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
27namespace clang {
28namespace doc {
29
30// SHA1'd hash of a USR.
31using SymbolID = std::array<uint8_t, 20>;
32
33struct BaseRecordInfo;
34struct EnumInfo;
35struct FunctionInfo;
36struct Info;
37struct TypedefInfo;
38
39enum 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.
49struct 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
87struct 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(args: USR, args: Name, args: QualName, args: RefType) ==
102 std::tie(args: Other.USR, args: Other.Name, args: QualName, args: Other.RefType);
103 }
104
105 bool mergeable(const Reference &Other);
106 void merge(Reference &&I);
107 bool operator<(const Reference &Other) const { return Name < Other.Name; }
108
109 /// Returns the path for this Reference relative to CurrentPath.
110 llvm::SmallString<64> getRelativeFilePath(const StringRef &CurrentPath) const;
111
112 /// Returns the basename that should be used for this Reference.
113 llvm::SmallString<16> getFileBaseName() const;
114
115 SymbolID USR = SymbolID(); // Unique identifier for referenced decl
116
117 // Name of type (possibly unresolved). Not including namespaces or template
118 // parameters (so for a std::vector<int> this would be "vector"). See also
119 // QualName.
120 SmallString<16> Name;
121
122 // Full qualified name of this type, including namespaces and template
123 // parameter (for example this could be "std::vector<int>"). Contrast to
124 // Name.
125 SmallString<16> QualName;
126
127 InfoType RefType = InfoType::IT_default; // Indicates the type of this
128 // Reference (namespace, record,
129 // function, enum, default).
130 // Path of directory where the clang-doc generated file will be saved
131 // (possibly unresolved)
132 llvm::SmallString<128> Path;
133};
134
135// Holds the children of a record or namespace.
136struct ScopeChildren {
137 // Namespaces and Records are references because they will be properly
138 // documented in their own info, while the entirety of Functions and Enums are
139 // included here because they should not have separate documentation from
140 // their scope.
141 //
142 // Namespaces are not syntactically valid as children of records, but making
143 // this general for all possible container types reduces code complexity.
144 std::vector<Reference> Namespaces;
145 std::vector<Reference> Records;
146 std::vector<FunctionInfo> Functions;
147 std::vector<EnumInfo> Enums;
148 std::vector<TypedefInfo> Typedefs;
149
150 void sort();
151};
152
153// A base struct for TypeInfos
154struct TypeInfo {
155 TypeInfo() = default;
156 TypeInfo(const Reference &R) : Type(R) {}
157
158 // Convenience constructor for when there is no symbol ID or info type
159 // (normally used for built-in types in tests).
160 TypeInfo(StringRef Name, StringRef Path = StringRef())
161 : Type(SymbolID(), Name, InfoType::IT_default, Name, Path) {}
162
163 bool operator==(const TypeInfo &Other) const { return Type == Other.Type; }
164
165 Reference Type; // Referenced type in this info.
166};
167
168// Represents one template parameter.
169//
170// This is a very simple serialization of the text of the source code of the
171// template parameter. It is saved in a struct so there is a place to add the
172// name and default values in the future if needed.
173struct TemplateParamInfo {
174 TemplateParamInfo() = default;
175 explicit TemplateParamInfo(StringRef Contents) : Contents(Contents) {}
176
177 // The literal contents of the code for that specifies this template parameter
178 // for this declaration. Typical values will be "class T" and
179 // "typename T = int".
180 SmallString<16> Contents;
181};
182
183struct TemplateSpecializationInfo {
184 // Indicates the declaration that this specializes.
185 SymbolID SpecializationOf;
186
187 // Template parameters applying to the specialized record/function.
188 std::vector<TemplateParamInfo> Params;
189};
190
191// Records the template information for a struct or function that is a template
192// or an explicit template specialization.
193struct TemplateInfo {
194 // May be empty for non-partial specializations.
195 std::vector<TemplateParamInfo> Params;
196
197 // Set when this is a specialization of another record/function.
198 std::optional<TemplateSpecializationInfo> Specialization;
199};
200
201// Info for field types.
202struct FieldTypeInfo : public TypeInfo {
203 FieldTypeInfo() = default;
204 FieldTypeInfo(const TypeInfo &TI, StringRef Name = StringRef(),
205 StringRef DefaultValue = StringRef())
206 : TypeInfo(TI), Name(Name), DefaultValue(DefaultValue) {}
207
208 bool operator==(const FieldTypeInfo &Other) const {
209 return std::tie(args: Type, args: Name, args: DefaultValue) ==
210 std::tie(args: Other.Type, args: Other.Name, args: Other.DefaultValue);
211 }
212
213 SmallString<16> Name; // Name associated with this info.
214
215 // When used for function parameters, contains the string representing the
216 // expression of the default value, if any.
