1.1
|
(pascal 03-Sep-16)
| 1
|
: //===- FileCheck.cpp - Check that File's Contents match what is expected --===//
|
|
| 2
|
: //
|
|
| 3
|
: // The LLVM Compiler Infrastructure
|
|
| 4
|
: //
|
|
| 5
|
: // This file is distributed under the University of Illinois Open Source
|
|
| 6
|
: // License. See LICENSE.TXT for details.
|
|
| 7
|
: //
|
|
| 8
|
: //===----------------------------------------------------------------------===//
|
|
| 9
|
: //
|
|
| 10
|
: // FileCheck does a line-by line check of a file that validates whether it
|
|
| 11
|
: // contains the expected content. This is useful for regression tests etc.
|
|
| 12
|
: //
|
1.1.1.4
|
(patrick 04-Oct-17)
| 13
|
: // This program exits with an exit status of 2 on error, exit status of 0 if
|
1.1
|
(pascal 03-Sep-16)
| 14
|
: // the file matched the expected contents, and exit status of 1 if it did not
|
|
| 15
|
: // contain the expected contents.
|
|
| 16
|
: //
|
|
| 17
|
: //===----------------------------------------------------------------------===//
|
|
| 18
|
:
|
|
| 19
|
: #include "llvm/ADT/SmallString.h"
|
|
| 20
|
: #include "llvm/ADT/StringExtras.h"
|
|
| 21
|
: #include "llvm/ADT/StringMap.h"
|
|
| 22
|
: #include "llvm/ADT/StringSet.h"
|
|
| 23
|
: #include "llvm/Support/CommandLine.h"
|
|
| 24
|
: #include "llvm/Support/MemoryBuffer.h"
|
|
| 25
|
: #include "llvm/Support/PrettyStackTrace.h"
|
|
| 26
|
: #include "llvm/Support/Regex.h"
|
|
| 27
|
: #include "llvm/Support/Signals.h"
|
|
| 28
|
: #include "llvm/Support/SourceMgr.h"
|
|
| 29
|
: #include "llvm/Support/raw_ostream.h"
|
|
| 30
|
: #include <algorithm>
|
|
| 31
|
: #include <cctype>
|
|
| 32
|
: #include <map>
|
|
| 33
|
: #include <string>
|
|
| 34
|
: #include <system_error>
|
|
| 35
|
: #include <vector>
|
|
| 36
|
: using namespace llvm;
|
|
| 37
|
:
|
|
| 38
|
: static cl::opt<std::string>
|
1.1.1.3
|
(patrick 24-Jan-17)
| 39
|
: CheckFilename(cl::Positional, cl::desc("<check-file>"), cl::Required);
|
1.1
|
(pascal 03-Sep-16)
| 40
|
:
|
|
| 41
|
: static cl::opt<std::string>
|
1.1.1.3
|
(patrick 24-Jan-17)
| 42
|
: InputFilename("input-file", cl::desc("File to check (defaults to stdin)"),
|
|
| 43
|
: cl::init("-"), cl::value_desc("filename"));
|
1.1
|
(pascal 03-Sep-16)
| 44
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 45
|
: static cl::list<std::string> CheckPrefixes(
|
|
| 46
|
: "check-prefix",
|
|
| 47
|
: cl::desc("Prefix to use from check file (defaults to 'CHECK')"));
|
1.1.1.2
|
(patrick 14-Jan-17)
| 48
|
: static cl::alias CheckPrefixesAlias(
|
|
| 49
|
: "check-prefixes", cl::aliasopt(CheckPrefixes), cl::CommaSeparated,
|
|
| 50
|
: cl::NotHidden,
|
|
| 51
|
: cl::desc(
|
|
| 52
|
: "Alias for -check-prefix permitting multiple comma separated values"));
|
1.1
|
(pascal 03-Sep-16)
| 53
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 54
|
: static cl::opt<bool> NoCanonicalizeWhiteSpace(
|
|
| 55
|
: "strict-whitespace",
|
|
| 56
|
: cl::desc("Do not treat all horizontal whitespace as equivalent"));
|
1.1
|
(pascal 03-Sep-16)
| 57
|
:
|
|
| 58
|
: static cl::list<std::string> ImplicitCheckNot(
|
|
| 59
|
: "implicit-check-not",
|
|
| 60
|
: cl::desc("Add an implicit negative check with this pattern to every\n"
|
|
| 61
|
: "positive check. This can be used to ensure that no instances of\n"
|
|
| 62
|
: "this pattern occur which are not matched by a positive pattern"),
|
|
| 63
|
: cl::value_desc("pattern"));
|
|
| 64
|
:
|
|
| 65
|
: static cl::opt<bool> AllowEmptyInput(
|
|
| 66
|
: "allow-empty", cl::init(false),
|
|
| 67
|
: cl::desc("Allow the input file to be empty. This is useful when making\n"
|
|
| 68
|
: "checks that some error message does not occur, for example."));
|
|
| 69
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 70
|
: static cl::opt<bool> MatchFullLines(
|
|
| 71
|
: "match-full-lines", cl::init(false),
|
|
| 72
|
: cl::desc("Require all positive matches to cover an entire input line.\n"
|
|
| 73
|
: "Allows leading and trailing whitespace if --strict-whitespace\n"
|
|
| 74
|
: "is not also passed."));
|
|
| 75
|
:
|
1.1.1.4
|
(patrick 04-Oct-17)
| 76
|
: static cl::opt<bool> EnableVarScope(
|
|
| 77
|
: "enable-var-scope", cl::init(false),
|
|
| 78
|
: cl::desc("Enables scope for regex variables. Variables with names that\n"
|
|
| 79
|
: "do not start with '$' will be reset at the beginning of\n"
|
|
| 80
|
: "each CHECK-LABEL block."));
|
|
| 81
|
:
|
1.1
|
(pascal 03-Sep-16)
| 82
|
: typedef cl::list<std::string>::const_iterator prefix_iterator;
|
|
| 83
|
:
|
|
| 84
|
: //===----------------------------------------------------------------------===//
|
|
| 85
|
: // Pattern Handling Code.
|
|
| 86
|
: //===----------------------------------------------------------------------===//
|
|
| 87
|
:
|
|
| 88
|
: namespace Check {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 89
|
: enum CheckType {
|
|
| 90
|
: CheckNone = 0,
|
|
| 91
|
: CheckPlain,
|
|
| 92
|
: CheckNext,
|
|
| 93
|
: CheckSame,
|
|
| 94
|
: CheckNot,
|
|
| 95
|
: CheckDAG,
|
|
| 96
|
: CheckLabel,
|
|
| 97
|
:
|
|
| 98
|
: /// Indicates the pattern only matches the end of file. This is used for
|
|
| 99
|
: /// trailing CHECK-NOTs.
|
|
| 100
|
: CheckEOF,
|
|
| 101
|
:
|
|
| 102
|
: /// Marks when parsing found a -NOT check combined with another CHECK suffix.
|
|
| 103
|
: CheckBadNot
|
|
| 104
|
: };
|
1.1
|
(pascal 03-Sep-16)
| 105
|
: }
|
|
| 106
|
:
|
|
| 107
|
: class Pattern {
|
|
| 108
|
: SMLoc PatternLoc;
|
|
| 109
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 110
|
: /// A fixed string to match as the pattern or empty if this pattern requires
|
|
| 111
|
: /// a regex match.
|
1.1
|
(pascal 03-Sep-16)
| 112
|
: StringRef FixedStr;
|
|
| 113
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 114
|
: /// A regex string to match as the pattern or empty if this pattern requires
|
|
| 115
|
: /// a fixed string to match.
|
1.1
|
(pascal 03-Sep-16)
| 116
|
: std::string RegExStr;
|
|
| 117
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 118
|
: /// Entries in this vector map to uses of a variable in the pattern, e.g.
|
|
| 119
|
: /// "foo[[bar]]baz". In this case, the RegExStr will contain "foobaz" and
|
|
| 120
|
: /// we'll get an entry in this vector that tells us to insert the value of
|
|
| 121
|
: /// bar at offset 3.
|
|
| 122
|
: std::vector<std::pair<StringRef, unsigned>> VariableUses;
|
|
| 123
|
:
|
|
| 124
|
: /// Maps definitions of variables to their parenthesized capture numbers.
|
|
| 125
|
: ///
|
|
| 126
|
: /// E.g. for the pattern "foo[[bar:.*]]baz", VariableDefs will map "bar" to
|
|
| 127
|
: /// 1.
|
1.1
|
(pascal 03-Sep-16)
| 128
|
: std::map<StringRef, unsigned> VariableDefs;
|
|
| 129
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 130
|
: Check::CheckType CheckTy;
|
1.1
|
(pascal 03-Sep-16)
| 131
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 132
|
: /// Contains the number of line this pattern is in.
|
|
| 133
|
: unsigned LineNumber;
|
|
| 134
|
:
|
|
| 135
|
: public:
|
|
| 136
|
: explicit Pattern(Check::CheckType Ty) : CheckTy(Ty) {}
|
1.1
|
(pascal 03-Sep-16)
| 137
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 138
|
: /// Returns the location in source code.
|
1.1
|
(pascal 03-Sep-16)
| 139
|
: SMLoc getLoc() const { return PatternLoc; }
|
|
| 140
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 141
|
: bool ParsePattern(StringRef PatternStr, StringRef Prefix, SourceMgr &SM,
|
1.1
|
(pascal 03-Sep-16)
| 142
|
: unsigned LineNumber);
|
|
| 143
|
: size_t Match(StringRef Buffer, size_t &MatchLen,
|
|
| 144
|
: StringMap<StringRef> &VariableTable) const;
|
|
| 145
|
: void PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
|
|
| 146
|
: const StringMap<StringRef> &VariableTable) const;
|
|
| 147
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 148
|
: bool hasVariable() const {
|
|
| 149
|
: return !(VariableUses.empty() && VariableDefs.empty());
|
|
| 150
|
: }
|
1.1
|
(pascal 03-Sep-16)
| 151
|
:
|
|
| 152
|
: Check::CheckType getCheckTy() const { return CheckTy; }
|
|
| 153
|
:
|
|
| 154
|
: private:
|
|
| 155
|
: bool AddRegExToRegEx(StringRef RS, unsigned &CurParen, SourceMgr &SM);
|
|
| 156
|
: void AddBackrefToRegEx(unsigned BackrefNum);
|
1.1.1.3
|
(patrick 24-Jan-17)
| 157
|
: unsigned
|
|
| 158
|
: ComputeMatchDistance(StringRef Buffer,
|
|
| 159
|
: const StringMap<StringRef> &VariableTable) const;
|
1.1
|
(pascal 03-Sep-16)
| 160
|
: bool EvaluateExpression(StringRef Expr, std::string &Value) const;
|
|
| 161
|
: size_t FindRegexVarEnd(StringRef Str, SourceMgr &SM);
|
|
| 162
|
: };
|
|
| 163
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 164
|
: /// Parses the given string into the Pattern.
|
|
| 165
|
: ///
|
|
| 166
|
: /// \p Prefix provides which prefix is being matched, \p SM provides the
|
|
| 167
|
: /// SourceMgr used for error reports, and \p LineNumber is the line number in
|
|
| 168
|
: /// the input file from which the pattern string was read. Returns true in
|
|
| 169
|
: /// case of an error, false otherwise.
