1 //===--- SarifDiagnostics.cpp - Sarif Diagnostics for Paths -----*- C++ -*-===//
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
7 //===----------------------------------------------------------------------===//
9 // This file defines the SarifDiagnostics object.
11 //===----------------------------------------------------------------------===//
13 #include "clang/Basic/Version.h"
14 #include "clang/Lex/Preprocessor.h"
15 #include "clang/StaticAnalyzer/Core/AnalyzerOptions.h"
16 #include "clang/StaticAnalyzer/Core/BugReporter/PathDiagnostic.h"
17 #include "clang/StaticAnalyzer/Core/PathDiagnosticConsumers.h"
18 #include "llvm/ADT/STLExtras.h"
19 #include "llvm/ADT/StringMap.h"
20 #include "llvm/Support/JSON.h"
21 #include "llvm/Support/Path.h"
24 using namespace clang;
28 class SarifDiagnostics : public PathDiagnosticConsumer {
29 std::string OutputFile;
32 SarifDiagnostics(AnalyzerOptions &, const std::string &Output)
33 : OutputFile(Output) {}
34 ~SarifDiagnostics() override = default;
36 void FlushDiagnosticsImpl(std::vector<const PathDiagnostic *> &Diags,
37 FilesMade *FM) override;
39 StringRef getName() const override { return "SarifDiagnostics"; }
40 PathGenerationScheme getGenerationScheme() const override { return Minimal; }
41 bool supportsLogicalOpControlFlow() const override { return true; }
42 bool supportsCrossFileDiagnostics() const override { return true; }
44 } // end anonymous namespace
46 void ento::createSarifDiagnosticConsumer(AnalyzerOptions &AnalyzerOpts,
47 PathDiagnosticConsumers &C,
48 const std::string &Output,
49 const Preprocessor &) {
50 C.push_back(new SarifDiagnostics(AnalyzerOpts, Output));
53 static StringRef getFileName(const FileEntry &FE) {
54 StringRef Filename = FE.tryGetRealPathName();
56 Filename = FE.getName();
60 static std::string percentEncodeURICharacter(char C) {
61 // RFC 3986 claims alpha, numeric, and this handful of
62 // characters are not reserved for the path component and
63 // should be written out directly. Otherwise, percent
64 // encode the character and write that out instead of the
65 // reserved character.
66 if (llvm::isAlnum(C) ||
67 StringRef::npos != StringRef("-._~:@!$&'()*+,;=").find(C))
68 return std::string(&C, 1);
69 return "%" + llvm::toHex(StringRef(&C, 1));
72 static std::string fileNameToURI(StringRef Filename) {
73 llvm::SmallString<32> Ret = StringRef("file://");
75 // Get the root name to see if it has a URI authority.
76 StringRef Root = sys::path::root_name(Filename);
77 if (Root.startswith("//")) {
78 // There is an authority, so add it to the URI.
79 Ret += Root.drop_front(2).str();
80 } else if (!Root.empty()) {
81 // There is no authority, so end the component and add the root to the URI.
82 Ret += Twine("/" + Root).str();
85 auto Iter = sys::path::begin(Filename), End = sys::path::end(Filename);
86 assert(Iter != End && "Expected there to be a non-root path component.");
87 // Add the rest of the path components, encoding any reserved characters;
88 // we skip past the first path component, as it was handled it above.
89 std::for_each(++Iter, End, [&Ret](StringRef Component) {
90 // For reasons unknown to me, we may get a backslash with Windows native
91 // paths for the initial backslash following the drive component, which
92 // we need to ignore as a URI path part.
93 if (Component == "\\")
96 // Add the separator between the previous path part and the one being
97 // currently processed.
100 // URI encode the part.
101 for (char C : Component) {
102 Ret += percentEncodeURICharacter(C);
106 return Ret.str().str();
109 static json::Object createFileLocation(const FileEntry &FE) {
110 return json::Object{{"uri", fileNameToURI(getFileName(FE))}};
113 static json::Object createFile(const FileEntry &FE) {
114 return json::Object{{"fileLocation", createFileLocation(FE)},
115 {"roles", json::Array{"resultFile"}},
116 {"length", FE.getSize()},
117 {"mimeType", "text/plain"}};
120 static json::Object createFileLocation(const FileEntry &FE,
121 json::Array &Files) {
122 std::string FileURI = fileNameToURI(getFileName(FE));
124 // See if the Files array contains this URI already. If it does not, create
125 // a new file object to add to the array.