217 SmallString<16> DefaultValue;
218};
219
220// Info for member types.
221struct MemberTypeInfo : public FieldTypeInfo {
222 MemberTypeInfo() = default;
223 MemberTypeInfo(const TypeInfo &TI, StringRef Name, AccessSpecifier Access)
224 : FieldTypeInfo(TI, Name), Access(Access) {}
225
226 bool operator==(const MemberTypeInfo &Other) const {
227 return std::tie(args: Type, args: Name, args: Access, args: Description) ==
228 std::tie(args: Other.Type, args: Other.Name, args: Other.Access, args: Other.Description);
229 }
230
231 // Access level associated with this info (public, protected, private, none).
232 // AS_public is set as default because the bitcode writer requires the enum
233 // with value 0 to be used as the default.
234 // (AS_public = 0, AS_protected = 1, AS_private = 2, AS_none = 3)
235 AccessSpecifier Access = AccessSpecifier::AS_public;
236
237 std::vector<CommentInfo> Description; // Comment description of this field.
238};
239
240struct Location {
241 Location(int LineNumber = 0, StringRef Filename = StringRef(),
242 bool IsFileInRootDir = false)
243 : LineNumber(LineNumber), Filename(Filename),
244 IsFileInRootDir(IsFileInRootDir) {}
245
246 bool operator==(const Location &Other) const {
247 return std::tie(args: LineNumber, args: Filename) ==
248 std::tie(args: Other.LineNumber, args: Other.Filename);
249 }
250
251 bool operator!=(const Location &Other) const {
252 return std::tie(args: LineNumber, args: Filename) !=
253 std::tie(args: Other.LineNumber, args: Other.Filename);
254 }
255
256 // This operator is used to sort a vector of Locations.
257 // No specific order (attributes more important than others) is required. Any
258 // sort is enough, the order is only needed to call std::unique after sorting
259 // the vector.
260 bool operator<(const Location &Other) const {
261 return std::tie(args: LineNumber, args: Filename) <
262 std::tie(args: Other.LineNumber, args: Other.Filename);
263 }
264
265 int LineNumber = 0; // Line number of this Location.
266 SmallString<32> Filename; // File for this Location.
267 bool IsFileInRootDir = false; // Indicates if file is inside root directory
268};
269
270/// A base struct for Infos.
271struct Info {
272 Info(InfoType IT = InfoType::IT_default, SymbolID USR = SymbolID(),
273 StringRef Name = StringRef(), StringRef Path = StringRef())
274 : USR(USR), IT(IT), Name(Name), Path(Path) {}
275
276 Info(const Info &Other) = delete;
277 Info(Info &&Other) = default;
278
279 virtual ~Info() = default;
280
281 Info &operator=(Info &&Other) = default;
282
283 SymbolID USR =
284 SymbolID(); // Unique identifier for the decl described by this Info.
285 InfoType IT = InfoType::IT_default; // InfoType of this particular Info.
286 SmallString<16> Name; // Unqualified name of the decl.
287 llvm::SmallVector<Reference, 4>
288 Namespace; // List of parent namespaces for this decl.
289 std::vector<CommentInfo> Description; // Comment description of this decl.
290 llvm::SmallString<128> Path; // Path of directory where the clang-doc
291 // generated file will be saved
292
293 void mergeBase(Info &&I);
294 bool mergeable(const Info &Other);
295
296 llvm::SmallString<16> extractName() const;
297
298 /// Returns the file path for this Info relative to CurrentPath.
299 llvm::SmallString<64> getRelativeFilePath(const StringRef &CurrentPath) const;
300
301 /// Returns the basename that should be used for this Info.
302 llvm::SmallString<16> getFileBaseName() const;
303};
304
305// Info for namespaces.
306struct NamespaceInfo : public Info {
307 NamespaceInfo(SymbolID USR = SymbolID(), StringRef Name = StringRef(),
308 StringRef Path = StringRef());
309
310 void merge(NamespaceInfo &&I);
311
312 ScopeChildren Children;
313};
314
315// Info for symbols.