|
|
| 170
|
: bool Pattern::ParsePattern(StringRef PatternStr, StringRef Prefix,
|
|
| 171
|
: SourceMgr &SM, unsigned LineNumber) {
|
1.1.1.2
|
(patrick 14-Jan-17)
| 172
|
: bool MatchFullLinesHere = MatchFullLines && CheckTy != Check::CheckNot;
|
|
| 173
|
:
|
1.1
|
(pascal 03-Sep-16)
| 174
|
: this->LineNumber = LineNumber;
|
|
| 175
|
: PatternLoc = SMLoc::getFromPointer(PatternStr.data());
|
|
| 176
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 177
|
: if (!(NoCanonicalizeWhiteSpace && MatchFullLines))
|
|
| 178
|
: // Ignore trailing whitespace.
|
|
| 179
|
: while (!PatternStr.empty() &&
|
|
| 180
|
: (PatternStr.back() == ' ' || PatternStr.back() == '\t'))
|
|
| 181
|
: PatternStr = PatternStr.substr(0, PatternStr.size() - 1);
|
1.1
|
(pascal 03-Sep-16)
| 182
|
:
|
|
| 183
|
: // Check that there is something on the line.
|
|
| 184
|
: if (PatternStr.empty()) {
|
|
| 185
|
: SM.PrintMessage(PatternLoc, SourceMgr::DK_Error,
|
1.1.1.3
|
(patrick 24-Jan-17)
| 186
|
: "found empty check string with prefix '" + Prefix + ":'");
|
1.1
|
(pascal 03-Sep-16)
| 187
|
: return true;
|
|
| 188
|
: }
|
|
| 189
|
:
|
|
| 190
|
: // Check to see if this is a fixed string, or if it has regex pieces.
|
1.1.1.2
|
(patrick 14-Jan-17)
| 191
|
: if (!MatchFullLinesHere &&
|
|
| 192
|
: (PatternStr.size() < 2 || (PatternStr.find("{{") == StringRef::npos &&
|
|
| 193
|
: PatternStr.find("[[") == StringRef::npos))) {
|
1.1
|
(pascal 03-Sep-16)
| 194
|
: FixedStr = PatternStr;
|
|
| 195
|
: return false;
|
|
| 196
|
: }
|
|
| 197
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 198
|
: if (MatchFullLinesHere) {
|
|
| 199
|
: RegExStr += '^';
|
|
| 200
|
: if (!NoCanonicalizeWhiteSpace)
|
|
| 201
|
: RegExStr += " *";
|
|
| 202
|
: }
|
|
| 203
|
:
|
1.1
|
(pascal 03-Sep-16)
| 204
|
: // Paren value #0 is for the fully matched string. Any new parenthesized
|
|
| 205
|
: // values add from there.
|
|
| 206
|
: unsigned CurParen = 1;
|
|
| 207
|
:
|
|
| 208
|
: // Otherwise, there is at least one regex piece. Build up the regex pattern
|
|
| 209
|
: // by escaping scary characters in fixed strings, building up one big regex.
|
|
| 210
|
: while (!PatternStr.empty()) {
|
|
| 211
|
: // RegEx matches.
|
|
| 212
|
: if (PatternStr.startswith("{{")) {
|
|
| 213
|
: // This is the start of a regex match. Scan for the }}.
|
|
| 214
|
: size_t End = PatternStr.find("}}");
|
|
| 215
|
: if (End == StringRef::npos) {
|
|
| 216
|
: SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
|
|
| 217
|
: SourceMgr::DK_Error,
|
|
| 218
|
: "found start of regex string with no end '}}'");
|
|
| 219
|
: return true;
|
|
| 220
|
: }
|
|
| 221
|
:
|
|
| 222
|
: // Enclose {{}} patterns in parens just like [[]] even though we're not
|
|
| 223
|
: // capturing the result for any purpose. This is required in case the
|
|
| 224
|
: // expression contains an alternation like: CHECK: abc{{x|z}}def. We
|
|
| 225
|
: // want this to turn into: "abc(x|z)def" not "abcx|zdef".
|
|
| 226
|
: RegExStr += '(';
|
|
| 227
|
: ++CurParen;
|
|
| 228
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 229
|
: if (AddRegExToRegEx(PatternStr.substr(2, End - 2), CurParen, SM))
|
1.1
|
(pascal 03-Sep-16)
| 230
|
: return true;
|
|
| 231
|
: RegExStr += ')';
|
|
| 232
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 233
|
: PatternStr = PatternStr.substr(End + 2);
|
1.1
|
(pascal 03-Sep-16)
| 234
|
: continue;
|
|
| 235
|
: }
|
|
| 236
|
:
|
|
| 237
|
: // Named RegEx matches. These are of two forms: [[foo:.*]] which matches .*
|
|
| 238
|
: // (or some other regex) and assigns it to the FileCheck variable 'foo'. The
|
|
| 239
|
: // second form is [[foo]] which is a reference to foo. The variable name
|
|
| 240
|
: // itself must be of the form "[a-zA-Z_][0-9a-zA-Z_]*", otherwise we reject
|
|
| 241
|
: // it. This is to catch some common errors.
|
|
| 242
|
: if (PatternStr.startswith("[[")) {
|
|
| 243
|
: // Find the closing bracket pair ending the match. End is going to be an
|
|
| 244
|
: // offset relative to the beginning of the match string.
|
|
| 245
|
: size_t End = FindRegexVarEnd(PatternStr.substr(2), SM);
|
|
| 246
|
:
|
|
| 247
|
: if (End == StringRef::npos) {
|
|
| 248
|
: SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
|
|
| 249
|
: SourceMgr::DK_Error,
|
|
| 250
|
: "invalid named regex reference, no ]] found");
|
|
| 251
|
: return true;
|
|
| 252
|
: }
|
|
| 253
|
:
|
|
| 254
|
: StringRef MatchStr = PatternStr.substr(2, End);
|
1.1.1.3
|
(patrick 24-Jan-17)
| 255
|
: PatternStr = PatternStr.substr(End + 4);
|
1.1
|
(pascal 03-Sep-16)
| 256
|
:
|
|
| 257
|
: // Get the regex name (e.g. "foo").
|
|
| 258
|
: size_t NameEnd = MatchStr.find(':');
|
|
| 259
|
: StringRef Name = MatchStr.substr(0, NameEnd);
|
|
| 260
|
:
|
|
| 261
|
: if (Name.empty()) {
|
|
| 262
|
: SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
|
|
| 263
|
: "invalid name in named regex: empty name");
|
|
| 264
|
: return true;
|
|
| 265
|
: }
|
|
| 266
|
:
|
|
| 267
|
: // Verify that the name/expression is well formed. FileCheck currently
|
|
| 268
|
: // supports @LINE, @LINE+number, @LINE-number expressions. The check here
|
|
| 269
|
: // is relaxed, more strict check is performed in \c EvaluateExpression.
|
|
| 270
|
: bool IsExpression = false;
|
|
| 271
|
: for (unsigned i = 0, e = Name.size(); i != e; ++i) {
|
1.1.1.4
|
(patrick 04-Oct-17)
| 272
|
: if (i == 0) {
|
|
| 273
|
: if (Name[i] == '$') // Global vars start with '$'
|
|
| 274
|
: continue;
|
|
| 275
|
: if (Name[i] == '@') {
|
|
| 276
|
: if (NameEnd != StringRef::npos) {
|
|
| 277
|
: SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
|
|
| 278
|
: SourceMgr::DK_Error,
|
|
| 279
|
: "invalid name in named regex definition");
|
|
| 280
|
: return true;
|
|
| 281
|
: }
|
|
| 282
|
: IsExpression = true;
|
|
| 283
|
: continue;
|
1.1
|
(pascal 03-Sep-16)
| 284
|
: }
|
|
| 285
|
: }
|
|
| 286
|
: if (Name[i] != '_' && !isalnum(Name[i]) &&
|
|
| 287
|
: (!IsExpression || (Name[i] != '+' && Name[i] != '-'))) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 288
|
: SM.PrintMessage(SMLoc::getFromPointer(Name.data() + i),
|
1.1
|
(pascal 03-Sep-16)
| 289
|
: SourceMgr::DK_Error, "invalid name in named regex");
|
|
| 290
|
: return true;
|
|
| 291
|
: }
|
|
| 292
|
: }
|
|
| 293
|
:
|
|
| 294
|
: // Name can't start with a digit.
|
|
| 295
|
: if (isdigit(static_cast<unsigned char>(Name[0]))) {
|
|
| 296
|
: SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
|
|
| 297
|
: "invalid name in named regex");
|
|
| 298
|
: return true;
|
|
| 299
|
: }
|
|
| 300
|
:
|
|
| 301
|
: // Handle [[foo]].
|
|
| 302
|
: if (NameEnd == StringRef::npos) {
|
|
| 303
|
: // Handle variables that were defined earlier on the same line by
|
|
| 304
|
: // emitting a backreference.
|
|
| 305
|
: if (VariableDefs.find(Name) != VariableDefs.end()) {
|
|
| 306
|
: unsigned VarParenNum = VariableDefs[Name];
|
|
| 307
|
: if (VarParenNum < 1 || VarParenNum > 9) {
|
|
| 308
|
: SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
|
|
| 309
|
: SourceMgr::DK_Error,
|
|
| 310
|
: "Can't back-reference more than 9 variables");
|
|
| 311
|
: return true;
|
|
| 312
|
: }
|
|
| 313
|
: AddBackrefToRegEx(VarParenNum);
|
|
| 314
|
: } else {
|
|
| 315
|
: VariableUses.push_back(std::make_pair(Name, RegExStr.size()));
|
|
| 316
|
: }
|
|
| 317
|
: continue;
|
|
| 318
|
: }
|
|
| 319
|
:
|
|
| 320
|
: // Handle [[foo:.*]].
|
|
| 321
|
: VariableDefs[Name] = CurParen;
|
|
| 322
|
: RegExStr += '(';
|
|
| 323
|
: ++CurParen;
|
|
| 324
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 325
|
: if (AddRegExToRegEx(MatchStr.substr(NameEnd + 1), CurParen, SM))
|
1.1
|
(pascal 03-Sep-16)
| 326
|
: return true;
|
|
| 327
|
:
|
|
| 328
|
: RegExStr += ')';
|
|
| 329
|
: }
|
|
| 330
|
:
|
|
| 331
|
: // Handle fixed string matches.
|
|
| 332
|
: // Find the end, which is the start of the next regex.
|
|
| 333
|
: size_t FixedMatchEnd = PatternStr.find("{{");
|
|
| 334
|
: FixedMatchEnd = std::min(FixedMatchEnd, PatternStr.find("[["));
|
|
| 335
|
: RegExStr += Regex::escape(PatternStr.substr(0, FixedMatchEnd));
|
|
| 336
|
: PatternStr = PatternStr.substr(FixedMatchEnd);
|
|
| 337
|
: }
|
|
| 338
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 339
|
: if (MatchFullLinesHere) {
|
|
| 340
|
: if (!NoCanonicalizeWhiteSpace)
|
|
| 341
|
: RegExStr += " *";
|
|
| 342
|
: RegExStr += '$';
|
|
| 343
|
: }
|
|
| 344
|
:
|
1.1
|
(pascal 03-Sep-16)
| 345
|
: return false;
|
|
| 346
|
: }
|
|
| 347
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 348
|
: bool Pattern::AddRegExToRegEx(StringRef RS, unsigned &CurParen, SourceMgr &SM) {
|
1.1
|
(pascal 03-Sep-16)
| 349
|
: Regex R(RS);
|
|
| 350
|
: std::string Error;
|
|
| 351
|
: if (!R.isValid(Error)) {
|
|
| 352
|
: SM.PrintMessage(SMLoc::getFromPointer(RS.data()), SourceMgr::DK_Error,
|
|
| 353
|
: "invalid regex: " + Error);
|
|
| 354
|
: return true;
|
|
| 355
|
: }
|
|
| 356
|
:
|
|
| 357
|
: RegExStr += RS.str();
|
|
| 358
|
: CurParen += R.getNumMatches();
|
|
| 359
|
: return false;
|
|
| 360
|
: }
|
|
| 361
|
:
|
|
| 362
|
: void Pattern::AddBackrefToRegEx(unsigned BackrefNum) {
|
|
| 363
|
: assert(BackrefNum >= 1 && BackrefNum <= 9 && "Invalid backref number");
|
1.1.1.3
|
(patrick 24-Jan-17)
| 364
|
: std::string Backref = std::string("\\") + std::string(1, '0' + BackrefNum);
|
1.1
|
(pascal 03-Sep-16)
| 365
|
: RegExStr += Backref;
|
|
| 366
|
: }
|
|
| 367
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 368
|
: /// Evaluates expression and stores the result to \p Value.
