FileCheck.cpp revision d9a84efe44db2f4d983e49bc7370fc8cef449214 
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//
13// This program exits with an error status of 2 on error, exit status of 0 if
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/OwningPtr.h"
20#include "llvm/ADT/SmallString.h"
21#include "llvm/ADT/StringExtras.h"
22#include "llvm/ADT/StringMap.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 "llvm/Support/system_error.h"
31#include <algorithm>
32#include <map>
33#include <string>
34#include <vector>
35using namespace llvm;
36
37static cl::opt<std::string>
38CheckFilename(cl::Positional, cl::desc("<check-file>"), cl::Required);
39
40static cl::opt<std::string>
41InputFilename("input-file", cl::desc("File to check (defaults to stdin)"),
42              cl::init("-"), cl::value_desc("filename"));
43
44static cl::opt<std::string>
45CheckPrefix("check-prefix", cl::init("CHECK"),
46            cl::desc("Prefix to use from check file (defaults to 'CHECK')"));
47
48static cl::opt<bool>
49NoCanonicalizeWhiteSpace("strict-whitespace",
50              cl::desc("Do not treat all horizontal whitespace as equivalent"));
51
52//===----------------------------------------------------------------------===//
53// Pattern Handling Code.
54//===----------------------------------------------------------------------===//
55
56class Pattern {
57  SMLoc PatternLoc;
58
59  /// MatchEOF - When set, this pattern only matches the end of file. This is
60  /// used for trailing CHECK-NOTs.
61  bool MatchEOF;
62
63  /// MatchNot
64  bool MatchNot;
65
66  /// MatchDag
67  bool MatchDag;
68
69  /// FixedStr - If non-empty, this pattern is a fixed string match with the
70  /// specified fixed string.
71  StringRef FixedStr;
72
73  /// RegEx - If non-empty, this is a regex pattern.
74  std::string RegExStr;
75
76  /// \brief Contains the number of line this pattern is in.
77  unsigned LineNumber;
78
79  /// VariableUses - Entries in this vector map to uses of a variable in the
80  /// pattern, e.g. "foo[[bar]]baz".  In this case, the RegExStr will contain
81  /// "foobaz" and we'll get an entry in this vector that tells us to insert the
82  /// value of bar at offset 3.
83  std::vector<std::pair<StringRef, unsigned> > VariableUses;
84
85  /// VariableDefs - Maps definitions of variables to their parenthesized
86  /// capture numbers.
87  /// E.g. for the pattern "foo[[bar:.*]]baz", VariableDefs will map "bar" to 1.
88  std::map<StringRef, unsigned> VariableDefs;
89
90public:
91
92  Pattern(bool matchEOF = false)
93    : MatchEOF(matchEOF), MatchNot(false), MatchDag(false) { }
94
95  /// getLoc - Return the location in source code.
96  SMLoc getLoc() const { return PatternLoc; }
97
98  /// ParsePattern - Parse the given string into the Pattern.  SM provides the
99  /// SourceMgr used for error reports, and LineNumber is the line number in
100  /// the input file from which the pattern string was read.
101  /// Returns true in case of an error, false otherwise.
102  bool ParsePattern(StringRef PatternStr, SourceMgr &SM, unsigned LineNumber);
103
104  /// Match - Match the pattern string against the input buffer Buffer.  This
105  /// returns the position that is matched or npos if there is no match.  If
106  /// there is a match, the size of the matched string is returned in MatchLen.
107  ///
108  /// The VariableTable StringMap provides the current values of filecheck
109  /// variables and is updated if this match defines new values.
110  size_t Match(StringRef Buffer, size_t &MatchLen,
111               StringMap<StringRef> &VariableTable) const;
112
113  /// PrintFailureInfo - Print additional information about a failure to match
114  /// involving this pattern.
115  void PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
116                        const StringMap<StringRef> &VariableTable) const;
117
118  bool hasVariable() const { return !(VariableUses.empty() &&
119                                      VariableDefs.empty()); }
120
121  void setMatchNot(bool Not) { MatchNot = Not; }
122  bool getMatchNot() const { return MatchNot; }
123
124  void setMatchDag(bool Dag) { MatchDag = Dag; }
125  bool getMatchDag() const { return MatchDag; }
126
127private:
128  static void AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr);
129  bool AddRegExToRegEx(StringRef RS, unsigned &CurParen, SourceMgr &SM);
130  void AddBackrefToRegEx(unsigned BackrefNum);
131
132  /// ComputeMatchDistance - Compute an arbitrary estimate for the quality of
133  /// matching this pattern at the start of \arg Buffer; a distance of zero
134  /// should correspond to a perfect match.
135  unsigned ComputeMatchDistance(StringRef Buffer,
136                               const StringMap<StringRef> &VariableTable) const;
137
138  /// \brief Evaluates expression and stores the result to \p Value.
139  /// \return true on success. false when the expression has invalid syntax.
140  bool EvaluateExpression(StringRef Expr, std::string &Value) const;
141
142  /// \brief Finds the closing sequence of a regex variable usage or
143  /// definition. Str has to point in the beginning of the definition
144  /// (right after the opening sequence).
145  /// \return offset of the closing sequence within Str, or npos if it was not
146  /// found.
147  size_t FindRegexVarEnd(StringRef Str);
148};
149
150
151bool Pattern::ParsePattern(StringRef PatternStr, SourceMgr &SM,
152                           unsigned LineNumber) {
153  this->LineNumber = LineNumber;
154  PatternLoc = SMLoc::getFromPointer(PatternStr.data());
155
156  // Ignore trailing whitespace.
