FileCheck.cpp revision 9fc6678bea11f32acba4e9b6624ae9a0d3a9287b 
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/Support/CommandLine.h"
20#include "llvm/Support/MemoryBuffer.h"
21#include "llvm/Support/PrettyStackTrace.h"
22#include "llvm/Support/SourceMgr.h"
23#include "llvm/Support/raw_ostream.h"
24#include "llvm/System/Signals.h"
25using namespace llvm;
26
27static cl::opt<std::string>
28CheckFilename(cl::Positional, cl::desc("<check-file>"), cl::Required);
29
30static cl::opt<std::string>
31InputFilename("input-file", cl::desc("File to check (defaults to stdin)"),
32              cl::init("-"), cl::value_desc("filename"));
33
34static cl::opt<std::string>
35CheckPrefix("check-prefix", cl::init("CHECK"),
36            cl::desc("Prefix to use from check file (defaults to 'CHECK')"));
37
38static cl::opt<bool>
39NoCanonicalizeWhiteSpace("strict-whitespace",
40              cl::desc("Do not treat all horizontal whitespace as equivalent"));
41
42class Pattern {
43  /// Str - The string to match.
44  std::string Str;
45public:
46
47  Pattern(StringRef S) : Str(S.str()) {
48    // Remove duplicate spaces in the check strings if requested.
49    if (!NoCanonicalizeWhiteSpace)
50      CanonicalizeCheckString();
51  }
52
53  /// Match - Match the pattern string against the input buffer Buffer.  This
54  /// returns the position that is matched or npos if there is no match.  If
55  /// there is a match, the size of the matched string is returned in MatchLen.
56  size_t Match(StringRef Buffer, size_t &MatchLen) const {
57    MatchLen = Str.size();
58    return Buffer.find(Str);
59  }
60
61private:
62  /// CanonicalizeCheckString - Replace all sequences of horizontal whitespace
63  /// in the check strings with a single space.
64  void CanonicalizeCheckString() {
65    for (unsigned C = 0; C != Str.size(); ++C) {
66      // If C is not a horizontal whitespace, skip it.
67      if (Str[C] != ' ' && Str[C] != '\t')
68        continue;
69
70      // Replace the character with space, then remove any other space
71      // characters after it.
72      Str[C] = ' ';
73
74      while (C+1 != Str.size() &&
75             (Str[C+1] == ' ' || Str[C+1] == '\t'))
76        Str.erase(Str.begin()+C+1);
77    }
78  }
79};
80
81
82/// CheckString - This is a check that we found in the input file.
83struct CheckString {
84  /// Pat - The pattern to match.
85  Pattern Pat;
86
87  /// Loc - The location in the match file that the check string was specified.
88  SMLoc Loc;
89
90  /// IsCheckNext - This is true if this is a CHECK-NEXT: directive (as opposed
91  /// to a CHECK: directive.
92  bool IsCheckNext;
93
94  /// NotStrings - These are all of the strings that are disallowed from
95  /// occurring between this match string and the previous one (or start of
96  /// file).
97  std::vector<std::pair<SMLoc, std::string> > NotStrings;
98
99  CheckString(const Pattern &P, SMLoc L, bool isCheckNext)
100    : Pat(P), Loc(L), IsCheckNext(isCheckNext) {}
101};
102
103
104/// ReadCheckFile - Read the check file, which specifies the sequence of
105/// expected strings.  The strings are added to the CheckStrings vector.
106static bool ReadCheckFile(SourceMgr &SM,
107                          std::vector<CheckString> &CheckStrings) {
108  // Open the check file, and tell SourceMgr about it.
109  std::string ErrorStr;
110  MemoryBuffer *F =
111    MemoryBuffer::getFileOrSTDIN(CheckFilename.c_str(), &ErrorStr);
112  if (F == 0) {
113    errs() << "Could not open check file '" << CheckFilename << "': "
114           << ErrorStr << '\n';
115    return true;
116  }
117  SM.AddNewSourceBuffer(F, SMLoc());
118
119  // Find all instances of CheckPrefix followed by : in the file.
120  StringRef Buffer = F->getBuffer();
121
122  std::vector<std::pair<SMLoc, std::string> > NotMatches;
123
124  while (1) {
125    // See if Prefix occurs in the memory buffer.
126    Buffer = Buffer.substr(Buffer.find(CheckPrefix));
127
128    // If we didn't find a match, we're done.
129    if (Buffer.empty())
130      break;
131
132    const char *CheckPrefixStart = Buffer.data();
133
134    // When we find a check prefix, keep track of whether we find CHECK: or
135    // CHECK-NEXT:
136    bool IsCheckNext = false, IsCheckNot = false;
137
138    // Verify that the : is present after the prefix.
139    if (Buffer[CheckPrefix.size()] == ':') {
140      Buffer = Buffer.substr(CheckPrefix.size()+1);
141    } else if (Buffer.size() > CheckPrefix.size()+6 &&
142               memcmp(Buffer.data()+CheckPrefix.size(), "-NEXT:", 6) == 0) {
143      Buffer = Buffer.substr(CheckPrefix.size()+7);
144      IsCheckNext = true;
145    } else if (Buffer.size() > CheckPrefix.size()+5 &&
146               memcmp(Buffer.data()+CheckPrefix.size(), "-NOT:", 5) == 0) {
147      Buffer = Buffer.substr(CheckPrefix.size()+6);
148      IsCheckNot = true;
149    } else {
150      Buffer = Buffer.substr(1);
151      continue;
152    }
153
154    // Okay, we found the prefix, yay.  Remember the rest of the line, but
155    // ignore leading and trailing whitespace.