126 auto I = llvm::find_if(Files, [&](const json::Value &File) {
127 if (const json::Object *Obj = File.getAsObject()) {
128 if (const json::Object *FileLoc = Obj->getObject("fileLocation")) {
129 Optional<StringRef> URI = FileLoc->getString("uri");
130 return URI && URI->equals(FileURI);
136 // Calculate the index within the file location array so it can be stored in
138 auto Index = static_cast<unsigned>(std::distance(Files.begin(), I));
139 if (I == Files.end())
140 Files.push_back(createFile(FE));
142 return json::Object{{"uri", FileURI}, {"fileIndex", Index}};
145 static json::Object createTextRegion(SourceRange R, const SourceManager &SM) {
147 {"startLine", SM.getExpansionLineNumber(R.getBegin())},
148 {"endLine", SM.getExpansionLineNumber(R.getEnd())},
149 {"startColumn", SM.getExpansionColumnNumber(R.getBegin())},
150 {"endColumn", SM.getExpansionColumnNumber(R.getEnd())}};
153 static json::Object createPhysicalLocation(SourceRange R, const FileEntry &FE,
154 const SourceManager &SMgr,
155 json::Array &Files) {
156 return json::Object{{{"fileLocation", createFileLocation(FE, Files)},
157 {"region", createTextRegion(R, SMgr)}}};
160 enum class Importance { Important, Essential, Unimportant };
162 static StringRef importanceToStr(Importance I) {
164 case Importance::Important:
166 case Importance::Essential:
168 case Importance::Unimportant:
169 return "unimportant";
171 llvm_unreachable("Fully covered switch is not so fully covered");
174 static json::Object createThreadFlowLocation(json::Object &&Location,
176 return json::Object{{"location", std::move(Location)},
177 {"importance", importanceToStr(I)}};
180 static json::Object createMessage(StringRef Text) {
181 return json::Object{{"text", Text.str()}};
184 static json::Object createLocation(json::Object &&PhysicalLocation,
185 StringRef Message = "") {
186 json::Object Ret{{"physicalLocation", std::move(PhysicalLocation)}};
187 if (!Message.empty())
188 Ret.insert({"message", createMessage(Message)});
192 static Importance calculateImportance(const PathDiagnosticPiece &Piece) {
193 switch (Piece.getKind()) {
194 case PathDiagnosticPiece::Call:
195 case PathDiagnosticPiece::Macro:
196 case PathDiagnosticPiece::Note:
197 case PathDiagnosticPiece::PopUp:
198 // FIXME: What should be reported here?
200 case PathDiagnosticPiece::Event:
201 return Piece.getTagStr() == "ConditionBRVisitor" ? Importance::Important
202 : Importance::Essential;
203 case PathDiagnosticPiece::ControlFlow:
204 return Importance::Unimportant;
206 return Importance::Unimportant;
209 static json::Object createThreadFlow(const PathPieces &Pieces,
210 json::Array &Files) {
211 const SourceManager &SMgr = Pieces.front()->getLocation().getManager();
212 json::Array Locations;
213 for (const auto &Piece : Pieces) {
214 const PathDiagnosticLocation &P = Piece->getLocation();
215 Locations.push_back(createThreadFlowLocation(
216 createLocation(createPhysicalLocation(P.asRange(),
217 *P.asLocation().getFileEntry(),
220 calculateImportance(*Piece)));
222 return json::Object{{"locations", std::move(Locations)}};
225 static json::Object createCodeFlow(const PathPieces &Pieces,
226 json::Array &Files) {
228 {"threadFlows", json::Array{createThreadFlow(Pieces, Files)}}};
231 static json::Object createTool() {
232 return json::Object{{"name", "clang"},
233 {"fullName", "clang static analyzer"},
234 {"language", "en-US"},
235 {"version", getClangFullVersion()}};
238 static json::Object createResult(const PathDiagnostic &Diag, json::Array &Files,