157  while (!PatternStr.empty() &&
158         (PatternStr.back() == ' ' || PatternStr.back() == '\t'))
159    PatternStr = PatternStr.substr(0, PatternStr.size()-1);
160
161  // Check that there is something on the line.
162  if (PatternStr.empty()) {
163    SM.PrintMessage(PatternLoc, SourceMgr::DK_Error,
164                    "found empty check string with prefix '" +
165                    CheckPrefix+":'");
166    return true;
167  }
168
169  // Check to see if this is a fixed string, or if it has regex pieces.
170  if (PatternStr.size() < 2 ||
171      (PatternStr.find("{{") == StringRef::npos &&
172       PatternStr.find("[[") == StringRef::npos)) {
173    FixedStr = PatternStr;
174    return false;
175  }
176
177  // Paren value #0 is for the fully matched string.  Any new parenthesized
178  // values add from there.
179  unsigned CurParen = 1;
180
181  // Otherwise, there is at least one regex piece.  Build up the regex pattern
182  // by escaping scary characters in fixed strings, building up one big regex.
183  while (!PatternStr.empty()) {
184    // RegEx matches.
185    if (PatternStr.startswith("{{")) {
186      // This is the start of a regex match.  Scan for the }}.
187      size_t End = PatternStr.find("}}");
188      if (End == StringRef::npos) {
189        SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
190                        SourceMgr::DK_Error,
191                        "found start of regex string with no end '}}'");
192        return true;
193      }
194
195      // Enclose {{}} patterns in parens just like [[]] even though we're not
196      // capturing the result for any purpose.  This is required in case the
197      // expression contains an alternation like: CHECK:  abc{{x|z}}def.  We
198      // want this to turn into: "abc(x|z)def" not "abcx|zdef".
199      RegExStr += '(';
200      ++CurParen;
201
202      if (AddRegExToRegEx(PatternStr.substr(2, End-2), CurParen, SM))
203        return true;
204      RegExStr += ')';
205
206      PatternStr = PatternStr.substr(End+2);
207      continue;
208    }
209
210    // Named RegEx matches.  These are of two forms: [[foo:.*]] which matches .*
211    // (or some other regex) and assigns it to the FileCheck variable 'foo'. The
212    // second form is [[foo]] which is a reference to foo.  The variable name
213    // itself must be of the form "[a-zA-Z_][0-9a-zA-Z_]*", otherwise we reject
214    // it.  This is to catch some common errors.
215    if (PatternStr.startswith("[[")) {
216      // Find the closing bracket pair ending the match.  End is going to be an
217      // offset relative to the beginning of the match string.
218      size_t End = FindRegexVarEnd(PatternStr.substr(2));
219
220      if (End == StringRef::npos) {
221        SM.PrintMessage(SMLoc::getFromPointer(PatternStr.data()),
222                        SourceMgr::DK_Error,
223                        "invalid named regex reference, no ]] found");
224        return true;
225      }
226
227      StringRef MatchStr = PatternStr.substr(2, End);
228      PatternStr = PatternStr.substr(End+4);
229
230      // Get the regex name (e.g. "foo").
231      size_t NameEnd = MatchStr.find(':');
232      StringRef Name = MatchStr.substr(0, NameEnd);
233
234      if (Name.empty()) {
235        SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
236                        "invalid name in named regex: empty name");
237        return true;
238      }
239
240      // Verify that the name/expression is well formed. FileCheck currently
241      // supports @LINE, @LINE+number, @LINE-number expressions. The check here
242      // is relaxed, more strict check is performed in \c EvaluateExpression.
243      bool IsExpression = false;
244      for (unsigned i = 0, e = Name.size(); i != e; ++i) {
245        if (i == 0 && Name[i] == '@') {
246          if (NameEnd != StringRef::npos) {
247            SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
248                            SourceMgr::DK_Error,
249                            "invalid name in named regex definition");
250            return true;
251          }
252          IsExpression = true;
253          continue;
254        }
255        if (Name[i] != '_' && !isalnum(Name[i]) &&
256            (!IsExpression || (Name[i] != '+' && Name[i] != '-'))) {
257          SM.PrintMessage(SMLoc::getFromPointer(Name.data()+i),
258                          SourceMgr::DK_Error, "invalid name in named regex");
259          return true;
260        }
261      }
262
263      // Name can't start with a digit.
264      if (isdigit(static_cast<unsigned char>(Name[0]))) {
265        SM.PrintMessage(SMLoc::getFromPointer(Name.data()), SourceMgr::DK_Error,
266                        "invalid name in named regex");
267        return true;
268      }
269
270      // Handle [[foo]].
271      if (NameEnd == StringRef::npos) {
272        // Handle variables that were defined earlier on the same line by
273        // emitting a backreference.
274        if (VariableDefs.find(Name) != VariableDefs.end()) {
275          unsigned VarParenNum = VariableDefs[Name];
276          if (VarParenNum < 1 || VarParenNum > 9) {
277            SM.PrintMessage(SMLoc::getFromPointer(Name.data()),
278                            SourceMgr::DK_Error,
279                            "Can't back-reference more than 9 variables");
280            return true;
281          }
282          AddBackrefToRegEx(VarParenNum);
283        } else {
284          VariableUses.push_back(std::make_pair(Name, RegExStr.size()));
285        }
286        continue;
287      }
288
289      // Handle [[foo:.*]].
290      VariableDefs[Name] = CurParen;
291      RegExStr += '(';
292      ++CurParen;
293
294      if (AddRegExToRegEx(MatchStr.substr(NameEnd+1), CurParen, SM))
295        return true;
296
297      RegExStr += ')';
298    }
299
300    // Handle fixed string matches.
