Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 1 | //==-- llvm/Support/FileCheck.h ---------------------------*- C++ -*-==// |
| 2 | // |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 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 |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 6 | // |
| 7 | //===----------------------------------------------------------------------===// |
| 8 | // |
| 9 | /// \file This file has some utilities to use FileCheck as an API |
| 10 | // |
| 11 | //===----------------------------------------------------------------------===// |
| 12 | |
| 13 | #ifndef LLVM_SUPPORT_FILECHECK_H |
| 14 | #define LLVM_SUPPORT_FILECHECK_H |
| 15 | |
| 16 | #include "llvm/ADT/StringMap.h" |
| 17 | #include "llvm/Support/MemoryBuffer.h" |
| 18 | #include "llvm/Support/Regex.h" |
| 19 | #include "llvm/Support/SourceMgr.h" |
| 20 | #include <vector> |
| 21 | #include <map> |
| 22 | |
| 23 | namespace llvm { |
| 24 | |
| 25 | /// Contains info about various FileCheck options. |
| 26 | struct FileCheckRequest { |
| 27 | std::vector<std::string> CheckPrefixes; |
| 28 | bool NoCanonicalizeWhiteSpace = false; |
| 29 | std::vector<std::string> ImplicitCheckNot; |
| 30 | std::vector<std::string> GlobalDefines; |
| 31 | bool AllowEmptyInput = false; |
| 32 | bool MatchFullLines = false; |
| 33 | bool EnableVarScope = false; |
| 34 | bool AllowDeprecatedDagOverlap = false; |
| 35 | bool Verbose = false; |
| 36 | bool VerboseVerbose = false; |
| 37 | }; |
| 38 | |
| 39 | |
| 40 | //===----------------------------------------------------------------------===// |
| 41 | // Pattern Handling Code. |
| 42 | //===----------------------------------------------------------------------===// |
| 43 | |
| 44 | namespace Check { |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 45 | |
| 46 | enum FileCheckKind { |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 47 | CheckNone = 0, |
| 48 | CheckPlain, |
| 49 | CheckNext, |
| 50 | CheckSame, |
| 51 | CheckNot, |
| 52 | CheckDAG, |
| 53 | CheckLabel, |
| 54 | CheckEmpty, |
| 55 | |
| 56 | /// Indicates the pattern only matches the end of file. This is used for |
| 57 | /// trailing CHECK-NOTs. |
| 58 | CheckEOF, |
| 59 | |
| 60 | /// Marks when parsing found a -NOT check combined with another CHECK suffix. |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 61 | CheckBadNot, |
| 62 | |
| 63 | /// Marks when parsing found a -COUNT directive with invalid count value. |
| 64 | CheckBadCount |
| 65 | }; |
| 66 | |
| 67 | class FileCheckType { |
| 68 | FileCheckKind Kind; |
| 69 | int Count; ///< optional Count for some checks |
| 70 | |
| 71 | public: |
| 72 | FileCheckType(FileCheckKind Kind = CheckNone) : Kind(Kind), Count(1) {} |
| 73 | FileCheckType(const FileCheckType &) = default; |
| 74 | |
| 75 | operator FileCheckKind() const { return Kind; } |
| 76 | |
| 77 | int getCount() const { return Count; } |
| 78 | FileCheckType &setCount(int C); |
| 79 | |
| 80 | std::string getDescription(StringRef Prefix) const; |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 81 | }; |
| 82 | } |
| 83 | |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 84 | struct FileCheckDiag; |
| 85 | |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 86 | class FileCheckPattern { |
| 87 | SMLoc PatternLoc; |
| 88 | |
| 89 | /// A fixed string to match as the pattern or empty if this pattern requires |
| 90 | /// a regex match. |
| 91 | StringRef FixedStr; |
| 92 | |
| 93 | /// A regex string to match as the pattern or empty if this pattern requires |
| 94 | /// a fixed string to match. |
| 95 | std::string RegExStr; |
| 96 | |
| 97 | /// Entries in this vector map to uses of a variable in the pattern, e.g. |
| 98 | /// "foo[[bar]]baz". In this case, the RegExStr will contain "foobaz" and |
| 99 | /// we'll get an entry in this vector that tells us to insert the value of |
| 100 | /// bar at offset 3. |
| 101 | std::vector<std::pair<StringRef, unsigned>> VariableUses; |
| 102 | |
| 103 | /// Maps definitions of variables to their parenthesized capture numbers. |
Andrew Scull | 0372a57 | 2018-11-16 15:47:06 +0000 | [diff] [blame] | 104 | /// |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 105 | /// E.g. for the pattern "foo[[bar:.*]]baz", VariableDefs will map "bar" to |
| 106 | /// 1. |
| 107 | std::map<StringRef, unsigned> VariableDefs; |
| 108 | |
| 109 | Check::FileCheckType CheckTy; |