|
|
| 369
|
: ///
|
|
| 370
|
: /// Returns true on success and false when the expression has invalid syntax.
|
1.1
|
(pascal 03-Sep-16)
| 371
|
: bool Pattern::EvaluateExpression(StringRef Expr, std::string &Value) const {
|
|
| 372
|
: // The only supported expression is @LINE([\+-]\d+)?
|
|
| 373
|
: if (!Expr.startswith("@LINE"))
|
|
| 374
|
: return false;
|
|
| 375
|
: Expr = Expr.substr(StringRef("@LINE").size());
|
|
| 376
|
: int Offset = 0;
|
|
| 377
|
: if (!Expr.empty()) {
|
|
| 378
|
: if (Expr[0] == '+')
|
|
| 379
|
: Expr = Expr.substr(1);
|
|
| 380
|
: else if (Expr[0] != '-')
|
|
| 381
|
: return false;
|
|
| 382
|
: if (Expr.getAsInteger(10, Offset))
|
|
| 383
|
: return false;
|
|
| 384
|
: }
|
|
| 385
|
: Value = llvm::itostr(LineNumber + Offset);
|
|
| 386
|
: return true;
|
|
| 387
|
: }
|
|
| 388
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 389
|
: /// Matches the pattern string against the input buffer \p Buffer
|
|
| 390
|
: ///
|
|
| 391
|
: /// This returns the position that is matched or npos if there is no match. If
|
|
| 392
|
: /// there is a match, the size of the matched string is returned in \p
|
|
| 393
|
: /// MatchLen.
|
|
| 394
|
: ///
|
|
| 395
|
: /// The \p VariableTable StringMap provides the current values of filecheck
|
|
| 396
|
: /// variables and is updated if this match defines new values.
|
1.1
|
(pascal 03-Sep-16)
| 397
|
: size_t Pattern::Match(StringRef Buffer, size_t &MatchLen,
|
|
| 398
|
: StringMap<StringRef> &VariableTable) const {
|
|
| 399
|
: // If this is the EOF pattern, match it immediately.
|
|
| 400
|
: if (CheckTy == Check::CheckEOF) {
|
|
| 401
|
: MatchLen = 0;
|
|
| 402
|
: return Buffer.size();
|
|
| 403
|
: }
|
|
| 404
|
:
|
|
| 405
|
: // If this is a fixed string pattern, just match it now.
|
|
| 406
|
: if (!FixedStr.empty()) {
|
|
| 407
|
: MatchLen = FixedStr.size();
|
|
| 408
|
: return Buffer.find(FixedStr);
|
|
| 409
|
: }
|
|
| 410
|
:
|
|
| 411
|
: // Regex match.
|
|
| 412
|
:
|
|
| 413
|
: // If there are variable uses, we need to create a temporary string with the
|
|
| 414
|
: // actual value.
|
|
| 415
|
: StringRef RegExToMatch = RegExStr;
|
|
| 416
|
: std::string TmpStr;
|
|
| 417
|
: if (!VariableUses.empty()) {
|
|
| 418
|
: TmpStr = RegExStr;
|
|
| 419
|
:
|
|
| 420
|
: unsigned InsertOffset = 0;
|
|
| 421
|
: for (const auto &VariableUse : VariableUses) {
|
|
| 422
|
: std::string Value;
|
|
| 423
|
:
|
|
| 424
|
: if (VariableUse.first[0] == '@') {
|
|
| 425
|
: if (!EvaluateExpression(VariableUse.first, Value))
|
|
| 426
|
: return StringRef::npos;
|
|
| 427
|
: } else {
|
|
| 428
|
: StringMap<StringRef>::iterator it =
|
|
| 429
|
: VariableTable.find(VariableUse.first);
|
|
| 430
|
: // If the variable is undefined, return an error.
|
|
| 431
|
: if (it == VariableTable.end())
|
|
| 432
|
: return StringRef::npos;
|
|
| 433
|
:
|
|
| 434
|
: // Look up the value and escape it so that we can put it into the regex.
|
|
| 435
|
: Value += Regex::escape(it->second);
|
|
| 436
|
: }
|
|
| 437
|
:
|
|
| 438
|
: // Plop it into the regex at the adjusted offset.
|
|
| 439
|
: TmpStr.insert(TmpStr.begin() + VariableUse.second + InsertOffset,
|
|
| 440
|
: Value.begin(), Value.end());
|
|
| 441
|
: InsertOffset += Value.size();
|
|
| 442
|
: }
|
|
| 443
|
:
|
|
| 444
|
: // Match the newly constructed regex.
|
|
| 445
|
: RegExToMatch = TmpStr;
|
|
| 446
|
: }
|
|
| 447
|
:
|
|
| 448
|
: SmallVector<StringRef, 4> MatchInfo;
|
|
| 449
|
: if (!Regex(RegExToMatch, Regex::Newline).match(Buffer, &MatchInfo))
|
|
| 450
|
: return StringRef::npos;
|
|
| 451
|
:
|
|
| 452
|
: // Successful regex match.
|
|
| 453
|
: assert(!MatchInfo.empty() && "Didn't get any match");
|
|
| 454
|
: StringRef FullMatch = MatchInfo[0];
|
|
| 455
|
:
|
|
| 456
|
: // If this defines any variables, remember their values.
|
|
| 457
|
: for (const auto &VariableDef : VariableDefs) {
|
|
| 458
|
: assert(VariableDef.second < MatchInfo.size() && "Internal paren error");
|
|
| 459
|
: VariableTable[VariableDef.first] = MatchInfo[VariableDef.second];
|
|
| 460
|
: }
|
|
| 461
|
:
|
|
| 462
|
: MatchLen = FullMatch.size();
|
1.1.1.3
|
(patrick 24-Jan-17)
| 463
|
: return FullMatch.data() - Buffer.data();
|
1.1
|
(pascal 03-Sep-16)
| 464
|
: }
|
|
| 465
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 466
|
:
|
|
| 467
|
: /// Computes an arbitrary estimate for the quality of matching this pattern at
|
|
| 468
|
: /// the start of \p Buffer; a distance of zero should correspond to a perfect
|
|
| 469
|
: /// match.
|
|
| 470
|
: unsigned
|
|
| 471
|
: Pattern::ComputeMatchDistance(StringRef Buffer,
|
1.1
|
(pascal 03-Sep-16)
| 472
|
: const StringMap<StringRef> &VariableTable) const {
|
|
| 473
|
: // Just compute the number of matching characters. For regular expressions, we
|
|
| 474
|
: // just compare against the regex itself and hope for the best.
|
|
| 475
|
: //
|
|
| 476
|
: // FIXME: One easy improvement here is have the regex lib generate a single
|
|
| 477
|
: // example regular expression which matches, and use that as the example
|
|
| 478
|
: // string.
|
|
| 479
|
: StringRef ExampleString(FixedStr);
|
|
| 480
|
: if (ExampleString.empty())
|
|
| 481
|
: ExampleString = RegExStr;
|
|
| 482
|
:
|
|
| 483
|
: // Only compare up to the first line in the buffer, or the string size.
|
|
| 484
|
: StringRef BufferPrefix = Buffer.substr(0, ExampleString.size());
|
|
| 485
|
: BufferPrefix = BufferPrefix.split('\n').first;
|
|
| 486
|
: return BufferPrefix.edit_distance(ExampleString);
|
|
| 487
|
: }
|
|
| 488
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 489
|
: /// Prints additional information about a failure to match involving this
|
|
| 490
|
: /// pattern.
|
|
| 491
|
: void Pattern::PrintFailureInfo(
|
|
| 492
|
: const SourceMgr &SM, StringRef Buffer,
|
|
| 493
|
: const StringMap<StringRef> &VariableTable) const {
|
1.1
|
(pascal 03-Sep-16)
| 494
|
: // If this was a regular expression using variables, print the current
|
|
| 495
|
: // variable values.
|
|
| 496
|
: if (!VariableUses.empty()) {
|
|
| 497
|
: for (const auto &VariableUse : VariableUses) {
|
|
| 498
|
: SmallString<256> Msg;
|
|
| 499
|
: raw_svector_ostream OS(Msg);
|
|
| 500
|
: StringRef Var = VariableUse.first;
|
|
| 501
|
: if (Var[0] == '@') {
|
|
| 502
|
: std::string Value;
|
|
| 503
|
: if (EvaluateExpression(Var, Value)) {
|
|
| 504
|
: OS << "with expression \"";
|
|
| 505
|
: OS.write_escaped(Var) << "\" equal to \"";
|
|
| 506
|
: OS.write_escaped(Value) << "\"";
|
|
| 507
|
: } else {
|
|
| 508
|
: OS << "uses incorrect expression \"";
|
|
| 509
|
: OS.write_escaped(Var) << "\"";
|
|
| 510
|
: }
|
|
| 511
|
: } else {
|
|
| 512
|
: StringMap<StringRef>::const_iterator it = VariableTable.find(Var);
|
|
| 513
|
:
|
|
| 514
|
: // Check for undefined variable references.
|
|
| 515
|
: if (it == VariableTable.end()) {
|
|
| 516
|
: OS << "uses undefined variable \"";
|
|
| 517
|
: OS.write_escaped(Var) << "\"";
|
|
| 518
|
: } else {
|
|
| 519
|
: OS << "with variable \"";
|
|
| 520
|
: OS.write_escaped(Var) << "\" equal to \"";
|
|
| 521
|
: OS.write_escaped(it->second) << "\"";
|
|
| 522
|
: }
|
|
| 523
|
: }
|
|
| 524
|
:
|
|
| 525
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
|
|
| 526
|
: OS.str());
|
|
| 527
|
: }
|
|
| 528
|
: }
|
|
| 529
|
:
|
|
| 530
|
: // Attempt to find the closest/best fuzzy match. Usually an error happens
|
|
| 531
|
: // because some string in the output didn't exactly match. In these cases, we
|
|
| 532
|
: // would like to show the user a best guess at what "should have" matched, to
|
|
| 533
|
: // save them having to actually check the input manually.
|
|
| 534
|
: size_t NumLinesForward = 0;
|
|
| 535
|
: size_t Best = StringRef::npos;
|
|
| 536
|
: double BestQuality = 0;
|
|
| 537
|
:
|
|
| 538
|
: // Use an arbitrary 4k limit on how far we will search.
|
|
| 539
|
: for (size_t i = 0, e = std::min(size_t(4096), Buffer.size()); i != e; ++i) {
|
|
| 540
|
: if (Buffer[i] == '\n')
|
|
| 541
|
: ++NumLinesForward;
|
|
| 542
|
:
|
|
| 543
|
: // Patterns have leading whitespace stripped, so skip whitespace when
|
|
| 544
|
: // looking for something which looks like a pattern.