301    // Find the end, which is the start of the next regex.
302    size_t FixedMatchEnd = PatternStr.find("{{");
303    FixedMatchEnd = std::min(FixedMatchEnd, PatternStr.find("[["));
304    AddFixedStringToRegEx(PatternStr.substr(0, FixedMatchEnd), RegExStr);
305    PatternStr = PatternStr.substr(FixedMatchEnd);
306  }
307
308  return false;
309}
310
311void Pattern::AddFixedStringToRegEx(StringRef FixedStr, std::string &TheStr) {
312  // Add the characters from FixedStr to the regex, escaping as needed.  This
313  // avoids "leaning toothpicks" in common patterns.
314  for (unsigned i = 0, e = FixedStr.size(); i != e; ++i) {
315    switch (FixedStr[i]) {
316    // These are the special characters matched in "p_ere_exp".
317    case '(':
318    case ')':
319    case '^':
320    case '$':
321    case '|':
322    case '*':
323    case '+':
324    case '?':
325    case '.':
326    case '[':
327    case '\\':
328    case '{':
329      TheStr += '\\';
330      // FALL THROUGH.
331    default:
332      TheStr += FixedStr[i];
333      break;
334    }
335  }
336}
337
338bool Pattern::AddRegExToRegEx(StringRef RS, unsigned &CurParen,
339                              SourceMgr &SM) {
340  Regex R(RS);
341  std::string Error;
342  if (!R.isValid(Error)) {
343    SM.PrintMessage(SMLoc::getFromPointer(RS.data()), SourceMgr::DK_Error,
344                    "invalid regex: " + Error);
345    return true;
346  }
347
348  RegExStr += RS.str();
349  CurParen += R.getNumMatches();
350  return false;
351}
352
353void Pattern::AddBackrefToRegEx(unsigned BackrefNum) {
354  assert(BackrefNum >= 1 && BackrefNum <= 9 && "Invalid backref number");
355  std::string Backref = std::string("\\") +
356                        std::string(1, '0' + BackrefNum);
357  RegExStr += Backref;
358}
359
360bool Pattern::EvaluateExpression(StringRef Expr, std::string &Value) const {
361  // The only supported expression is @LINE([\+-]\d+)?
362  if (!Expr.startswith("@LINE"))
363    return false;
364  Expr = Expr.substr(StringRef("@LINE").size());
365  int Offset = 0;
366  if (!Expr.empty()) {
367    if (Expr[0] == '+')
368      Expr = Expr.substr(1);
369    else if (Expr[0] != '-')
370      return false;
371    if (Expr.getAsInteger(10, Offset))
372      return false;
373  }
374  Value = llvm::itostr(LineNumber + Offset);
375  return true;
376}
377
378/// Match - Match the pattern string against the input buffer Buffer.  This
379/// returns the position that is matched or npos if there is no match.  If
380/// there is a match, the size of the matched string is returned in MatchLen.
381size_t Pattern::Match(StringRef Buffer, size_t &MatchLen,
382                      StringMap<StringRef> &VariableTable) const {
383  // If this is the EOF pattern, match it immediately.
384  if (MatchEOF) {
385    MatchLen = 0;
386    return Buffer.size();
387  }
388
389  // If this is a fixed string pattern, just match it now.
390  if (!FixedStr.empty()) {
391    MatchLen = FixedStr.size();
392    return Buffer.find(FixedStr);
393  }
394
395  // Regex match.
396
397  // If there are variable uses, we need to create a temporary string with the
398  // actual value.
399  StringRef RegExToMatch = RegExStr;
400  std::string TmpStr;
401  if (!VariableUses.empty()) {
402    TmpStr = RegExStr;
403
404    unsigned InsertOffset = 0;
405    for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
406      std::string Value;
407
408      if (VariableUses[i].first[0] == '@') {
409        if (!EvaluateExpression(VariableUses[i].first, Value))
410          return StringRef::npos;
411      } else {
412        StringMap<StringRef>::iterator it =
413          VariableTable.find(VariableUses[i].first);
414        // If the variable is undefined, return an error.
415        if (it == VariableTable.end())
416          return StringRef::npos;
417
418        // Look up the value and escape it so that we can plop it into the regex.
419        AddFixedStringToRegEx(it->second, Value);
420      }
421
422      // Plop it into the regex at the adjusted offset.
423      TmpStr.insert(TmpStr.begin()+VariableUses[i].second+InsertOffset,
424                    Value.begin(), Value.end());
425      InsertOffset += Value.size();
426    }
427
428    // Match the newly constructed regex.
429    RegExToMatch = TmpStr;
430  }
431
432
433  SmallVector<StringRef, 4> MatchInfo;
434  if (!Regex(RegExToMatch, Regex::Newline).match(Buffer, &MatchInfo))
435    return StringRef::npos;
436
437  // Successful regex match.
438  assert(!MatchInfo.empty() && "Didn't get any match");
439  StringRef FullMatch = MatchInfo[0];
440
441  // If this defines any variables, remember their values.
442  for (std::map<StringRef, unsigned>::const_iterator I = VariableDefs.begin(),
443                                                     E = VariableDefs.end();
444       I != E; ++I) {
445    assert(I->second < MatchInfo.size() && "Internal paren error");
446    VariableTable[I->first] = MatchInfo[I->second];
447  }
448
449  MatchLen = FullMatch.size();
450  return FullMatch.data()-Buffer.data();
451}
452
453unsigned Pattern::ComputeMatchDistance(StringRef Buffer,
454                              const StringMap<StringRef> &VariableTable) const {
455  // Just compute the number of matching characters. For regular expressions, we
456  // just compare against the regex itself and hope for the best.