| 110 | |
| 111 | /// Contains the number of line this pattern is in. |
| 112 | unsigned LineNumber; |
| 113 | |
| 114 | public: |
| 115 | explicit FileCheckPattern(Check::FileCheckType Ty) |
| 116 | : CheckTy(Ty) {} |
| 117 | |
| 118 | /// Returns the location in source code. |
| 119 | SMLoc getLoc() const { return PatternLoc; } |
| 120 | |
| 121 | bool ParsePattern(StringRef PatternStr, StringRef Prefix, SourceMgr &SM, |
| 122 | unsigned LineNumber, const FileCheckRequest &Req); |
| 123 | size_t Match(StringRef Buffer, size_t &MatchLen, |
| 124 | StringMap<StringRef> &VariableTable) const; |
| 125 | void PrintVariableUses(const SourceMgr &SM, StringRef Buffer, |
| 126 | const StringMap<StringRef> &VariableTable, |
| 127 | SMRange MatchRange = None) const; |
| 128 | void PrintFuzzyMatch(const SourceMgr &SM, StringRef Buffer, |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 129 | const StringMap<StringRef> &VariableTable, |
| 130 | std::vector<FileCheckDiag> *Diags) const; |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 131 | |
| 132 | bool hasVariable() const { |
| 133 | return !(VariableUses.empty() && VariableDefs.empty()); |
| 134 | } |
| 135 | |
| 136 | Check::FileCheckType getCheckTy() const { return CheckTy; } |
| 137 | |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 138 | int getCount() const { return CheckTy.getCount(); } |
| 139 | |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 140 | private: |
| 141 | bool AddRegExToRegEx(StringRef RS, unsigned &CurParen, SourceMgr &SM); |
| 142 | void AddBackrefToRegEx(unsigned BackrefNum); |
| 143 | unsigned |
| 144 | ComputeMatchDistance(StringRef Buffer, |
| 145 | const StringMap<StringRef> &VariableTable) const; |
| 146 | bool EvaluateExpression(StringRef Expr, std::string &Value) const; |
| 147 | size_t FindRegexVarEnd(StringRef Str, SourceMgr &SM); |
| 148 | }; |
| 149 | |
| 150 | //===----------------------------------------------------------------------===// |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 151 | /// Summary of a FileCheck diagnostic. |
| 152 | //===----------------------------------------------------------------------===// |
| 153 | |
| 154 | struct FileCheckDiag { |
| 155 | /// What is the FileCheck directive for this diagnostic? |
| 156 | Check::FileCheckType CheckTy; |
| 157 | /// Where is the FileCheck directive for this diagnostic? |
| 158 | unsigned CheckLine, CheckCol; |
| 159 | /// What type of match result does this diagnostic describe? |
| 160 | /// |
| 161 | /// A directive's supplied pattern is said to be either expected or excluded |
| 162 | /// depending on whether the pattern must have or must not have a match in |
| 163 | /// order for the directive to succeed. For example, a CHECK directive's |
| 164 | /// pattern is expected, and a CHECK-NOT directive's pattern is excluded. |
| 165 | /// All match result types whose names end with "Excluded" are for excluded |
| 166 | /// patterns, and all others are for expected patterns. |
| 167 | /// |
| 168 | /// There might be more than one match result for a single pattern. For |
| 169 | /// example, there might be several discarded matches |
| 170 | /// (MatchFoundButDiscarded) before either a good match |
| 171 | /// (MatchFoundAndExpected) or a failure to match (MatchNoneButExpected), |
| 172 | /// and there might be a fuzzy match (MatchFuzzy) after the latter. |
| 173 | enum MatchType { |
| 174 | /// Indicates a good match for an expected pattern. |
| 175 | MatchFoundAndExpected, |
| 176 | /// Indicates a match for an excluded pattern. |
| 177 | MatchFoundButExcluded, |
| 178 | /// Indicates a match for an expected pattern, but the match is on the |
| 179 | /// wrong line. |
| 180 | MatchFoundButWrongLine, |
| 181 | /// Indicates a discarded match for an expected pattern. |
| 182 | MatchFoundButDiscarded, |
| 183 | /// Indicates no match for an excluded pattern. |
| 184 | MatchNoneAndExcluded, |
| 185 | /// Indicates no match for an expected pattern, but this might follow good |
| 186 | /// matches when multiple matches are expected for the pattern, or it might |
| 187 | /// follow discarded matches for the pattern. |
| 188 | MatchNoneButExpected, |
| 189 | /// Indicates a fuzzy match that serves as a suggestion for the next |
| 190 | /// intended match for an expected pattern with too few or no good matches. |
| 191 | MatchFuzzy, |
| 192 | } MatchTy; |
| 193 | /// The search range if MatchTy is MatchNoneAndExcluded or |