|
|
| 545
|
: if (Buffer[i] == ' ' || Buffer[i] == '\t')
|
|
| 546
|
: continue;
|
|
| 547
|
:
|
|
| 548
|
: // Compute the "quality" of this match as an arbitrary combination of the
|
|
| 549
|
: // match distance and the number of lines skipped to get to this match.
|
|
| 550
|
: unsigned Distance = ComputeMatchDistance(Buffer.substr(i), VariableTable);
|
|
| 551
|
: double Quality = Distance + (NumLinesForward / 100.);
|
|
| 552
|
:
|
|
| 553
|
: if (Quality < BestQuality || Best == StringRef::npos) {
|
|
| 554
|
: Best = i;
|
|
| 555
|
: BestQuality = Quality;
|
|
| 556
|
: }
|
|
| 557
|
: }
|
|
| 558
|
:
|
|
| 559
|
: // Print the "possible intended match here" line if we found something
|
|
| 560
|
: // reasonable and not equal to what we showed in the "scanning from here"
|
|
| 561
|
: // line.
|
|
| 562
|
: if (Best && Best != StringRef::npos && BestQuality < 50) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 563
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + Best),
|
|
| 564
|
: SourceMgr::DK_Note, "possible intended match here");
|
1.1
|
(pascal 03-Sep-16)
| 565
|
:
|
|
| 566
|
: // FIXME: If we wanted to be really friendly we would show why the match
|
|
| 567
|
: // failed, as it can be hard to spot simple one character differences.
|
|
| 568
|
: }
|
|
| 569
|
: }
|
|
| 570
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 571
|
: /// Finds the closing sequence of a regex variable usage or definition.
|
|
| 572
|
: ///
|
|
| 573
|
: /// \p Str has to point in the beginning of the definition (right after the
|
|
| 574
|
: /// opening sequence). Returns the offset of the closing sequence within Str,
|
|
| 575
|
: /// or npos if it was not found.
|
1.1
|
(pascal 03-Sep-16)
| 576
|
: size_t Pattern::FindRegexVarEnd(StringRef Str, SourceMgr &SM) {
|
|
| 577
|
: // Offset keeps track of the current offset within the input Str
|
|
| 578
|
: size_t Offset = 0;
|
|
| 579
|
: // [...] Nesting depth
|
|
| 580
|
: size_t BracketDepth = 0;
|
|
| 581
|
:
|
|
| 582
|
: while (!Str.empty()) {
|
|
| 583
|
: if (Str.startswith("]]") && BracketDepth == 0)
|
|
| 584
|
: return Offset;
|
|
| 585
|
: if (Str[0] == '\\') {
|
|
| 586
|
: // Backslash escapes the next char within regexes, so skip them both.
|
|
| 587
|
: Str = Str.substr(2);
|
|
| 588
|
: Offset += 2;
|
|
| 589
|
: } else {
|
|
| 590
|
: switch (Str[0]) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 591
|
: default:
|
|
| 592
|
: break;
|
|
| 593
|
: case '[':
|
|
| 594
|
: BracketDepth++;
|
|
| 595
|
: break;
|
|
| 596
|
: case ']':
|
|
| 597
|
: if (BracketDepth == 0) {
|
|
| 598
|
: SM.PrintMessage(SMLoc::getFromPointer(Str.data()),
|
|
| 599
|
: SourceMgr::DK_Error,
|
|
| 600
|
: "missing closing \"]\" for regex variable");
|
|
| 601
|
: exit(1);
|
|
| 602
|
: }
|
|
| 603
|
: BracketDepth--;
|
|
| 604
|
: break;
|
1.1
|
(pascal 03-Sep-16)
| 605
|
: }
|
|
| 606
|
: Str = Str.substr(1);
|
|
| 607
|
: Offset++;
|
|
| 608
|
: }
|
|
| 609
|
: }
|
|
| 610
|
:
|
|
| 611
|
: return StringRef::npos;
|
|
| 612
|
: }
|
|
| 613
|
:
|
|
| 614
|
: //===----------------------------------------------------------------------===//
|
|
| 615
|
: // Check Strings.
|
|
| 616
|
: //===----------------------------------------------------------------------===//
|
|
| 617
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 618
|
: /// A check that we found in the input file.
|
1.1
|
(pascal 03-Sep-16)
| 619
|
: struct CheckString {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 620
|
: /// The pattern to match.
|
1.1
|
(pascal 03-Sep-16)
| 621
|
: Pattern Pat;
|
|
| 622
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 623
|
: /// Which prefix name this check matched.
|
1.1
|
(pascal 03-Sep-16)
| 624
|
: StringRef Prefix;
|
|
| 625
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 626
|
: /// The location in the match file that the check string was specified.
|
1.1
|
(pascal 03-Sep-16)
| 627
|
: SMLoc Loc;
|
|
| 628
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 629
|
: /// All of the strings that are disallowed from occurring between this match
|
|
| 630
|
: /// string and the previous one (or start of file).
|
1.1
|
(pascal 03-Sep-16)
| 631
|
: std::vector<Pattern> DagNotStrings;
|
|
| 632
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 633
|
: CheckString(const Pattern &P, StringRef S, SMLoc L)
|
|
| 634
|
: : Pat(P), Prefix(S), Loc(L) {}
|
1.1
|
(pascal 03-Sep-16)
| 635
|
:
|
|
| 636
|
: size_t Check(const SourceMgr &SM, StringRef Buffer, bool IsLabelScanMode,
|
|
| 637
|
: size_t &MatchLen, StringMap<StringRef> &VariableTable) const;
|
|
| 638
|
:
|
|
| 639
|
: bool CheckNext(const SourceMgr &SM, StringRef Buffer) const;
|
|
| 640
|
: bool CheckSame(const SourceMgr &SM, StringRef Buffer) const;
|
|
| 641
|
: bool CheckNot(const SourceMgr &SM, StringRef Buffer,
|
|
| 642
|
: const std::vector<const Pattern *> &NotStrings,
|
|
| 643
|
: StringMap<StringRef> &VariableTable) const;
|
|
| 644
|
: size_t CheckDag(const SourceMgr &SM, StringRef Buffer,
|
|
| 645
|
: std::vector<const Pattern *> &NotStrings,
|
|
| 646
|
: StringMap<StringRef> &VariableTable) const;
|
|
| 647
|
: };
|
|
| 648
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 649
|
: /// Canonicalize whitespaces in the file. Line endings are replaced with
|
|
| 650
|
: /// UNIX-style '\n'.
|
|
| 651
|
: static StringRef CanonicalizeFile(MemoryBuffer &MB,
|
|
| 652
|
: SmallVectorImpl<char> &OutputBuffer) {
|
|
| 653
|
: OutputBuffer.reserve(MB.getBufferSize());
|
1.1
|
(pascal 03-Sep-16)
| 654
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 655
|
: for (const char *Ptr = MB.getBufferStart(), *End = MB.getBufferEnd();
|
1.1
|
(pascal 03-Sep-16)
| 656
|
: Ptr != End; ++Ptr) {
|
|
| 657
|
: // Eliminate trailing dosish \r.
|
|
| 658
|
: if (Ptr <= End - 2 && Ptr[0] == '\r' && Ptr[1] == '\n') {
|
|
| 659
|
: continue;
|
|
| 660
|
: }
|
|
| 661
|
:
|
|
| 662
|
: // If current char is not a horizontal whitespace or if horizontal
|
|
| 663
|
: // whitespace canonicalization is disabled, dump it to output as is.
|
1.1.1.3
|
(patrick 24-Jan-17)
| 664
|
: if (NoCanonicalizeWhiteSpace || (*Ptr != ' ' && *Ptr != '\t')) {
|
|
| 665
|
: OutputBuffer.push_back(*Ptr);
|
1.1
|
(pascal 03-Sep-16)
| 666
|
: continue;
|
|
| 667
|
: }
|
|
| 668
|
:
|
|
| 669
|
: // Otherwise, add one space and advance over neighboring space.
|
1.1.1.3
|
(patrick 24-Jan-17)
| 670
|
: OutputBuffer.push_back(' ');
|
|
| 671
|
: while (Ptr + 1 != End && (Ptr[1] == ' ' || Ptr[1] == '\t'))
|
1.1
|
(pascal 03-Sep-16)
| 672
|
: ++Ptr;
|
|
| 673
|
: }
|
|
| 674
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 675
|
: // Add a null byte and then return all but that byte.
|
|
| 676
|
: OutputBuffer.push_back('\0');
|
|
| 677
|
: return StringRef(OutputBuffer.data(), OutputBuffer.size() - 1);
|
1.1
|
(pascal 03-Sep-16)
| 678
|
: }
|
|
| 679
|
:
|
|
| 680
|
: static bool IsPartOfWord(char c) {
|
|
| 681
|
: return (isalnum(c) || c == '-' || c == '_');
|
|
| 682
|
: }
|
|
| 683
|
:
|
|
| 684
|
: // Get the size of the prefix extension.
|
|
| 685
|
: static size_t CheckTypeSize(Check::CheckType Ty) {
|
|
| 686
|
: switch (Ty) {
|
|
| 687
|
: case Check::CheckNone:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 688
|
: case Check::CheckBadNot:
|
1.1
|
(pascal 03-Sep-16)
| 689
|
: return 0;
|
|
| 690
|
:
|
|
| 691
|
: case Check::CheckPlain:
|
|
| 692
|
: return sizeof(":") - 1;
|
|
| 693
|
:
|
|
| 694
|
: case Check::CheckNext:
|
|
| 695
|
: return sizeof("-NEXT:") - 1;
|
|
| 696
|
:
|
|
| 697
|
: case Check::CheckSame:
|
|
| 698
|
: return sizeof("-SAME:") - 1;
|
|
| 699
|
:
|
|
| 700
|
: case Check::CheckNot:
|
|
| 701
|
: return sizeof("-NOT:") - 1;
|
|
| 702
|
:
|
|
| 703
|
: case Check::CheckDAG:
|
|
| 704
|
: return sizeof("-DAG:") - 1;
|
|
| 705
|
:
|
|
| 706
|
: case Check::CheckLabel:
|
|
| 707
|
: return sizeof("-LABEL:") - 1;
|
|
| 708
|
:
|
|
| 709
|
: case Check::CheckEOF:
|
|
| 710
|
: llvm_unreachable("Should not be using EOF size");
|
|
| 711
|
: }
|
|
| 712
|
:
|
|
| 713
|
: llvm_unreachable("Bad check type");
|
|
| 714
|
: }
|
|
| 715
|
:
|
|
| 716
|
: static Check::CheckType FindCheckType(StringRef Buffer, StringRef Prefix) {
|
|
| 717
|
: char NextChar = Buffer[Prefix.size()];
|
|
| 718
|
:
|
|
| 719
|
: // Verify that the : is present after the prefix.
|
|
| 720
|
: if (NextChar == ':')
|
|
| 721
|
: return Check::CheckPlain;
|
|
| 722
|
:
|
|
| 723
|
: if (NextChar != '-')
|
|
| 724
|
: return Check::CheckNone;
|
|
| 725
|
:
|
|
| 726
|
: StringRef Rest = Buffer.drop_front(Prefix.size() + 1);
|
|
| 727
|
: if (Rest.startswith("NEXT:"))
|
|
| 728
|
: return Check::CheckNext;
|
|
| 729
|
:
|
|
| 730
|
: if (Rest.startswith("SAME:"))
|
|
| 731
|
: return Check::CheckSame;
|
|
| 732
|
:
|
|
| 733
|
: if (Rest.startswith("NOT:"))
|
|
| 734
|
: return Check::CheckNot;
|
|
| 735
|
:
|
|
| 736
|
: if (Rest.startswith("DAG:"))
|
|
| 737
|
: return Check::CheckDAG;
|
|
| 738
|
:
|
|
| 739
|
: if (Rest.startswith("LABEL:"))
|
|
| 740
|
: return Check::CheckLabel;
|
|
| 741
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 742
|
: // You can't combine -NOT with another suffix.