457  //
458  // FIXME: One easy improvement here is have the regex lib generate a single
459  // example regular expression which matches, and use that as the example
460  // string.
461  StringRef ExampleString(FixedStr);
462  if (ExampleString.empty())
463    ExampleString = RegExStr;
464
465  // Only compare up to the first line in the buffer, or the string size.
466  StringRef BufferPrefix = Buffer.substr(0, ExampleString.size());
467  BufferPrefix = BufferPrefix.split('\n').first;
468  return BufferPrefix.edit_distance(ExampleString);
469}
470
471void Pattern::PrintFailureInfo(const SourceMgr &SM, StringRef Buffer,
472                               const StringMap<StringRef> &VariableTable) const{
473  // If this was a regular expression using variables, print the current
474  // variable values.
475  if (!VariableUses.empty()) {
476    for (unsigned i = 0, e = VariableUses.size(); i != e; ++i) {
477      SmallString<256> Msg;
478      raw_svector_ostream OS(Msg);
479      StringRef Var = VariableUses[i].first;
480      if (Var[0] == '@') {
481        std::string Value;
482        if (EvaluateExpression(Var, Value)) {
483          OS << "with expression \"";
484          OS.write_escaped(Var) << "\" equal to \"";
485          OS.write_escaped(Value) << "\"";
486        } else {
487          OS << "uses incorrect expression \"";
488          OS.write_escaped(Var) << "\"";
489        }
490      } else {
491        StringMap<StringRef>::const_iterator it = VariableTable.find(Var);
492
493        // Check for undefined variable references.
494        if (it == VariableTable.end()) {
495          OS << "uses undefined variable \"";
496          OS.write_escaped(Var) << "\"";
497        } else {
498          OS << "with variable \"";
499          OS.write_escaped(Var) << "\" equal to \"";
500          OS.write_escaped(it->second) << "\"";
501        }
502      }
503
504      SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
505                      OS.str());
506    }
507  }
508
509  // Attempt to find the closest/best fuzzy match.  Usually an error happens
510  // because some string in the output didn't exactly match. In these cases, we
511  // would like to show the user a best guess at what "should have" matched, to
512  // save them having to actually check the input manually.
513  size_t NumLinesForward = 0;
514  size_t Best = StringRef::npos;
515  double BestQuality = 0;
516
517  // Use an arbitrary 4k limit on how far we will search.
518  for (size_t i = 0, e = std::min(size_t(4096), Buffer.size()); i != e; ++i) {
519    if (Buffer[i] == '\n')
520      ++NumLinesForward;
521
522    // Patterns have leading whitespace stripped, so skip whitespace when
523    // looking for something which looks like a pattern.
524    if (Buffer[i] == ' ' || Buffer[i] == '\t')
525      continue;
526
527    // Compute the "quality" of this match as an arbitrary combination of the
528    // match distance and the number of lines skipped to get to this match.
529    unsigned Distance = ComputeMatchDistance(Buffer.substr(i), VariableTable);
530    double Quality = Distance + (NumLinesForward / 100.);
531
532    if (Quality < BestQuality || Best == StringRef::npos) {
533      Best = i;
534      BestQuality = Quality;
535    }
536  }
537
538  // Print the "possible intended match here" line if we found something
539  // reasonable and not equal to what we showed in the "scanning from here"
540  // line.
541  if (Best && Best != StringRef::npos && BestQuality < 50) {
542      SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + Best),
543                      SourceMgr::DK_Note, "possible intended match here");
544
545    // FIXME: If we wanted to be really friendly we would show why the match
546    // failed, as it can be hard to spot simple one character differences.
547  }
548}
549
550size_t Pattern::FindRegexVarEnd(StringRef Str) {
551  // Offset keeps track of the current offset within the input Str
552  size_t Offset = 0;
553  // [...] Nesting depth
554  size_t BracketDepth = 0;
555
556  while (!Str.empty()) {
557    if (Str.startswith("]]") && BracketDepth == 0)
558      return Offset;
559    if (Str[0] == '\\') {
560      // Backslash escapes the next char within regexes, so skip them both.
561      Str = Str.substr(2);
562      Offset += 2;
563    } else {
564      switch (Str[0]) {
565        default:
566          break;
567        case '[':
568          BracketDepth++;
569          break;
570        case ']':
571          assert(BracketDepth > 0 && "Invalid regex");
572          BracketDepth--;
573          break;
574      }
575      Str = Str.substr(1);
576      Offset++;
577    }
578  }
579
580  return StringRef::npos;
581}
582
583
584//===----------------------------------------------------------------------===//
585// Check Strings.
586//===----------------------------------------------------------------------===//
587
588/// CheckString - This is a check that we found in the input file.
589struct CheckString {
590  /// Pat - The pattern to match.
591  Pattern Pat;
592
593  /// Loc - The location in the match file that the check string was specified.
594  SMLoc Loc;
595
596  /// IsCheckNext - This is true if this is a CHECK-NEXT: directive (as opposed
597  /// to a CHECK: directive.
598  bool IsCheckNext;
599
600  /// IsCheckLabel - This is true if this is a CHECK-LABEL: directive (as
601  /// opposed to a CHECK: directive.
602  bool IsCheckLabel;
603
604  /// DagNotStrings - These are all of the strings that are disallowed from
605  /// occurring between this match string and the previous one (or start of
606  /// file).