316struct SymbolInfo : public Info {
317 SymbolInfo(InfoType IT, SymbolID USR = SymbolID(),
318 StringRef Name = StringRef(), StringRef Path = StringRef())
319 : Info(IT, USR, Name, Path) {}
320
321 void merge(SymbolInfo &&I);
322
323 std::optional<Location> DefLoc; // Location where this decl is defined.
324 llvm::SmallVector<Location, 2> Loc; // Locations where this decl is declared.
325
326 bool operator<(const SymbolInfo &Other) const {
327 // Sort by declaration location since we want the doc to be
328 // generated in the order of the source code.
329 // If the declaration location is the same, or not present
330 // we sort by defined location otherwise fallback to the extracted name
331 if (Loc.size() > 0 && Other.Loc.size() > 0 && Loc[0] != Other.Loc[0])
332 return Loc[0] < Other.Loc[0];
333
334 if (DefLoc && Other.DefLoc && *DefLoc != *Other.DefLoc)
335 return *DefLoc < *Other.DefLoc;
336
337 return extractName() < Other.extractName();
338 }
339};
340
341// TODO: Expand to allow for documenting templating and default args.
342// Info for functions.
343struct FunctionInfo : public SymbolInfo {
344 FunctionInfo(SymbolID USR = SymbolID())
345 : SymbolInfo(InfoType::IT_function, USR) {}
346
347 void merge(FunctionInfo &&I);
348
349 bool IsMethod = false; // Indicates whether this function is a class method.
350 Reference Parent; // Reference to the parent class decl for this method.
351 TypeInfo ReturnType; // Info about the return type of this function.
352 llvm::SmallVector<FieldTypeInfo, 4> Params; // List of parameters.
353 // Access level for this method (public, private, protected, none).
354 // AS_public is set as default because the bitcode writer requires the enum
355 // with value 0 to be used as the default.
356 // (AS_public = 0, AS_protected = 1, AS_private = 2, AS_none = 3)
357 AccessSpecifier Access = AccessSpecifier::AS_public;
358
359 // Full qualified name of this function, including namespaces and template
360 // specializations.
361 SmallString<16> FullName;
362
363 // When present, this function is a template or specialization.
364 std::optional<TemplateInfo> Template;
365};
366
367// TODO: Expand to allow for documenting templating, inheritance access,
368// friend classes
369// Info for types.
370struct RecordInfo : public SymbolInfo {
371 RecordInfo(SymbolID USR = SymbolID(), StringRef Name = StringRef(),
372 StringRef Path = StringRef());
373
374 void merge(RecordInfo &&I);
375
376 // Type of this record (struct, class, union, interface).
377 TagTypeKind TagType = TagTypeKind::Struct;
378
379 // Full qualified name of this record, including namespaces and template
380 // specializations.
381 SmallString<16> FullName;
382
383 // When present, this record is a template or specialization.
384 std::optional<TemplateInfo> Template;
385
386 // Indicates if the record was declared using a typedef. Things like anonymous
387 // structs in a typedef:
388 // typedef struct { ... } foo_t;
389 // are converted into records with the typedef as the Name + this flag set.
390 bool IsTypeDef = false;
391
392 llvm::SmallVector<MemberTypeInfo, 4>
393 Members; // List of info about record members.
394 llvm::SmallVector<Reference, 4> Parents; // List of base/parent records
395 // (does not include virtual
396 // parents).
397 llvm::SmallVector<Reference, 4>
398 VirtualParents; // List of virtual base/parent records.
399
400 std::vector<BaseRecordInfo>
401 Bases; // List of base/parent records; this includes inherited methods and
402 // attributes
403
404 ScopeChildren Children;
405};
406
407// Info for typedef and using statements.
408struct TypedefInfo : public SymbolInfo {
409 TypedefInfo(SymbolID USR = SymbolID())
410 : SymbolInfo(InfoType::IT_typedef, USR) {}
411
412 void merge(TypedefInfo &&I);
413
414 TypeInfo Underlying;
415
416 // Inidicates if this is a new C++ "using"-style typedef:
417 // using MyVector = std::vector<int>
418 // False means it's a C-style typedef:
419 // typedef std::vector<int> MyVector;
420 bool IsUsing = false;
421};
422
423struct BaseRecordInfo : public RecordInfo {
424 BaseRecordInfo();
425 BaseRecordInfo(SymbolID USR, StringRef Name, StringRef Path, bool IsVirtual,
426 AccessSpecifier Access, bool IsParent);
427
428 // Indicates if base corresponds to a virtual inheritance
429 bool IsVirtual = false;
430 // Access level associated with this inherited info (public, protected,
431 // private).