|
|
| 743
|
: if (Rest.startswith("DAG-NOT:") || Rest.startswith("NOT-DAG:") ||
|
|
| 744
|
: Rest.startswith("NEXT-NOT:") || Rest.startswith("NOT-NEXT:") ||
|
|
| 745
|
: Rest.startswith("SAME-NOT:") || Rest.startswith("NOT-SAME:"))
|
|
| 746
|
: return Check::CheckBadNot;
|
|
| 747
|
:
|
1.1
|
(pascal 03-Sep-16)
| 748
|
: return Check::CheckNone;
|
|
| 749
|
: }
|
|
| 750
|
:
|
|
| 751
|
: // From the given position, find the next character after the word.
|
|
| 752
|
: static size_t SkipWord(StringRef Str, size_t Loc) {
|
|
| 753
|
: while (Loc < Str.size() && IsPartOfWord(Str[Loc]))
|
|
| 754
|
: ++Loc;
|
|
| 755
|
: return Loc;
|
|
| 756
|
: }
|
|
| 757
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 758
|
: /// Search the buffer for the first prefix in the prefix regular expression.
|
|
| 759
|
: ///
|
|
| 760
|
: /// This searches the buffer using the provided regular expression, however it
|
|
| 761
|
: /// enforces constraints beyond that:
|
|
| 762
|
: /// 1) The found prefix must not be a suffix of something that looks like
|
|
| 763
|
: /// a valid prefix.
|
|
| 764
|
: /// 2) The found prefix must be followed by a valid check type suffix using \c
|
|
| 765
|
: /// FindCheckType above.
|
|
| 766
|
: ///
|
|
| 767
|
: /// The first match of the regular expression to satisfy these two is returned,
|
|
| 768
|
: /// otherwise an empty StringRef is returned to indicate failure.
|
|
| 769
|
: ///
|
|
| 770
|
: /// If this routine returns a valid prefix, it will also shrink \p Buffer to
|
|
| 771
|
: /// start at the beginning of the returned prefix, increment \p LineNumber for
|
|
| 772
|
: /// each new line consumed from \p Buffer, and set \p CheckTy to the type of
|
|
| 773
|
: /// check found by examining the suffix.
|
|
| 774
|
: ///
|
|
| 775
|
: /// If no valid prefix is found, the state of Buffer, LineNumber, and CheckTy
|
|
| 776
|
: /// is unspecified.
|
|
| 777
|
: static StringRef FindFirstMatchingPrefix(Regex &PrefixRE, StringRef &Buffer,
|
1.1
|
(pascal 03-Sep-16)
| 778
|
: unsigned &LineNumber,
|
1.1.1.3
|
(patrick 24-Jan-17)
| 779
|
: Check::CheckType &CheckTy) {
|
|
| 780
|
: SmallVector<StringRef, 2> Matches;
|
1.1
|
(pascal 03-Sep-16)
| 781
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 782
|
: while (!Buffer.empty()) {
|
|
| 783
|
: // Find the first (longest) match using the RE.
|
|
| 784
|
: if (!PrefixRE.match(Buffer, &Matches))
|
|
| 785
|
: // No match at all, bail.
|
1.1
|
(pascal 03-Sep-16)
| 786
|
: return StringRef();
|
|
| 787
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 788
|
: StringRef Prefix = Matches[0];
|
|
| 789
|
: Matches.clear();
|
1.1
|
(pascal 03-Sep-16)
| 790
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 791
|
: assert(Prefix.data() >= Buffer.data() &&
|
|
| 792
|
: Prefix.data() < Buffer.data() + Buffer.size() &&
|
|
| 793
|
: "Prefix doesn't start inside of buffer!");
|
|
| 794
|
: size_t Loc = Prefix.data() - Buffer.data();
|
|
| 795
|
: StringRef Skipped = Buffer.substr(0, Loc);
|
|
| 796
|
: Buffer = Buffer.drop_front(Loc);
|
|
| 797
|
: LineNumber += Skipped.count('\n');
|
|
| 798
|
:
|
|
| 799
|
: // Check that the matched prefix isn't a suffix of some other check-like
|
|
| 800
|
: // word.
|
|
| 801
|
: // FIXME: This is a very ad-hoc check. it would be better handled in some
|
|
| 802
|
: // other way. Among other things it seems hard to distinguish between
|
|
| 803
|
: // intentional and unintentional uses of this feature.
|
|
| 804
|
: if (Skipped.empty() || !IsPartOfWord(Skipped.back())) {
|
|
| 805
|
: // Now extract the type.
|
|
| 806
|
: CheckTy = FindCheckType(Buffer, Prefix);
|
|
| 807
|
:
|
|
| 808
|
: // If we've found a valid check type for this prefix, we're done.
|
|
| 809
|
: if (CheckTy != Check::CheckNone)
|
|
| 810
|
: return Prefix;
|
|
| 811
|
: }
|
|
| 812
|
:
|
|
| 813
|
: // If we didn't successfully find a prefix, we need to skip this invalid
|
|
| 814
|
: // prefix and continue scanning. We directly skip the prefix that was
|
|
| 815
|
: // matched and any additional parts of that check-like word.
|
|
| 816
|
: Buffer = Buffer.drop_front(SkipWord(Buffer, Prefix.size()));
|
1.1
|
(pascal 03-Sep-16)
| 817
|
: }
|
|
| 818
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 819
|
: // We ran out of buffer while skipping partial matches so give up.
|
1.1
|
(pascal 03-Sep-16)
| 820
|
: return StringRef();
|
|
| 821
|
: }
|
|
| 822
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 823
|
: /// Read the check file, which specifies the sequence of expected strings.
|
|
| 824
|
: ///
|
|
| 825
|
: /// The strings are added to the CheckStrings vector. Returns true in case of
|
|
| 826
|
: /// an error, false otherwise.
|
|
| 827
|
: static bool ReadCheckFile(SourceMgr &SM, StringRef Buffer, Regex &PrefixRE,
|
1.1
|
(pascal 03-Sep-16)
| 828
|
: std::vector<CheckString> &CheckStrings) {
|
|
| 829
|
: std::vector<Pattern> ImplicitNegativeChecks;
|
|
| 830
|
: for (const auto &PatternString : ImplicitCheckNot) {
|
|
| 831
|
: // Create a buffer with fake command line content in order to display the
|
|
| 832
|
: // command line option responsible for the specific implicit CHECK-NOT.
|
|
| 833
|
: std::string Prefix = (Twine("-") + ImplicitCheckNot.ArgStr + "='").str();
|
|
| 834
|
: std::string Suffix = "'";
|
|
| 835
|
: std::unique_ptr<MemoryBuffer> CmdLine = MemoryBuffer::getMemBufferCopy(
|
|
| 836
|
: Prefix + PatternString + Suffix, "command line");
|
|
| 837
|
:
|
|
| 838
|
: StringRef PatternInBuffer =
|
|
| 839
|
: CmdLine->getBuffer().substr(Prefix.size(), PatternString.size());
|
|
| 840
|
: SM.AddNewSourceBuffer(std::move(CmdLine), SMLoc());
|
|
| 841
|
:
|
|
| 842
|
: ImplicitNegativeChecks.push_back(Pattern(Check::CheckNot));
|
|
| 843
|
: ImplicitNegativeChecks.back().ParsePattern(PatternInBuffer,
|
|
| 844
|
: "IMPLICIT-CHECK", SM, 0);
|
|
| 845
|
: }
|
|
| 846
|
:
|
|
| 847
|
: std::vector<Pattern> DagNotMatches = ImplicitNegativeChecks;
|
|
| 848
|
:
|
|
| 849
|
: // LineNumber keeps track of the line on which CheckPrefix instances are
|
|
| 850
|
: // found.
|
|
| 851
|
: unsigned LineNumber = 1;
|
|
| 852
|
:
|
|
| 853
|
: while (1) {
|
|
| 854
|
: Check::CheckType CheckTy;
|
|
| 855
|
:
|
|
| 856
|
: // See if a prefix occurs in the memory buffer.
|
1.1.1.3
|
(patrick 24-Jan-17)
| 857
|
: StringRef UsedPrefix = FindFirstMatchingPrefix(PrefixRE, Buffer, LineNumber,
|
|
| 858
|
: CheckTy);
|
1.1
|
(pascal 03-Sep-16)
| 859
|
: if (UsedPrefix.empty())
|
|
| 860
|
: break;
|
1.1.1.3
|
(patrick 24-Jan-17)
| 861
|
: assert(UsedPrefix.data() == Buffer.data() &&
|
|
| 862
|
: "Failed to move Buffer's start forward, or pointed prefix outside "
|
|
| 863
|
: "of the buffer!");
|
1.1
|
(pascal 03-Sep-16)
| 864
|
:
|
|
| 865
|
: // Location to use for error messages.
|
1.1.1.3
|
(patrick 24-Jan-17)
| 866
|
: const char *UsedPrefixStart = UsedPrefix.data();
|
1.1
|
(pascal 03-Sep-16)
| 867
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 868
|
: // Skip the buffer to the end.
|
1.1
|
(pascal 03-Sep-16)
| 869
|
: Buffer = Buffer.drop_front(UsedPrefix.size() + CheckTypeSize(CheckTy));
|
|
| 870
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 871
|
: // Complain about useful-looking but unsupported suffixes.
|
|
| 872
|
: if (CheckTy == Check::CheckBadNot) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 873
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Error,
|
1.1.1.2
|
(patrick 14-Jan-17)
| 874
|
: "unsupported -NOT combo on prefix '" + UsedPrefix + "'");
|
|
| 875
|
: return true;
|
|
| 876
|
: }
|
|
| 877
|
:
|
1.1
|
(pascal 03-Sep-16)
| 878
|
: // Okay, we found the prefix, yay. Remember the rest of the line, but ignore
|
1.1.1.3
|
(patrick 24-Jan-17)
| 879
|
: // leading whitespace.
|
|
| 880
|
: if (!(NoCanonicalizeWhiteSpace && MatchFullLines))
|
|
| 881
|
: Buffer = Buffer.substr(Buffer.find_first_not_of(" \t"));
|
1.1
|
(pascal 03-Sep-16)
| 882
|
:
|
|
| 883
|
: // Scan ahead to the end of line.
|
|
| 884
|
: size_t EOL = Buffer.find_first_of("\n\r");
|
|
| 885
|
:
|
|
| 886
|
: // Remember the location of the start of the pattern, for diagnostics.
|
|
| 887
|
: SMLoc PatternLoc = SMLoc::getFromPointer(Buffer.data());
|
|
| 888
|
:
|
|
| 889
|
: // Parse the pattern.
|
|
| 890
|
: Pattern P(CheckTy);
|
|
| 891
|
: if (P.ParsePattern(Buffer.substr(0, EOL), UsedPrefix, SM, LineNumber))
|
|
| 892
|
: return true;
|
|
| 893
|
:
|
|
| 894
|
: // Verify that CHECK-LABEL lines do not define or use variables
|
|
| 895
|
: if ((CheckTy == Check::CheckLabel) && P.hasVariable()) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 896
|
: SM.PrintMessage(
|
|
| 897
|
: SMLoc::getFromPointer(UsedPrefixStart), SourceMgr::DK_Error,
|
|
| 898
|
: "found '" + UsedPrefix + "-LABEL:'"
|
|
| 899
|
: " with variable definition or use");
|
1.1
|
(pascal 03-Sep-16)
| 900
|
: return true;
|
|
| 901
|
: }
|
|
| 902
|
:
|
|
| 903
|
: Buffer = Buffer.substr(EOL);
|
|
| 904
|
:
|
|
| 905
|
: // Verify that CHECK-NEXT lines have at least one CHECK line before them.