607  std::vector<Pattern> DagNotStrings;
608
609  CheckString(const Pattern &P, SMLoc L, bool isCheckNext, bool isCheckLabel)
610    : Pat(P), Loc(L), IsCheckNext(isCheckNext), IsCheckLabel(isCheckLabel) {}
611
612  /// Check - Match check string and its "not strings" and/or "dag strings".
613  size_t Check(const SourceMgr &SM, StringRef Buffer, bool IsLabel,
614               size_t &MatchLen, StringMap<StringRef> &VariableTable) const;
615
616  /// CheckNext - Verify there is a single line in the given buffer.
617  bool CheckNext(const SourceMgr &SM, StringRef Buffer) const;
618
619  /// CheckNot - Verify there's no "not strings" in the given buffer.
620  bool CheckNot(const SourceMgr &SM, StringRef Buffer,
621                const std::vector<const Pattern *> &NotStrings,
622                StringMap<StringRef> &VariableTable) const;
623
624  /// CheckDag - Match "dag strings" and their mixed "not strings".
625  size_t CheckDag(const SourceMgr &SM, StringRef Buffer,
626                  std::vector<const Pattern *> &NotStrings,
627                  StringMap<StringRef> &VariableTable) const;
628};
629
630/// Canonicalize whitespaces in the input file. Line endings are replaced
631/// with UNIX-style '\n'.
632///
633/// \param PreserveHorizontal Don't squash consecutive horizontal whitespace
634/// characters to a single space.
635static MemoryBuffer *CanonicalizeInputFile(MemoryBuffer *MB,
636                                           bool PreserveHorizontal) {
637  SmallString<128> NewFile;
638  NewFile.reserve(MB->getBufferSize());
639
640  for (const char *Ptr = MB->getBufferStart(), *End = MB->getBufferEnd();
641       Ptr != End; ++Ptr) {
642    // Eliminate trailing dosish \r.
643    if (Ptr <= End - 2 && Ptr[0] == '\r' && Ptr[1] == '\n') {
644      continue;
645    }
646
647    // If current char is not a horizontal whitespace or if horizontal
648    // whitespace canonicalization is disabled, dump it to output as is.
649    if (PreserveHorizontal || (*Ptr != ' ' && *Ptr != '\t')) {
650      NewFile.push_back(*Ptr);
651      continue;
652    }
653
654    // Otherwise, add one space and advance over neighboring space.
655    NewFile.push_back(' ');
656    while (Ptr+1 != End &&
657           (Ptr[1] == ' ' || Ptr[1] == '\t'))
658      ++Ptr;
659  }
660
661  // Free the old buffer and return a new one.
662  MemoryBuffer *MB2 =
663    MemoryBuffer::getMemBufferCopy(NewFile.str(), MB->getBufferIdentifier());
664
665  delete MB;
666  return MB2;
667}
668
669
670/// ReadCheckFile - Read the check file, which specifies the sequence of
671/// expected strings.  The strings are added to the CheckStrings vector.
672/// Returns true in case of an error, false otherwise.
673static bool ReadCheckFile(SourceMgr &SM,
674                          std::vector<CheckString> &CheckStrings) {
675  OwningPtr<MemoryBuffer> File;
676  if (error_code ec =
677        MemoryBuffer::getFileOrSTDIN(CheckFilename, File)) {
678    errs() << "Could not open check file '" << CheckFilename << "': "
679           << ec.message() << '\n';
680    return true;
681  }
682
683  // If we want to canonicalize whitespace, strip excess whitespace from the
684  // buffer containing the CHECK lines. Remove DOS style line endings.
685  MemoryBuffer *F =
686    CanonicalizeInputFile(File.take(), NoCanonicalizeWhiteSpace);
687
688  SM.AddNewSourceBuffer(F, SMLoc());
689
690  // Find all instances of CheckPrefix followed by : in the file.
691  StringRef Buffer = F->getBuffer();
692  std::vector<Pattern> DagNotMatches;
693
694  // LineNumber keeps track of the line on which CheckPrefix instances are
695  // found.
696  unsigned LineNumber = 1;
697
698  while (1) {
699    // See if Prefix occurs in the memory buffer.
700    size_t PrefixLoc = Buffer.find(CheckPrefix);
701    // If we didn't find a match, we're done.
702    if (PrefixLoc == StringRef::npos)
703      break;
704
705    LineNumber += Buffer.substr(0, PrefixLoc).count('\n');
706
707    // Keep the charcter before our prefix so we can validate that we have
708    // found our prefix, and account for cases when PrefixLoc is 0.
709    Buffer = Buffer.substr(std::min(PrefixLoc-1, PrefixLoc));
710
711    const char *CheckPrefixStart = Buffer.data() + (PrefixLoc == 0 ? 0 : 1);
712
713    // When we find a check prefix, keep track of whether we find CHECK: or
714    // CHECK-NEXT:
715    bool IsCheckNext = false, IsCheckNot = false, IsCheckDag = false,
716         IsCheckLabel = false;
717
718    // Make sure we have actually found our prefix, and not a word containing
719    // our prefix.
720    if (PrefixLoc != 0 && (isalnum(Buffer[0]) ||
721                           Buffer[0] == '-'   ||
722                           Buffer[0] == '_')) {
723      Buffer = Buffer.substr(CheckPrefix.size());
724      continue;
725    }
726
727    // Verify that the : is present after the prefix.
