
 333   AccessSpecifier Access = AccessSpecifier::AS_public;
 334
 335   // Full qualified name of this function, including namespaces and template
 336   // specializations.
 337   SmallString<16> FullName;
 338
 339   // When present, this function is a template or specialization.
 340   std::optional<TemplateInfo> Template;
 341 };
 342
 343 // TODO: Expand to allow for documenting templating, inheritance access,
 344 // friend classes
 345 // Info for types.
 346 struct RecordInfo : public SymbolInfo {
 347   RecordInfo(SymbolID USR = SymbolID(), StringRef Name = StringRef(),
 348              StringRef Path = StringRef());
 349
 350   void merge(RecordInfo &&I);
 351
 352   // Type of this record (struct, class, union, interface).
 353   TagTypeKind TagType = TagTypeKind::Struct;
 354
 355   // Full qualified name of this record, including namespaces and template
 356   // specializations.
 357   SmallString<16> FullName;
 358
 359   // When present, this record is a template or specialization.
 360   std::optional<TemplateInfo> Template;
 361
 362   // Indicates if the record was declared using a typedef. Things like anonymous
 363   // structs in a typedef:
 364   //   typedef struct { ... } foo_t;
 365   // are converted into records with the typedef as the Name + this flag set.
 366   bool IsTypeDef = false;
 367
 368   llvm::SmallVector<MemberTypeInfo, 4>
 369       Members;                             // List of info about record members.
 370   llvm::SmallVector<Reference, 4> Parents; // List of base/parent records
 371                                            // (does not include virtual
 372                                            // parents).
 373   llvm::SmallVector<Reference, 4>
 374       VirtualParents; // List of virtual base/parent records.
 375
 376   std::vector<BaseRecordInfo>
 377       Bases; // List of base/parent records; this includes inherited methods and
 378              // attributes
 379
 380   ScopeChildren Children;
 381 };
 382
 383 // Info for typedef and using statements.
 384 struct TypedefInfo : public SymbolInfo {
 385   TypedefInfo(SymbolID USR = SymbolID())
 386       : SymbolInfo(InfoType::IT_typedef, USR) {}
 387
 388   void merge(TypedefInfo &&I);
 389
 390   TypeInfo Underlying;
 391
 392   // Inidicates if this is a new C++ "using"-style typedef:
 393   //   using MyVector = std::vector<int>
 394   // False means it's a C-style typedef:
 395   //   typedef std::vector<int> MyVector;
 396   bool IsUsing = false;
 397 };
 398
 399 struct BaseRecordInfo : public RecordInfo {
 400   BaseRecordInfo();
 401   BaseRecordInfo(SymbolID USR, StringRef Name, StringRef Path, bool IsVirtual,
 402                  AccessSpecifier Access, bool IsParent);
 403
 404   // Indicates if base corresponds to a virtual inheritance
 405   bool IsVirtual = false;
 406   // Access level associated with this inherited info (public, protected,
 407   // private).
 408   AccessSpecifier Access = AccessSpecifier::AS_public;
 409   bool IsParent = false; // Indicates if this base is a direct parent
 410 };
 411
 412 // Information for a single possible value of an enumeration.
 413 struct EnumValueInfo {
 414   explicit EnumValueInfo(StringRef Name = StringRef(),
 415                          StringRef Value = StringRef("0"),
 416                          StringRef ValueExpr = StringRef())
 417       : Name(Name), Value(Value), ValueExpr(ValueExpr) {}
 418
 419   bool operator==(const EnumValueInfo &Other) const {
 420     return std::tie(Name, Value, ValueExpr) ==
 421            std::tie(Other.Name, Other.Value, Other.ValueExpr);
 422   }
 423
 424   SmallString<16> Name;
 425
 426   // The computed value of the enumeration constant. This could be the result of
 427   // evaluating the ValueExpr, or it could be automatically generated according
 428   // to C rules.
 429   SmallString<16> Value;
 430
 431   // Stores the user-supplied initialization expression for this enumeration
 432   // constant. This will be empty for implicit enumeration values.
 433   SmallString<16> ValueExpr;
 434 };
 435
 436 // TODO: Expand to allow for documenting templating.
 437 // Info for types.
 438 struct EnumInfo : public SymbolInfo {
 439   EnumInfo() : SymbolInfo(InfoType::IT_enum) {}
 440   EnumInfo(SymbolID USR) : SymbolInfo(InfoType::IT_enum, USR) {}
 441
 442   void merge(EnumInfo &&I);
 443
 444   // Indicates whether this enum is scoped (e.g. enum class).
 445   bool Scoped = false;
 446
 447   // Set to nonempty to the type when this is an explicitly typed enum. For
 448   //   enum Foo : short { ... };
 449   // this will be "short".
 450   std::optional<TypeInfo> BaseType;
 451
 452   llvm::SmallVector<EnumValueInfo, 4> Members; // List of enum members.
 453 };
 454
 455 struct Index : public Reference {
 456   Index() = default;
 457   Index(StringRef Name) : Reference(SymbolID(), Name) {}
 458   Index(StringRef Name, StringRef JumpToSection)
 459       : Reference(SymbolID(), Name), JumpToSection(JumpToSection) {}
 460   Index(SymbolID USR, StringRef Name, InfoType IT, StringRef Path)
 461       : Reference(USR, Name, IT, Name, Path) {}
 462   // This is used to look for a USR in a vector of Indexes using std::find
 463   bool operator==(const SymbolID &Other) const { return USR == Other; }
 464   bool operator<(const Index &Other) const;
 465
 466   std::optional<SmallString<16>> JumpToSection;
 467   std::vector<Index> Children;
 468
 469   void sort();
 470 };
 471
 472 // TODO: Add functionality to include separate markdown pages.
 473
 474 // A standalone function to call to merge a vector of infos into one.
 475 // This assumes that all infos in the vector are of the same type, and will fail
 476 // if they are different.
 477 llvm::Expected<std::unique_ptr<Info>>
 478 mergeInfos(std::vector<std::unique_ptr<Info>> &Values);
 479
 480 struct ClangDocContext {
 481   ClangDocContext() = default;
 482   ClangDocContext(tooling::ExecutionContext *ECtx, StringRef ProjectName,
 483                   bool PublicOnly, StringRef OutDirectory, StringRef SourceRoot,
 484                   StringRef RepositoryUrl,
 485                   std::vector<std::string> UserStylesheets);
 486   tooling::ExecutionContext *ECtx;
 487   std::string ProjectName; // Name of project clang-doc is documenting.
 488   bool PublicOnly; // Indicates if only public declarations are documented.
 489   std::string OutDirectory; // Directory for outputting generated files.
 490   std::string SourceRoot;   // Directory where processed files are stored. Links
 491                             // to definition locations will only be generated if
 492                             // the file is in this dir.
 493   // URL of repository that hosts code used for links to definition locations.
 494   std::optional<std::string> RepositoryUrl;
 495   // Path of CSS stylesheets that will be copied to OutDirectory and used to
 496   // style all HTML files.
 497   std::vector<std::string> UserStylesheets;
 498   // JavaScript files that will be imported in allHTML file.
 499   std::vector<std::string> JsScripts;
 500   Index Idx;
 501 };
 502
 503 } // namespace doc
 504 } // namespace clang
 505
 506 #endif // LLVM_CLANG_TOOLS_EXTRA_CLANG_DOC_REPRESENTATION_H