|
|
| 906
|
: if ((CheckTy == Check::CheckNext || CheckTy == Check::CheckSame) &&
|
|
| 907
|
: CheckStrings.empty()) {
|
|
| 908
|
: StringRef Type = CheckTy == Check::CheckNext ? "NEXT" : "SAME";
|
|
| 909
|
: SM.PrintMessage(SMLoc::getFromPointer(UsedPrefixStart),
|
|
| 910
|
: SourceMgr::DK_Error,
|
1.1.1.3
|
(patrick 24-Jan-17)
| 911
|
: "found '" + UsedPrefix + "-" + Type +
|
|
| 912
|
: "' without previous '" + UsedPrefix + ": line");
|
1.1
|
(pascal 03-Sep-16)
| 913
|
: return true;
|
|
| 914
|
: }
|
|
| 915
|
:
|
|
| 916
|
: // Handle CHECK-DAG/-NOT.
|
|
| 917
|
: if (CheckTy == Check::CheckDAG || CheckTy == Check::CheckNot) {
|
|
| 918
|
: DagNotMatches.push_back(P);
|
|
| 919
|
: continue;
|
|
| 920
|
: }
|
|
| 921
|
:
|
|
| 922
|
: // Okay, add the string we captured to the output vector and move on.
|
1.1.1.2
|
(patrick 14-Jan-17)
| 923
|
: CheckStrings.emplace_back(P, UsedPrefix, PatternLoc);
|
1.1
|
(pascal 03-Sep-16)
| 924
|
: std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
|
|
| 925
|
: DagNotMatches = ImplicitNegativeChecks;
|
|
| 926
|
: }
|
|
| 927
|
:
|
|
| 928
|
: // Add an EOF pattern for any trailing CHECK-DAG/-NOTs, and use the first
|
|
| 929
|
: // prefix as a filler for the error message.
|
|
| 930
|
: if (!DagNotMatches.empty()) {
|
|
| 931
|
: CheckStrings.emplace_back(Pattern(Check::CheckEOF), *CheckPrefixes.begin(),
|
1.1.1.2
|
(patrick 14-Jan-17)
| 932
|
: SMLoc::getFromPointer(Buffer.data()));
|
1.1
|
(pascal 03-Sep-16)
| 933
|
: std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
|
|
| 934
|
: }
|
|
| 935
|
:
|
|
| 936
|
: if (CheckStrings.empty()) {
|
|
| 937
|
: errs() << "error: no check strings found with prefix"
|
|
| 938
|
: << (CheckPrefixes.size() > 1 ? "es " : " ");
|
|
| 939
|
: prefix_iterator I = CheckPrefixes.begin();
|
|
| 940
|
: prefix_iterator E = CheckPrefixes.end();
|
|
| 941
|
: if (I != E) {
|
|
| 942
|
: errs() << "\'" << *I << ":'";
|
|
| 943
|
: ++I;
|
|
| 944
|
: }
|
1.1.1.2
|
(patrick 14-Jan-17)
| 945
|
: for (; I != E; ++I)
|
1.1
|
(pascal 03-Sep-16)
| 946
|
: errs() << ", \'" << *I << ":'";
|
|
| 947
|
:
|
|
| 948
|
: errs() << '\n';
|
|
| 949
|
: return true;
|
|
| 950
|
: }
|
|
| 951
|
:
|
|
| 952
|
: return false;
|
|
| 953
|
: }
|
|
| 954
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 955
|
: static void PrintCheckFailed(const SourceMgr &SM, SMLoc Loc, const Pattern &Pat,
|
|
| 956
|
: StringRef Buffer,
|
1.1
|
(pascal 03-Sep-16)
| 957
|
: StringMap<StringRef> &VariableTable) {
|
|
| 958
|
: // Otherwise, we have an error, emit an error message.
|
|
| 959
|
: SM.PrintMessage(Loc, SourceMgr::DK_Error,
|
|
| 960
|
: "expected string not found in input");
|
|
| 961
|
:
|
|
| 962
|
: // Print the "scanning from here" line. If the current position is at the
|
|
| 963
|
: // end of a line, advance to the start of the next line.
|
|
| 964
|
: Buffer = Buffer.substr(Buffer.find_first_not_of(" \t\n\r"));
|
|
| 965
|
:
|
|
| 966
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
|
|
| 967
|
: "scanning from here");
|
|
| 968
|
:
|
|
| 969
|
: // Allow the pattern to print additional information if desired.
|
|
| 970
|
: Pat.PrintFailureInfo(SM, Buffer, VariableTable);
|
|
| 971
|
: }
|
|
| 972
|
:
|
|
| 973
|
: static void PrintCheckFailed(const SourceMgr &SM, const CheckString &CheckStr,
|
|
| 974
|
: StringRef Buffer,
|
|
| 975
|
: StringMap<StringRef> &VariableTable) {
|
|
| 976
|
: PrintCheckFailed(SM, CheckStr.Loc, CheckStr.Pat, Buffer, VariableTable);
|
|
| 977
|
: }
|
|
| 978
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 979
|
: /// Count the number of newlines in the specified range.
|
1.1
|
(pascal 03-Sep-16)
| 980
|
: static unsigned CountNumNewlinesBetween(StringRef Range,
|
|
| 981
|
: const char *&FirstNewLine) {
|
|
| 982
|
: unsigned NumNewLines = 0;
|
|
| 983
|
: while (1) {
|
|
| 984
|
: // Scan for newline.
|
|
| 985
|
: Range = Range.substr(Range.find_first_of("\n\r"));
|
1.1.1.3
|
(patrick 24-Jan-17)
| 986
|
: if (Range.empty())
|
|
| 987
|
: return NumNewLines;
|
1.1
|
(pascal 03-Sep-16)
| 988
|
:
|
|
| 989
|
: ++NumNewLines;
|
|
| 990
|
:
|
|
| 991
|
: // Handle \n\r and \r\n as a single newline.
|
1.1.1.3
|
(patrick 24-Jan-17)
| 992
|
: if (Range.size() > 1 && (Range[1] == '\n' || Range[1] == '\r') &&
|
1.1
|
(pascal 03-Sep-16)
| 993
|
: (Range[0] != Range[1]))
|
|
| 994
|
: Range = Range.substr(1);
|
|
| 995
|
: Range = Range.substr(1);
|
|
| 996
|
:
|
|
| 997
|
: if (NumNewLines == 1)
|
|
| 998
|
: FirstNewLine = Range.begin();
|
|
| 999
|
: }
|
|
| 1000
|
: }
|
|
| 1001
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1002
|
: /// Match check string and its "not strings" and/or "dag strings".
|
1.1
|
(pascal 03-Sep-16)
| 1003
|
: size_t CheckString::Check(const SourceMgr &SM, StringRef Buffer,
|
|
| 1004
|
: bool IsLabelScanMode, size_t &MatchLen,
|
|
| 1005
|
: StringMap<StringRef> &VariableTable) const {
|
|
| 1006
|
: size_t LastPos = 0;
|
|
| 1007
|
: std::vector<const Pattern *> NotStrings;
|
|
| 1008
|
:
|
|
| 1009
|
: // IsLabelScanMode is true when we are scanning forward to find CHECK-LABEL
|
|
| 1010
|
: // bounds; we have not processed variable definitions within the bounded block
|
|
| 1011
|
: // yet so cannot handle any final CHECK-DAG yet; this is handled when going
|
|
| 1012
|
: // over the block again (including the last CHECK-LABEL) in normal mode.
|
|
| 1013
|
: if (!IsLabelScanMode) {
|
|
| 1014
|
: // Match "dag strings" (with mixed "not strings" if any).
|
|
| 1015
|
: LastPos = CheckDag(SM, Buffer, NotStrings, VariableTable);
|
|
| 1016
|
: if (LastPos == StringRef::npos)
|
|
| 1017
|
: return StringRef::npos;
|
|
| 1018
|
: }
|
|
| 1019
|
:
|
|
| 1020
|
: // Match itself from the last position after matching CHECK-DAG.
|
|
| 1021
|
: StringRef MatchBuffer = Buffer.substr(LastPos);
|
|
| 1022
|
: size_t MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
|
|
| 1023
|
: if (MatchPos == StringRef::npos) {
|
|
| 1024
|
: PrintCheckFailed(SM, *this, MatchBuffer, VariableTable);
|
|
| 1025
|
: return StringRef::npos;
|
|
| 1026
|
: }
|
|
| 1027
|
:
|
|
| 1028
|
: // Similar to the above, in "label-scan mode" we can't yet handle CHECK-NEXT
|
|
| 1029
|
: // or CHECK-NOT
|
|
| 1030
|
: if (!IsLabelScanMode) {
|
|
| 1031
|
: StringRef SkippedRegion = Buffer.substr(LastPos, MatchPos);
|
|
| 1032
|
:
|
|
| 1033
|
: // If this check is a "CHECK-NEXT", verify that the previous match was on
|
|
| 1034
|
: // the previous line (i.e. that there is one newline between them).
|
|
| 1035
|
: if (CheckNext(SM, SkippedRegion))
|
|
| 1036
|
: return StringRef::npos;
|
|
| 1037
|
:
|
|
| 1038
|
: // If this check is a "CHECK-SAME", verify that the previous match was on
|
|
| 1039
|
: // the same line (i.e. that there is no newline between them).
|
|
| 1040
|
: if (CheckSame(SM, SkippedRegion))
|
|
| 1041
|
: return StringRef::npos;
|
|
| 1042
|
:
|
|
| 1043
|
: // If this match had "not strings", verify that they don't exist in the
|
|
| 1044
|
: // skipped region.
|
|
| 1045
|
: if (CheckNot(SM, SkippedRegion, NotStrings, VariableTable))
|
|
| 1046
|
: return StringRef::npos;
|
|
| 1047
|
: }
|
|
| 1048
|
:
|
|
| 1049
|
: return LastPos + MatchPos;
|
|
| 1050
|
: }
|
|
| 1051
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1052
|
: /// Verify there is a single line in the given buffer.
|
1.1
|
(pascal 03-Sep-16)
| 1053
|
: bool CheckString::CheckNext(const SourceMgr &SM, StringRef Buffer) const {
|
1.1.1.2
|
(patrick 14-Jan-17)
| 1054
|
: if (Pat.getCheckTy() != Check::CheckNext)
|
1.1
|
(pascal 03-Sep-16)
| 1055
|
: return false;
|
|
| 1056
|
:
|
|
| 1057
|
: // Count the number of newlines between the previous match and this one.