728    if (Buffer[CheckPrefix.size()] == ':') {
729      Buffer = Buffer.substr(CheckPrefix.size()+1);
730    } else if (Buffer.size() > CheckPrefix.size()+6 &&
731               memcmp(Buffer.data()+CheckPrefix.size(), "-NEXT:", 6) == 0) {
732      Buffer = Buffer.substr(CheckPrefix.size()+6);
733      IsCheckNext = true;
734    } else if (Buffer.size() > CheckPrefix.size()+5 &&
735               memcmp(Buffer.data()+CheckPrefix.size(), "-NOT:", 5) == 0) {
736      Buffer = Buffer.substr(CheckPrefix.size()+5);
737      IsCheckNot = true;
738    } else if (Buffer.size() > CheckPrefix.size()+5 &&
739               memcmp(Buffer.data()+CheckPrefix.size(), "-DAG:", 5) == 0) {
740      Buffer = Buffer.substr(CheckPrefix.size()+5);
741      IsCheckDag = true;
742    } else if (Buffer.size() > CheckPrefix.size()+7 &&
743               memcmp(Buffer.data()+CheckPrefix.size(), "-LABEL:", 7) == 0) {
744      Buffer = Buffer.substr(CheckPrefix.size()+7);
745      IsCheckLabel = true;
746    } else {
747      Buffer = Buffer.substr(1);
748      continue;
749    }
750
751    // Okay, we found the prefix, yay.  Remember the rest of the line, but
752    // ignore leading and trailing whitespace.
753    Buffer = Buffer.substr(Buffer.find_first_not_of(" \t"));
754
755    // Scan ahead to the end of line.
756    size_t EOL = Buffer.find_first_of("\n\r");
757
758    // Remember the location of the start of the pattern, for diagnostics.
759    SMLoc PatternLoc = SMLoc::getFromPointer(Buffer.data());
760
761    // Parse the pattern.
762    Pattern P;
763    if (P.ParsePattern(Buffer.substr(0, EOL), SM, LineNumber))
764      return true;
765
766    // Verify that CHECK-LABEL lines do not define or use variables
767    if (IsCheckLabel && P.hasVariable()) {
768      SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
769                      SourceMgr::DK_Error,
770                      "found '"+CheckPrefix+"-LABEL:' with variable definition"
771                      " or use'");
772      return true;
773    }
774
775    P.setMatchNot(IsCheckNot);
776    P.setMatchDag(IsCheckDag);
777
778    Buffer = Buffer.substr(EOL);
779
780    // Verify that CHECK-NEXT lines have at least one CHECK line before them.
781    if (IsCheckNext && CheckStrings.empty()) {
782      SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
783                      SourceMgr::DK_Error,
784                      "found '"+CheckPrefix+"-NEXT:' without previous '"+
785                      CheckPrefix+ ": line");
786      return true;
787    }
788
789    // Handle CHECK-DAG/-NOT.
790    if (IsCheckDag || IsCheckNot) {
791      DagNotMatches.push_back(P);
792      continue;
793    }
794
795    // Okay, add the string we captured to the output vector and move on.
796    CheckStrings.push_back(CheckString(P,
797                                       PatternLoc,
798                                       IsCheckNext,
799                                       IsCheckLabel));
800    std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
801  }
802
803  // Add an EOF pattern for any trailing CHECK-DAG/-NOTs.
804  if (!DagNotMatches.empty()) {
805    CheckStrings.push_back(CheckString(Pattern(true),
806                                       SMLoc::getFromPointer(Buffer.data()),
807                                       false,
808                                       false));
809    std::swap(DagNotMatches, CheckStrings.back().DagNotStrings);
810  }
811
812  if (CheckStrings.empty()) {
813    errs() << "error: no check strings found with prefix '" << CheckPrefix
814           << ":'\n";
815    return true;
816  }
817
818  return false;
819}
820
821static void PrintCheckFailed(const SourceMgr &SM, const SMLoc &Loc,
822                             const Pattern &Pat, StringRef Buffer,
823                             StringMap<StringRef> &VariableTable) {
824  // Otherwise, we have an error, emit an error message.
825  SM.PrintMessage(Loc, SourceMgr::DK_Error,
826                  "expected string not found in input");
827
828  // Print the "scanning from here" line.  If the current position is at the
829  // end of a line, advance to the start of the next line.
830  Buffer = Buffer.substr(Buffer.find_first_not_of(" \t\n\r"));
831
832  SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
833                  "scanning from here");
834
835  // Allow the pattern to print additional information if desired.
836  Pat.PrintFailureInfo(SM, Buffer, VariableTable);
837}
838
839static void PrintCheckFailed(const SourceMgr &SM, const CheckString &CheckStr,
840                             StringRef Buffer,
841                             StringMap<StringRef> &VariableTable) {
842  PrintCheckFailed(SM, CheckStr.Loc, CheckStr.Pat, Buffer, VariableTable);
843}
844
845/// CountNumNewlinesBetween - Count the number of newlines in the specified
846/// range.
847static unsigned CountNumNewlinesBetween(StringRef Range) {
848  unsigned NumNewLines = 0;
849  while (1) {
850    // Scan for newline.
851    Range = Range.substr(Range.find_first_of("\n\r"));
852    if (Range.empty()) return NumNewLines;
853
854    ++NumNewLines;
855
856    // Handle \n\r and \r\n as a single newline.
857    if (Range.size() > 1 &&
858        (Range[1] == '\n' || Range[1] == '\r') &&
859        (Range[0] != Range[1]))
860      Range = Range.substr(1);
861    Range = Range.substr(1);
862  }
863}
864
865size_t CheckString::Check(const SourceMgr &SM, StringRef Buffer,
866                          bool IsLabel, size_t &MatchLen,
867                          StringMap<StringRef> &VariableTable) const {
868  size_t LastPos = 0;
869  std::vector<const Pattern *> NotStrings;
870
871  if (!IsLabel) {
872    // Match "dag strings" (with mixed "not strings" if any).