| 194 | /// MatchNoneButExpected, or the match range otherwise. |
| 195 | unsigned InputStartLine; |
| 196 | unsigned InputStartCol; |
| 197 | unsigned InputEndLine; |
| 198 | unsigned InputEndCol; |
| 199 | FileCheckDiag(const SourceMgr &SM, const Check::FileCheckType &CheckTy, |
| 200 | SMLoc CheckLoc, MatchType MatchTy, SMRange InputRange); |
| 201 | }; |
| 202 | |
| 203 | //===----------------------------------------------------------------------===// |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 204 | // Check Strings. |
| 205 | //===----------------------------------------------------------------------===// |
| 206 | |
| 207 | /// A check that we found in the input file. |
| 208 | struct FileCheckString { |
| 209 | /// The pattern to match. |
| 210 | FileCheckPattern Pat; |
| 211 | |
| 212 | /// Which prefix name this check matched. |
| 213 | StringRef Prefix; |
| 214 | |
| 215 | /// The location in the match file that the check string was specified. |
| 216 | SMLoc Loc; |
| 217 | |
| 218 | /// All of the strings that are disallowed from occurring between this match |
| 219 | /// string and the previous one (or start of file). |
| 220 | std::vector<FileCheckPattern> DagNotStrings; |
| 221 | |
| 222 | FileCheckString(const FileCheckPattern &P, StringRef S, SMLoc L) |
| 223 | : Pat(P), Prefix(S), Loc(L) {} |
| 224 | |
| 225 | size_t Check(const SourceMgr &SM, StringRef Buffer, bool IsLabelScanMode, |
| 226 | size_t &MatchLen, StringMap<StringRef> &VariableTable, |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 227 | FileCheckRequest &Req, std::vector<FileCheckDiag> *Diags) const; |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 228 | |
| 229 | bool CheckNext(const SourceMgr &SM, StringRef Buffer) const; |
| 230 | bool CheckSame(const SourceMgr &SM, StringRef Buffer) const; |
| 231 | bool CheckNot(const SourceMgr &SM, StringRef Buffer, |
| 232 | const std::vector<const FileCheckPattern *> &NotStrings, |
| 233 | StringMap<StringRef> &VariableTable, |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 234 | const FileCheckRequest &Req, |
| 235 | std::vector<FileCheckDiag> *Diags) const; |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 236 | size_t CheckDag(const SourceMgr &SM, StringRef Buffer, |
| 237 | std::vector<const FileCheckPattern *> &NotStrings, |
| 238 | StringMap<StringRef> &VariableTable, |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 239 | const FileCheckRequest &Req, |
| 240 | std::vector<FileCheckDiag> *Diags) const; |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 241 | }; |
| 242 | |
| 243 | /// FileCheck class takes the request and exposes various methods that |
| 244 | /// use information from the request. |
| 245 | class FileCheck { |
| 246 | FileCheckRequest Req; |
| 247 | |
| 248 | public: |
| 249 | FileCheck(FileCheckRequest Req) : Req(Req) {} |
| 250 | |
| 251 | // Combines the check prefixes into a single regex so that we can efficiently |
| 252 | // scan for any of the set. |
| 253 | // |
| 254 | // The semantics are that the longest-match wins which matches our regex |
| 255 | // library. |
| 256 | Regex buildCheckPrefixRegex(); |
| 257 | |
| 258 | /// Read the check file, which specifies the sequence of expected strings. |
| 259 | /// |
| 260 | /// The strings are added to the CheckStrings vector. Returns true in case of |
| 261 | /// an error, false otherwise. |
| 262 | bool ReadCheckFile(SourceMgr &SM, StringRef Buffer, Regex &PrefixRE, |
| 263 | std::vector<FileCheckString> &CheckStrings); |
| 264 | |
| 265 | bool ValidateCheckPrefixes(); |
| 266 | |
| 267 | /// Canonicalize whitespaces in the file. Line endings are replaced with |
| 268 | /// UNIX-style '\n'. |
| 269 | StringRef CanonicalizeFile(MemoryBuffer &MB, |
| 270 | SmallVectorImpl<char> &OutputBuffer); |
| 271 | |
| 272 | /// Check the input to FileCheck provided in the \p Buffer against the \p |
| 273 | /// CheckStrings read from the check file. |
| 274 | /// |
| 275 | /// Returns false if the input fails to satisfy the checks. |
| 276 | bool CheckInput(SourceMgr &SM, StringRef Buffer, |
Andrew Walbran | 16937d0 | 2019-10-22 13:54:20 +0100 | [diff] [blame] | 277 | ArrayRef<FileCheckString> CheckStrings, |
| 278 | std::vector<FileCheckDiag> *Diags = nullptr); |
Andrew Scull | cdfcccc | 2018-10-05 20:58:37 +0100 | [diff] [blame] | 279 | }; |
| 280 | } // namespace llvm |
| 281 | #endif |