|
|
| 1058
|
: assert(Buffer.data() !=
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1059
|
: SM.getMemoryBuffer(SM.FindBufferContainingLoc(
|
|
| 1060
|
: SMLoc::getFromPointer(Buffer.data())))
|
|
| 1061
|
: ->getBufferStart() &&
|
1.1
|
(pascal 03-Sep-16)
| 1062
|
: "CHECK-NEXT can't be the first check in a file");
|
|
| 1063
|
:
|
|
| 1064
|
: const char *FirstNewLine = nullptr;
|
|
| 1065
|
: unsigned NumNewLines = CountNumNewlinesBetween(Buffer, FirstNewLine);
|
|
| 1066
|
:
|
|
| 1067
|
: if (NumNewLines == 0) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1068
|
: SM.PrintMessage(Loc, SourceMgr::DK_Error,
|
|
| 1069
|
: Prefix + "-NEXT: is on the same line as previous match");
|
|
| 1070
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()), SourceMgr::DK_Note,
|
|
| 1071
|
: "'next' match was here");
|
1.1
|
(pascal 03-Sep-16)
| 1072
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
|
|
| 1073
|
: "previous match ended here");
|
|
| 1074
|
: return true;
|
|
| 1075
|
: }
|
|
| 1076
|
:
|
|
| 1077
|
: if (NumNewLines != 1) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1078
|
: SM.PrintMessage(Loc, SourceMgr::DK_Error,
|
|
| 1079
|
: Prefix +
|
|
| 1080
|
: "-NEXT: is not on the line after the previous match");
|
|
| 1081
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()), SourceMgr::DK_Note,
|
|
| 1082
|
: "'next' match was here");
|
1.1
|
(pascal 03-Sep-16)
| 1083
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
|
|
| 1084
|
: "previous match ended here");
|
|
| 1085
|
: SM.PrintMessage(SMLoc::getFromPointer(FirstNewLine), SourceMgr::DK_Note,
|
|
| 1086
|
: "non-matching line after previous match is here");
|
|
| 1087
|
: return true;
|
|
| 1088
|
: }
|
|
| 1089
|
:
|
|
| 1090
|
: return false;
|
|
| 1091
|
: }
|
|
| 1092
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1093
|
: /// Verify there is no newline in the given buffer.
|
1.1
|
(pascal 03-Sep-16)
| 1094
|
: bool CheckString::CheckSame(const SourceMgr &SM, StringRef Buffer) const {
|
1.1.1.2
|
(patrick 14-Jan-17)
| 1095
|
: if (Pat.getCheckTy() != Check::CheckSame)
|
1.1
|
(pascal 03-Sep-16)
| 1096
|
: return false;
|
|
| 1097
|
:
|
|
| 1098
|
: // Count the number of newlines between the previous match and this one.
|
|
| 1099
|
: assert(Buffer.data() !=
|
|
| 1100
|
: SM.getMemoryBuffer(SM.FindBufferContainingLoc(
|
|
| 1101
|
: SMLoc::getFromPointer(Buffer.data())))
|
|
| 1102
|
: ->getBufferStart() &&
|
|
| 1103
|
: "CHECK-SAME can't be the first check in a file");
|
|
| 1104
|
:
|
|
| 1105
|
: const char *FirstNewLine = nullptr;
|
|
| 1106
|
: unsigned NumNewLines = CountNumNewlinesBetween(Buffer, FirstNewLine);
|
|
| 1107
|
:
|
|
| 1108
|
: if (NumNewLines != 0) {
|
|
| 1109
|
: SM.PrintMessage(Loc, SourceMgr::DK_Error,
|
|
| 1110
|
: Prefix +
|
|
| 1111
|
: "-SAME: is not on the same line as the previous match");
|
|
| 1112
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()), SourceMgr::DK_Note,
|
|
| 1113
|
: "'next' match was here");
|
|
| 1114
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
|
|
| 1115
|
: "previous match ended here");
|
|
| 1116
|
: return true;
|
|
| 1117
|
: }
|
|
| 1118
|
:
|
|
| 1119
|
: return false;
|
|
| 1120
|
: }
|
|
| 1121
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1122
|
: /// Verify there's no "not strings" in the given buffer.
|
1.1
|
(pascal 03-Sep-16)
| 1123
|
: bool CheckString::CheckNot(const SourceMgr &SM, StringRef Buffer,
|
|
| 1124
|
: const std::vector<const Pattern *> &NotStrings,
|
|
| 1125
|
: StringMap<StringRef> &VariableTable) const {
|
|
| 1126
|
: for (const Pattern *Pat : NotStrings) {
|
|
| 1127
|
: assert((Pat->getCheckTy() == Check::CheckNot) && "Expect CHECK-NOT!");
|
|
| 1128
|
:
|
|
| 1129
|
: size_t MatchLen = 0;
|
|
| 1130
|
: size_t Pos = Pat->Match(Buffer, MatchLen, VariableTable);
|
|
| 1131
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1132
|
: if (Pos == StringRef::npos)
|
|
| 1133
|
: continue;
|
1.1
|
(pascal 03-Sep-16)
| 1134
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1135
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + Pos),
|
|
| 1136
|
: SourceMgr::DK_Error, Prefix + "-NOT: string occurred!");
|
1.1
|
(pascal 03-Sep-16)
| 1137
|
: SM.PrintMessage(Pat->getLoc(), SourceMgr::DK_Note,
|
|
| 1138
|
: Prefix + "-NOT: pattern specified here");
|
|
| 1139
|
: return true;
|
|
| 1140
|
: }
|
|
| 1141
|
:
|
|
| 1142
|
: return false;
|
|
| 1143
|
: }
|
|
| 1144
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1145
|
: /// Match "dag strings" and their mixed "not strings".
|
1.1
|
(pascal 03-Sep-16)
| 1146
|
: size_t CheckString::CheckDag(const SourceMgr &SM, StringRef Buffer,
|
|
| 1147
|
: std::vector<const Pattern *> &NotStrings,
|
|
| 1148
|
: StringMap<StringRef> &VariableTable) const {
|
|
| 1149
|
: if (DagNotStrings.empty())
|
|
| 1150
|
: return 0;
|
|
| 1151
|
:
|
|
| 1152
|
: size_t LastPos = 0;
|
|
| 1153
|
: size_t StartPos = LastPos;
|
|
| 1154
|
:
|
|
| 1155
|
: for (const Pattern &Pat : DagNotStrings) {
|
|
| 1156
|
: assert((Pat.getCheckTy() == Check::CheckDAG ||
|
|
| 1157
|
: Pat.getCheckTy() == Check::CheckNot) &&
|
|
| 1158
|
: "Invalid CHECK-DAG or CHECK-NOT!");
|
|
| 1159
|
:
|
|
| 1160
|
: if (Pat.getCheckTy() == Check::CheckNot) {
|
|
| 1161
|
: NotStrings.push_back(&Pat);
|
|
| 1162
|
: continue;
|
|
| 1163
|
: }
|
|
| 1164
|
:
|
|
| 1165
|
: assert((Pat.getCheckTy() == Check::CheckDAG) && "Expect CHECK-DAG!");
|
|
| 1166
|
:
|
|
| 1167
|
: size_t MatchLen = 0, MatchPos;
|
|
| 1168
|
:
|
|
| 1169
|
: // CHECK-DAG always matches from the start.
|
|
| 1170
|
: StringRef MatchBuffer = Buffer.substr(StartPos);
|
|
| 1171
|
: MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
|
|
| 1172
|
: // With a group of CHECK-DAGs, a single mismatching means the match on
|
|
| 1173
|
: // that group of CHECK-DAGs fails immediately.
|
|
| 1174
|
: if (MatchPos == StringRef::npos) {
|
|
| 1175
|
: PrintCheckFailed(SM, Pat.getLoc(), Pat, MatchBuffer, VariableTable);
|
|
| 1176
|
: return StringRef::npos;
|
|
| 1177
|
: }
|
|
| 1178
|
: // Re-calc it as the offset relative to the start of the original string.
|
|
| 1179
|
: MatchPos += StartPos;
|
|
| 1180
|
:
|
|
| 1181
|
: if (!NotStrings.empty()) {
|
|
| 1182
|
: if (MatchPos < LastPos) {
|
|
| 1183
|
: // Reordered?
|
|
| 1184
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + MatchPos),
|
|
| 1185
|
: SourceMgr::DK_Error,
|
|
| 1186
|
: Prefix + "-DAG: found a match of CHECK-DAG"
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1187
|
: " reordering across a CHECK-NOT");
|
1.1
|
(pascal 03-Sep-16)
| 1188
|
: SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + LastPos),
|
|
| 1189
|
: SourceMgr::DK_Note,
|
|
| 1190
|
: Prefix + "-DAG: the farthest match of CHECK-DAG"
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1191
|
: " is found here");
|
1.1
|
(pascal 03-Sep-16)
| 1192
|
: SM.PrintMessage(NotStrings[0]->getLoc(), SourceMgr::DK_Note,
|
|
| 1193
|
: Prefix + "-NOT: the crossed pattern specified"
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1194
|
: " here");
|
1.1
|
(pascal 03-Sep-16)
| 1195
|
: SM.PrintMessage(Pat.getLoc(), SourceMgr::DK_Note,
|
|
| 1196
|
: Prefix + "-DAG: the reordered pattern specified"
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1197
|
: " here");
|
1.1
|
(pascal 03-Sep-16)
| 1198
|
: return StringRef::npos;
|
|
| 1199
|
: }
|
|
| 1200
|
: // All subsequent CHECK-DAGs should be matched from the farthest
|
|
| 1201
|
: // position of all precedent CHECK-DAGs (including this one.)
|
|
| 1202
|
: StartPos = LastPos;
|
|
| 1203
|
: // If there's CHECK-NOTs between two CHECK-DAGs or from CHECK to
|
|
| 1204
|
: // CHECK-DAG, verify that there's no 'not' strings occurred in that
|
|
| 1205
|
: // region.
|
1.1.1.4
|
(patrick 04-Oct-17)
| 1206
|
: StringRef SkippedRegion = Buffer.slice(LastPos, MatchPos);
|
1.1
|
(pascal 03-Sep-16)
| 1207
|
: if (CheckNot(SM, SkippedRegion, NotStrings, VariableTable))
|
|
| 1208
|
: return StringRef::npos;
|
|
| 1209
|
: // Clear "not strings".
|
|
| 1210
|
: NotStrings.clear();
|
|
| 1211
|
: }
|
|
| 1212
|
:
|
|
| 1213
|
: // Update the last position with CHECK-DAG matches.
|
|
| 1214
|
: LastPos = std::max(MatchPos + MatchLen, LastPos);
|
|
| 1215
|
: }
|
|
| 1216
|
:
|
|
| 1217
|
: return LastPos;
|
|
| 1218
|
: }
|
|
| 1219
|
:
|
|
| 1220
|
: // A check prefix must contain only alphanumeric, hyphens and underscores.
|
|
| 1221
|
: static bool ValidateCheckPrefix(StringRef CheckPrefix) {
|
|
| 1222
|
: Regex Validator("^[a-zA-Z0-9_-]*$");
|
|
| 1223
|
: return Validator.match(CheckPrefix);
|
|
| 1224
|
: }
|
|
| 1225
|
:
|
|
| 1226
|
: static bool ValidateCheckPrefixes() {
|
|
| 1227
|
: StringSet<> PrefixSet;
|
|
| 1228
|
:
|
|
| 1229
|
: for (StringRef Prefix : CheckPrefixes) {
|
|
| 1230
|
: // Reject empty prefixes.
|
|
| 1231
|
: if (Prefix == "")
|
|
| 1232
|
: return false;
|
|
| 1233
|
:
|
|
| 1234
|
: if (!PrefixSet.insert(Prefix).second)
|
|
| 1235
|
: return false;
|
|
| 1236
|
:
|
|
| 1237
|
: if (!ValidateCheckPrefix(Prefix))
|
|
| 1238
|
: return false;
|
|
| 1239
|
: }
|
|
| 1240
|
:
|
|
| 1241
|
: return true;
|
|
| 1242
|
: }
|
|
| 1243
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1244
|
: // Combines the check prefixes into a single regex so that we can efficiently
|
|
| 1245
|
: // scan for any of the set.
|
|
| 1246
|
: //
|
|
| 1247
|
: // The semantics are that the longest-match wins which matches our regex
|
|
| 1248
|
: // library.