432 AccessSpecifier Access = AccessSpecifier::AS_public;
433 bool IsParent = false; // Indicates if this base is a direct parent
434};
435
436// Information for a single possible value of an enumeration.
437struct EnumValueInfo {
438 explicit EnumValueInfo(StringRef Name = StringRef(),
439 StringRef Value = StringRef("0"),
440 StringRef ValueExpr = StringRef())
441 : Name(Name), Value(Value), ValueExpr(ValueExpr) {}
442
443 bool operator==(const EnumValueInfo &Other) const {
444 return std::tie(args: Name, args: Value, args: ValueExpr) ==
445 std::tie(args: Other.Name, args: Other.Value, args: Other.ValueExpr);
446 }
447
448 SmallString<16> Name;
449
450 // The computed value of the enumeration constant. This could be the result of
451 // evaluating the ValueExpr, or it could be automatically generated according
452 // to C rules.
453 SmallString<16> Value;
454
455 // Stores the user-supplied initialization expression for this enumeration
456 // constant. This will be empty for implicit enumeration values.
457 SmallString<16> ValueExpr;
458
459 std::vector<CommentInfo> Description; /// Comment description of this field.
460};
461
462// TODO: Expand to allow for documenting templating.
463// Info for types.
464struct EnumInfo : public SymbolInfo {
465 EnumInfo() : SymbolInfo(InfoType::IT_enum) {}
466 EnumInfo(SymbolID USR) : SymbolInfo(InfoType::IT_enum, USR) {}
467
468 void merge(EnumInfo &&I);
469
470 // Indicates whether this enum is scoped (e.g. enum class).
471 bool Scoped = false;
472
473 // Set to nonempty to the type when this is an explicitly typed enum. For
474 // enum Foo : short { ... };
475 // this will be "short".
476 std::optional<TypeInfo> BaseType;
477
478 llvm::SmallVector<EnumValueInfo, 4> Members; // List of enum members.
479};
480
481struct Index : public Reference {
482 Index() = default;
483 Index(StringRef Name) : Reference(SymbolID(), Name) {}
484 Index(StringRef Name, StringRef JumpToSection)
485 : Reference(SymbolID(), Name), JumpToSection(JumpToSection) {}
486 Index(SymbolID USR, StringRef Name, InfoType IT, StringRef Path)
487 : Reference(USR, Name, IT, Name, Path) {}
488 // This is used to look for a USR in a vector of Indexes using std::find
489 bool operator==(const SymbolID &Other) const { return USR == Other; }
490 bool operator<(const Index &Other) const;
491
492 std::optional<SmallString<16>> JumpToSection;
493 std::vector<Index> Children;
494
495 void sort();
496};
497
498// TODO: Add functionality to include separate markdown pages.
499
500// A standalone function to call to merge a vector of infos into one.
501// This assumes that all infos in the vector are of the same type, and will fail
502// if they are different.
503llvm::Expected<std::unique_ptr<Info>>
504mergeInfos(std::vector<std::unique_ptr<Info>> &Values);
505
506struct ClangDocContext {
507 ClangDocContext() = default;
508 ClangDocContext(tooling::ExecutionContext *ECtx, StringRef ProjectName,
509 bool PublicOnly, StringRef OutDirectory, StringRef SourceRoot,
510 StringRef RepositoryUrl,
511 std::vector<std::string> UserStylesheets);
512 tooling::ExecutionContext *ECtx;
513 std::string ProjectName; // Name of project clang-doc is documenting.
514 bool PublicOnly; // Indicates if only public declarations are documented.
515 std::string OutDirectory; // Directory for outputting generated files.
516 std::string SourceRoot; // Directory where processed files are stored. Links
517 // to definition locations will only be generated if
518 // the file is in this dir.
519 // URL of repository that hosts code used for links to definition locations.
520 std::optional<std::string> RepositoryUrl;
521 // Path of CSS stylesheets that will be copied to OutDirectory and used to
522 // style all HTML files.
523 std::vector<std::string> UserStylesheets;
524 // JavaScript files that will be imported in allHTML file.
525 std::vector<std::string> JsScripts;
526 Index Idx;
527};
528
529} // namespace doc
530} // namespace clang
531
532#endif // LLVM_CLANG_TOOLS_EXTRA_CLANG_DOC_REPRESENTATION_H
533