239 const StringMap<unsigned> &RuleMapping) {
240 const PathPieces &Path = Diag.path.flatten(false);
241 const SourceManager &SMgr = Path.front()->getLocation().getManager();
243 auto Iter = RuleMapping.find(Diag.getCheckName());
244 assert(Iter != RuleMapping.end() && "Rule ID is not in the array index map?");
247 {"message", createMessage(Diag.getVerboseDescription())},
248 {"codeFlows", json::Array{createCodeFlow(Path, Files)}},
250 json::Array{createLocation(createPhysicalLocation(
251 Diag.getLocation().asRange(),
252 *Diag.getLocation().asLocation().getFileEntry(), SMgr, Files))}},
253 {"ruleIndex", Iter->getValue()},
254 {"ruleId", Diag.getCheckName()}};
257 static StringRef getRuleDescription(StringRef CheckName) {
258 return llvm::StringSwitch<StringRef>(CheckName)
260 #define CHECKER(FULLNAME, CLASS, HELPTEXT, DOC_URI, IS_HIDDEN) \
261 .Case(FULLNAME, HELPTEXT)
262 #include "clang/StaticAnalyzer/Checkers/Checkers.inc"
268 static StringRef getRuleHelpURIStr(StringRef CheckName) {
269 return llvm::StringSwitch<StringRef>(CheckName)
271 #define CHECKER(FULLNAME, CLASS, HELPTEXT, DOC_URI, IS_HIDDEN) \
272 .Case(FULLNAME, DOC_URI)
273 #include "clang/StaticAnalyzer/Checkers/Checkers.inc"
279 static json::Object createRule(const PathDiagnostic &Diag) {
280 StringRef CheckName = Diag.getCheckName();
282 {"fullDescription", createMessage(getRuleDescription(CheckName))},
283 {"name", createMessage(CheckName)},
286 std::string RuleURI = getRuleHelpURIStr(CheckName);
287 if (!RuleURI.empty())
288 Ret["helpUri"] = RuleURI;
293 static json::Array createRules(std::vector<const PathDiagnostic *> &Diags,
294 StringMap<unsigned> &RuleMapping) {
296 llvm::StringSet<> Seen;
298 llvm::for_each(Diags, [&](const PathDiagnostic *D) {
299 StringRef RuleID = D->getCheckName();
300 std::pair<llvm::StringSet<>::iterator, bool> P = Seen.insert(RuleID);
302 RuleMapping[RuleID] = Rules.size(); // Maps RuleID to an Array Index.
303 Rules.push_back(createRule(*D));
310 static json::Object createResources(std::vector<const PathDiagnostic *> &Diags,
311 StringMap<unsigned> &RuleMapping) {
312 return json::Object{{"rules", createRules(Diags, RuleMapping)}};
315 static json::Object createRun(std::vector<const PathDiagnostic *> &Diags) {
316 json::Array Results, Files;
317 StringMap<unsigned> RuleMapping;
318 json::Object Resources = createResources(Diags, RuleMapping);
320 llvm::for_each(Diags, [&](const PathDiagnostic *D) {
321 Results.push_back(createResult(*D, Files, RuleMapping));
324 return json::Object{{"tool", createTool()},
325 {"resources", std::move(Resources)},
326 {"results", std::move(Results)},
327 {"files", std::move(Files)}};
330 void SarifDiagnostics::FlushDiagnosticsImpl(
331 std::vector<const PathDiagnostic *> &Diags, FilesMade *) {
332 // We currently overwrite the file if it already exists. However, it may be
333 // useful to add a feature someday that allows the user to append a run to an
334 // existing SARIF file. One danger from that approach is that the size of the
335 // file can become large very quickly, so decoding into JSON to append a run
336 // may be an expensive operation.
338 llvm::raw_fd_ostream OS(OutputFile, EC, llvm::sys::fs::F_Text);
340 llvm::errs() << "warning: could not create file: " << EC.message() << '\n';
345 "http://json.schemastore.org/sarif-2.0.0-csd.2.beta.2018-11-28"},
346 {"version", "2.0.0-csd.2.beta.2018-11-28"},
347 {"runs", json::Array{createRun(Diags)}}};
348 OS << llvm::formatv("{0:2}\n", json::Value(std::move(Sarif)));