|
|
| 1249
|
: static Regex buildCheckPrefixRegex() {
|
|
| 1250
|
: // I don't think there's a way to specify an initial value for cl::list,
|
|
| 1251
|
: // so if nothing was specified, add the default
|
1.1
|
(pascal 03-Sep-16)
| 1252
|
: if (CheckPrefixes.empty())
|
|
| 1253
|
: CheckPrefixes.push_back("CHECK");
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1254
|
:
|
|
| 1255
|
: // We already validated the contents of CheckPrefixes so just concatenate
|
|
| 1256
|
: // them as alternatives.
|
|
| 1257
|
: SmallString<32> PrefixRegexStr;
|
|
| 1258
|
: for (StringRef Prefix : CheckPrefixes) {
|
|
| 1259
|
: if (Prefix != CheckPrefixes.front())
|
|
| 1260
|
: PrefixRegexStr.push_back('|');
|
|
| 1261
|
:
|
|
| 1262
|
: PrefixRegexStr.append(Prefix);
|
|
| 1263
|
: }
|
|
| 1264
|
:
|
|
| 1265
|
: return Regex(PrefixRegexStr);
|
1.1
|
(pascal 03-Sep-16)
| 1266
|
: }
|
|
| 1267
|
:
|
1.1.1.2
|
(patrick 14-Jan-17)
| 1268
|
: static void DumpCommandLine(int argc, char **argv) {
|
|
| 1269
|
: errs() << "FileCheck command line: ";
|
|
| 1270
|
: for (int I = 0; I < argc; I++)
|
|
| 1271
|
: errs() << " " << argv[I];
|
|
| 1272
|
: errs() << "\n";
|
|
| 1273
|
: }
|
|
| 1274
|
:
|
1.1.1.4
|
(patrick 04-Oct-17)
| 1275
|
: // Remove local variables from \p VariableTable. Global variables
|
|
| 1276
|
: // (start with '$') are preserved.
|
|
| 1277
|
: static void ClearLocalVars(StringMap<StringRef> &VariableTable) {
|
|
| 1278
|
: SmallVector<StringRef, 16> LocalVars;
|
|
| 1279
|
: for (const auto &Var : VariableTable)
|
|
| 1280
|
: if (Var.first()[0] != '$')
|
|
| 1281
|
: LocalVars.push_back(Var.first());
|
|
| 1282
|
:
|
|
| 1283
|
: for (const auto &Var : LocalVars)
|
|
| 1284
|
: VariableTable.erase(Var);
|
|
| 1285
|
: }
|
|
| 1286
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1287
|
: /// Check the input to FileCheck provided in the \p Buffer against the \p
|
|
| 1288
|
: /// CheckStrings read from the check file.
|
|
| 1289
|
: ///
|
|
| 1290
|
: /// Returns false if the input fails to satisfy the checks.
|
|
| 1291
|
: bool CheckInput(SourceMgr &SM, StringRef Buffer,
|
|
| 1292
|
: ArrayRef<CheckString> CheckStrings) {
|
|
| 1293
|
: bool ChecksFailed = false;
|
1.1
|
(pascal 03-Sep-16)
| 1294
|
:
|
|
| 1295
|
: /// VariableTable - This holds all the current filecheck variables.
|
|
| 1296
|
: StringMap<StringRef> VariableTable;
|
|
| 1297
|
:
|
|
| 1298
|
: unsigned i = 0, j = 0, e = CheckStrings.size();
|
|
| 1299
|
: while (true) {
|
|
| 1300
|
: StringRef CheckRegion;
|
|
| 1301
|
: if (j == e) {
|
|
| 1302
|
: CheckRegion = Buffer;
|
|
| 1303
|
: } else {
|
|
| 1304
|
: const CheckString &CheckLabelStr = CheckStrings[j];
|
1.1.1.2
|
(patrick 14-Jan-17)
| 1305
|
: if (CheckLabelStr.Pat.getCheckTy() != Check::CheckLabel) {
|
1.1
|
(pascal 03-Sep-16)
| 1306
|
: ++j;
|
|
| 1307
|
: continue;
|
|
| 1308
|
: }
|
|
| 1309
|
:
|
|
| 1310
|
: // Scan to next CHECK-LABEL match, ignoring CHECK-NOT and CHECK-DAG
|
|
| 1311
|
: size_t MatchLabelLen = 0;
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1312
|
: size_t MatchLabelPos =
|
|
| 1313
|
: CheckLabelStr.Check(SM, Buffer, true, MatchLabelLen, VariableTable);
|
|
| 1314
|
: if (MatchLabelPos == StringRef::npos)
|
|
| 1315
|
: // Immediately bail of CHECK-LABEL fails, nothing else we can do.
|
|
| 1316
|
: return false;
|
1.1
|
(pascal 03-Sep-16)
| 1317
|
:
|
|
| 1318
|
: CheckRegion = Buffer.substr(0, MatchLabelPos + MatchLabelLen);
|
|
| 1319
|
: Buffer = Buffer.substr(MatchLabelPos + MatchLabelLen);
|
|
| 1320
|
: ++j;
|
|
| 1321
|
: }
|
1.1.1.4
|
(patrick 04-Oct-17)
| 1322
|
:
|
|
| 1323
|
: if (EnableVarScope)
|
|
| 1324
|
: ClearLocalVars(VariableTable);
|
1.1
|
(pascal 03-Sep-16)
| 1325
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1326
|
: for (; i != j; ++i) {
|
1.1
|
(pascal 03-Sep-16)
| 1327
|
: const CheckString &CheckStr = CheckStrings[i];
|
|
| 1328
|
:
|
|
| 1329
|
: // Check each string within the scanned region, including a second check
|
|
| 1330
|
: // of any final CHECK-LABEL (to verify CHECK-NOT and CHECK-DAG)
|
|
| 1331
|
: size_t MatchLen = 0;
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1332
|
: size_t MatchPos =
|
|
| 1333
|
: CheckStr.Check(SM, CheckRegion, false, MatchLen, VariableTable);
|
1.1
|
(pascal 03-Sep-16)
| 1334
|
:
|
|
| 1335
|
: if (MatchPos == StringRef::npos) {
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1336
|
: ChecksFailed = true;
|
1.1
|
(pascal 03-Sep-16)
| 1337
|
: i = j;
|
|
| 1338
|
: break;
|
|
| 1339
|
: }
|
|
| 1340
|
:
|
|
| 1341
|
: CheckRegion = CheckRegion.substr(MatchPos + MatchLen);
|
|
| 1342
|
: }
|
|
| 1343
|
:
|
|
| 1344
|
: if (j == e)
|
|
| 1345
|
: break;
|
|
| 1346
|
: }
|
|
| 1347
|
:
|
1.1.1.3
|
(patrick 24-Jan-17)
| 1348
|
: // Success if no checks failed.
|
|
| 1349
|
: return !ChecksFailed;
|
|
| 1350
|
: }
|
|
| 1351
|
:
|
|
| 1352
|
: int main(int argc, char **argv) {
|
|
| 1353
|
: sys::PrintStackTraceOnErrorSignal(argv[0]);
|
|
| 1354
|
: PrettyStackTraceProgram X(argc, argv);
|
|
| 1355
|
: cl::ParseCommandLineOptions(argc, argv);
|
|
| 1356
|
:
|
|
| 1357
|
: if (!ValidateCheckPrefixes()) {
|
|
| 1358
|
: errs() << "Supplied check-prefix is invalid! Prefixes must be unique and "
|
|
| 1359
|
: "start with a letter and contain only alphanumeric characters, "
|
|
| 1360
|
: "hyphens and underscores\n";
|
|
| 1361
|
: return 2;
|
|
| 1362
|
: }
|
|
| 1363
|
:
|
|
| 1364
|
: Regex PrefixRE = buildCheckPrefixRegex();
|
|
| 1365
|
: std::string REError;
|
|
| 1366
|
: if (!PrefixRE.isValid(REError)) {
|
|
| 1367
|
: errs() << "Unable to combine check-prefix strings into a prefix regular "
|
|
| 1368
|
: "expression! This is likely a bug in FileCheck's verification of "
|
|
| 1369
|
: "the check-prefix strings. Regular expression parsing failed "
|
|
| 1370
|
: "with the following error: "
|
|
| 1371
|
: << REError << "\n";
|
|
| 1372
|
: return 2;
|
|
| 1373
|
: }
|
|
| 1374
|
:
|
|
| 1375
|
: SourceMgr SM;
|
|
| 1376
|
:
|
|
| 1377
|
: // Read the expected strings from the check file.
|
|
| 1378
|
: ErrorOr<std::unique_ptr<MemoryBuffer>> CheckFileOrErr =
|
|
| 1379
|
: MemoryBuffer::getFileOrSTDIN(CheckFilename);
|
|
| 1380
|
: if (std::error_code EC = CheckFileOrErr.getError()) {
|
|
| 1381
|
: errs() << "Could not open check file '" << CheckFilename
|
|
| 1382
|
: << "': " << EC.message() << '\n';
|
|
| 1383
|
: return 2;
|
|
| 1384
|
: }
|
|
| 1385
|
: MemoryBuffer &CheckFile = *CheckFileOrErr.get();
|
|
| 1386
|
:
|
|
| 1387
|
: SmallString<4096> CheckFileBuffer;
|
|
| 1388
|
: StringRef CheckFileText = CanonicalizeFile(CheckFile, CheckFileBuffer);
|
|
| 1389
|
:
|
|
| 1390
|
: SM.AddNewSourceBuffer(MemoryBuffer::getMemBuffer(
|
|
| 1391
|
: CheckFileText, CheckFile.getBufferIdentifier()),
|
|
| 1392
|
: SMLoc());
|
|
| 1393
|
:
|
|
| 1394
|
: std::vector<CheckString> CheckStrings;
|
|
| 1395
|
: if (ReadCheckFile(SM, CheckFileText, PrefixRE, CheckStrings))
|
|
| 1396
|
: return 2;
|
|
| 1397
|
:
|
|
| 1398
|
: // Open the file to check and add it to SourceMgr.
|
|
| 1399
|
: ErrorOr<std::unique_ptr<MemoryBuffer>> InputFileOrErr =
|
|
| 1400
|
: MemoryBuffer::getFileOrSTDIN(InputFilename);
|
|
| 1401
|
: if (std::error_code EC = InputFileOrErr.getError()) {
|
|
| 1402
|
: errs() << "Could not open input file '" << InputFilename
|
|
| 1403
|
: << "': " << EC.message() << '\n';
|
|
| 1404
|
: return 2;
|
|
| 1405
|
: }
|
|
| 1406
|
: MemoryBuffer &InputFile = *InputFileOrErr.get();
|
|
| 1407
|
:
|
|
| 1408
|
: if (InputFile.getBufferSize() == 0 && !AllowEmptyInput) {
|
|
| 1409
|
: errs() << "FileCheck error: '" << InputFilename << "' is empty.\n";
|
|
| 1410
|
: DumpCommandLine(argc, argv);
|
|
| 1411
|
: return 2;
|
|
| 1412
|
: }
|
|
| 1413
|
:
|
|
| 1414
|
: SmallString<4096> InputFileBuffer;
|
|
| 1415
|
: StringRef InputFileText = CanonicalizeFile(InputFile, InputFileBuffer);
|
|
| 1416
|
:
|
|
| 1417
|
: SM.AddNewSourceBuffer(MemoryBuffer::getMemBuffer(
|
|
| 1418
|
: InputFileText, InputFile.getBufferIdentifier()),
|
|
| 1419
|
: SMLoc());
|
|
| 1420
|
:
|
|
| 1421
|
: return CheckInput(SM, InputFileText, CheckStrings) ? EXIT_SUCCESS : 1;
|
1.1
|
(pascal 03-Sep-16)
| 1422
|
: }
|
Back