873    LastPos = CheckDag(SM, Buffer, NotStrings, VariableTable);
874    if (LastPos == StringRef::npos)
875      return StringRef::npos;
876  }
877
878  // Match itself from the last position after matching CHECK-DAG.
879  StringRef MatchBuffer = Buffer.substr(LastPos);
880  size_t MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
881  if (MatchPos == StringRef::npos) {
882    PrintCheckFailed(SM, *this, MatchBuffer, VariableTable);
883    return StringRef::npos;
884  }
885  MatchPos += LastPos;
886
887  if (!IsLabel) {
888    StringRef SkippedRegion = Buffer.substr(LastPos, MatchPos);
889
890    // If this check is a "CHECK-NEXT", verify that the previous match was on
891    // the previous line (i.e. that there is one newline between them).
892    if (CheckNext(SM, SkippedRegion))
893      return StringRef::npos;
894
895    // If this match had "not strings", verify that they don't exist in the
896    // skipped region.
897    if (CheckNot(SM, SkippedRegion, NotStrings, VariableTable))
898      return StringRef::npos;
899  }
900
901  return MatchPos;
902}
903
904bool CheckString::CheckNext(const SourceMgr &SM, StringRef Buffer) const {
905  if (!IsCheckNext)
906    return false;
907
908  // Count the number of newlines between the previous match and this one.
909  assert(Buffer.data() !=
910         SM.getMemoryBuffer(
911           SM.FindBufferContainingLoc(
912             SMLoc::getFromPointer(Buffer.data())))->getBufferStart() &&
913         "CHECK-NEXT can't be the first check in a file");
914
915  unsigned NumNewLines = CountNumNewlinesBetween(Buffer);
916
917  if (NumNewLines == 0) {
918    SM.PrintMessage(Loc, SourceMgr::DK_Error, CheckPrefix+
919                    "-NEXT: is on the same line as previous match");
920    SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()),
921                    SourceMgr::DK_Note, "'next' match was here");
922    SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
923                    "previous match ended here");
924    return true;
925  }
926
927  if (NumNewLines != 1) {
928    SM.PrintMessage(Loc, SourceMgr::DK_Error, CheckPrefix+
929                    "-NEXT: is not on the line after the previous match");
930    SM.PrintMessage(SMLoc::getFromPointer(Buffer.end()),
931                    SourceMgr::DK_Note, "'next' match was here");
932    SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), SourceMgr::DK_Note,
933                    "previous match ended here");
934    return true;
935  }
936
937  return false;
938}
939
940bool CheckString::CheckNot(const SourceMgr &SM, StringRef Buffer,
941                           const std::vector<const Pattern *> &NotStrings,
942                           StringMap<StringRef> &VariableTable) const {
943  for (unsigned ChunkNo = 0, e = NotStrings.size();
944       ChunkNo != e; ++ChunkNo) {
945    const Pattern *Pat = NotStrings[ChunkNo];
946    assert(Pat->getMatchNot() && "Expect CHECK-NOT!");
947
948    size_t MatchLen = 0;
949    size_t Pos = Pat->Match(Buffer, MatchLen, VariableTable);
950
951    if (Pos == StringRef::npos) continue;
952
953    SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()+Pos),
954                    SourceMgr::DK_Error,
955                    CheckPrefix+"-NOT: string occurred!");
956    SM.PrintMessage(Pat->getLoc(), SourceMgr::DK_Note,
957                    CheckPrefix+"-NOT: pattern specified here");
958    return true;
959  }
960
961  return false;
962}
963
964size_t CheckString::CheckDag(const SourceMgr &SM, StringRef Buffer,
965                             std::vector<const Pattern *> &NotStrings,
966                             StringMap<StringRef> &VariableTable) const {
967  if (DagNotStrings.empty())
968    return 0;
969
970  size_t LastPos = 0;
971  size_t StartPos = LastPos;
972
973  for (unsigned ChunkNo = 0, e = DagNotStrings.size();
974       ChunkNo != e; ++ChunkNo) {
975    const Pattern &Pat = DagNotStrings[ChunkNo];
976
977    assert((Pat.getMatchDag() ^ Pat.getMatchNot()) &&
978           "Invalid CHECK-DAG or CHECK-NOT!");
979
980    if (Pat.getMatchNot()) {
981      NotStrings.push_back(&Pat);
982      continue;
983    }
984
985    assert(Pat.getMatchDag() && "Expect CHECK-DAG!");
986
987    size_t MatchLen = 0, MatchPos;
988
989    // CHECK-DAG always matches from the start.
990    StringRef MatchBuffer = Buffer.substr(StartPos);
991    MatchPos = Pat.Match(MatchBuffer, MatchLen, VariableTable);
992    // With a group of CHECK-DAGs, a single mismatching means the match on
993    // that group of CHECK-DAGs fails immediately.
994    if (MatchPos == StringRef::npos) {
995      PrintCheckFailed(SM, Pat.getLoc(), Pat, MatchBuffer, VariableTable);
996      return StringRef::npos;
997    }
998    // Re-calc it as the offset relative to the start of the original string.
999    MatchPos += StartPos;
1000
1001    if (!NotStrings.empty()) {
1002      if (MatchPos < LastPos) {
1003        // Reordered?
1004        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + MatchPos),
1005                        SourceMgr::DK_Error,
1006                        CheckPrefix+"-DAG: found a match of CHECK-DAG"
1007                        " reordering across a CHECK-NOT");
1008        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data() + LastPos),
1009                        SourceMgr::DK_Note,
1010                        CheckPrefix+"-DAG: the farthest match of CHECK-DAG"
1011                        " is found here");
1012        SM.PrintMessage(NotStrings[0]->getLoc(), SourceMgr::DK_Note,
1013                        CheckPrefix+"-NOT: the crossed pattern specified"
1014                        " here");
1015        SM.PrintMessage(Pat.getLoc(), SourceMgr::DK_Note,
1016                        CheckPrefix+"-DAG: the reordered pattern specified"
1017                        " here");
1018        return StringRef::npos;
1019      }
1020      // All subsequent CHECK-DAGs should be matched from the farthest
1021      // position of all precedent CHECK-DAGs (including this one.)
1022      StartPos = LastPos;
1023      // If there's CHECK-NOTs between two CHECK-DAGs or from CHECK to
1024      // CHECK-DAG, verify that there's no 'not' strings occurred in that
1025      // region.
1026      StringRef SkippedRegion = Buffer.substr(LastPos, MatchPos);
1027      if (CheckNot(SM, SkippedRegion, NotStrings, VariableTable))
1028        return StringRef::npos;
1029      // Clear "not strings".
1030      NotStrings.clear();
1031    }
1032
1033    // Update the last position with CHECK-DAG matches.
1034    LastPos = std::max(MatchPos + MatchLen, LastPos);
1035  }
1036
1037  return LastPos;
1038}
1039
1040bool ValidateCheckPrefix() {
1041  // The check prefix must contain only alphanumeric, hyphens and underscores.
1042  Regex prefixValidator("^[a-zA-Z0-9_-]*$");
1043  return prefixValidator.match(CheckPrefix);
1044}
1045
1046int main(int argc, char **argv) {
1047  sys::PrintStackTraceOnErrorSignal();
1048  PrettyStackTraceProgram X(argc, argv);
1049  cl::ParseCommandLineOptions(argc, argv);
1050
1051  if (!ValidateCheckPrefix()) {
1052    errs() << "Supplied check-prefix is invalid! Prefixes must start with a "
1053              "letter and contain only alphanumeric characters, hyphens and "
1054              "underscores\n";
1055    return 2;
1056  }
1057
1058  SourceMgr SM;
1059
1060  // Read the expected strings from the check file.
1061  std::vector<CheckString> CheckStrings;
1062  if (ReadCheckFile(SM, CheckStrings))
1063    return 2;
1064
1065  // Open the file to check and add it to SourceMgr.
1066  OwningPtr<MemoryBuffer> File;
1067  if (error_code ec =
1068        MemoryBuffer::getFileOrSTDIN(InputFilename, File)) {
1069    errs() << "Could not open input file '" << InputFilename << "': "
1070           << ec.message() << '\n';
1071    return 2;
1072  }
1073
1074  if (File->getBufferSize() == 0) {
1075    errs() << "FileCheck error: '" << InputFilename << "' is empty.\n";
1076    return 2;
1077  }
1078
1079  // Remove duplicate spaces in the input file if requested.
1080  // Remove DOS style line endings.
1081  MemoryBuffer *F =
1082    CanonicalizeInputFile(File.take(), NoCanonicalizeWhiteSpace);
1083
1084  SM.AddNewSourceBuffer(F, SMLoc());
1085
1086  /// VariableTable - This holds all the current filecheck variables.
1087  StringMap<StringRef> VariableTable;
1088
1089  // Check that we have all of the expected strings, in order, in the input
1090  // file.
1091  StringRef Buffer = F->getBuffer();
1092
1093  bool hasError = false;
1094
1095  unsigned i = 0, j = 0, e = CheckStrings.size();
1096
1097  while (true) {
1098    StringRef CheckRegion;
1099    if (j == e) {
1100      CheckRegion = Buffer;
1101    } else {
1102      const CheckString &CheckLabelStr = CheckStrings[j];
1103      if (!CheckLabelStr.IsCheckLabel) {
1104        ++j;
1105        continue;
1106      }
1107
1108      // Scan to next CHECK-LABEL match, ignoring CHECK-NOT and CHECK-DAG
1109      size_t MatchLabelLen = 0;
1110      size_t MatchLabelPos = CheckLabelStr.Check(SM, Buffer, true,
1111                                                 MatchLabelLen, VariableTable);
1112      if (MatchLabelPos == StringRef::npos) {
1113        hasError = true;
1114        break;
1115      }
1116
1117      CheckRegion = Buffer.substr(0, MatchLabelPos + MatchLabelLen);
1118      Buffer = Buffer.substr(MatchLabelPos + MatchLabelLen);
1119      ++j;
1120    }
1121
1122    for ( ; i != j; ++i) {
1123      const CheckString &CheckStr = CheckStrings[i];
1124
1125      // Check each string within the scanned region, including a second check
1126      // of any final CHECK-LABEL (to verify CHECK-NOT and CHECK-DAG)
1127      size_t MatchLen = 0;
1128      size_t MatchPos = CheckStr.Check(SM, CheckRegion, false, MatchLen,
1129                                       VariableTable);
1130
1131      if (MatchPos == StringRef::npos) {
1132        hasError = true;
1133        i = j;
1134        break;
1135      }
1136
1137      CheckRegion = CheckRegion.substr(MatchPos + MatchLen);
1138    }
1139
1140    if (j == e)
1141      break;
1142  }
1143
1144  return hasError ? 1 : 0;
1145}
1146