156    Buffer = Buffer.substr(Buffer.find_first_not_of(" \t"));
157
158    // Scan ahead to the end of line.
159    size_t EOL = Buffer.find_first_of("\n\r");
160    if (EOL == StringRef::npos) EOL = Buffer.size();
161
162    // Ignore trailing whitespace.
163    while (EOL && (Buffer[EOL-1] == ' ' || Buffer[EOL-1] == '\t'))
164      --EOL;
165
166    // Check that there is something on the line.
167    if (EOL == 0) {
168      SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()),
169                      "found empty check string with prefix '"+CheckPrefix+":'",
170                      "error");
171      return true;
172    }
173
174    StringRef PatternStr = Buffer.substr(0, EOL);
175
176    // Handle CHECK-NOT.
177    if (IsCheckNot) {
178      NotMatches.push_back(std::make_pair(SMLoc::getFromPointer(Buffer.data()),
179                                          PatternStr.str()));
180      Buffer = Buffer.substr(EOL);
181      continue;
182    }
183
184    // Verify that CHECK-NEXT lines have at least one CHECK line before them.
185    if (IsCheckNext && CheckStrings.empty()) {
186      SM.PrintMessage(SMLoc::getFromPointer(CheckPrefixStart),
187                      "found '"+CheckPrefix+"-NEXT:' without previous '"+
188                      CheckPrefix+ ": line", "error");
189      return true;
190    }
191
192    Pattern P(PatternStr);
193
194    // Okay, add the string we captured to the output vector and move on.
195    CheckStrings.push_back(CheckString(P,
196                                       SMLoc::getFromPointer(Buffer.data()),
197                                       IsCheckNext));
198    std::swap(NotMatches, CheckStrings.back().NotStrings);
199
200    Buffer = Buffer.substr(EOL);
201  }
202
203  if (CheckStrings.empty()) {
204    errs() << "error: no check strings found with prefix '" << CheckPrefix
205           << ":'\n";
206    return true;
207  }
208
209  if (!NotMatches.empty()) {
210    errs() << "error: '" << CheckPrefix
211           << "-NOT:' not supported after last check line.\n";
212    return true;
213  }
214
215  return false;
216}
217
218/// CanonicalizeInputFile - Remove duplicate horizontal space from the specified
219/// memory buffer, free it, and return a new one.
220static MemoryBuffer *CanonicalizeInputFile(MemoryBuffer *MB) {
221  SmallVector<char, 16> NewFile;
222  NewFile.reserve(MB->getBufferSize());
223
224  for (const char *Ptr = MB->getBufferStart(), *End = MB->getBufferEnd();
225       Ptr != End; ++Ptr) {
226    // If C is not a horizontal whitespace, skip it.
227    if (*Ptr != ' ' && *Ptr != '\t') {
228      NewFile.push_back(*Ptr);
229      continue;
230    }
231
232    // Otherwise, add one space and advance over neighboring space.
233    NewFile.push_back(' ');
234    while (Ptr+1 != End &&
235           (Ptr[1] == ' ' || Ptr[1] == '\t'))
236      ++Ptr;
237  }
238
239  // Free the old buffer and return a new one.
240  MemoryBuffer *MB2 =
241    MemoryBuffer::getMemBufferCopy(NewFile.data(),
242                                   NewFile.data() + NewFile.size(),
243                                   MB->getBufferIdentifier());
244
245  delete MB;
246  return MB2;
247}
248
249
250static void PrintCheckFailed(const SourceMgr &SM, const CheckString &CheckStr,
251                             StringRef Buffer) {
252  // Otherwise, we have an error, emit an error message.
253  SM.PrintMessage(CheckStr.Loc, "expected string not found in input",
254                  "error");
255
256  // Print the "scanning from here" line.  If the current position is at the
257  // end of a line, advance to the start of the next line.
258  Buffer = Buffer.substr(Buffer.find_first_not_of(" \t\n\r"));
259
260  SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()), "scanning from here",
261                  "note");
262}
263
264/// CountNumNewlinesBetween - Count the number of newlines in the specified
265/// range.
266static unsigned CountNumNewlinesBetween(StringRef Range) {
267  unsigned NumNewLines = 0;
268  while (1) {
269    // Scan for newline.
270    Range = Range.substr(Range.find_first_of("\n\r"));
271    if (Range.empty()) return NumNewLines;
272
273    ++NumNewLines;
274
275    // Handle \n\r and \r\n as a single newline.
276    if (Range.size() > 1 &&
277        (Range[1] == '\n' || Range[1] == '\r') &&
278        (Range[0] != Range[1]))
279      Range = Range.substr(1);
280    Range = Range.substr(1);
281  }
282}
283
284int main(int argc, char **argv) {
285  sys::PrintStackTraceOnErrorSignal();
286  PrettyStackTraceProgram X(argc, argv);
287  cl::ParseCommandLineOptions(argc, argv);
288
289  SourceMgr SM;
290
291  // Read the expected strings from the check file.
292  std::vector<CheckString> CheckStrings;
293  if (ReadCheckFile(SM, CheckStrings))
294    return 2;
295
296  // Open the file to check and add it to SourceMgr.
297  std::string ErrorStr;
298  MemoryBuffer *F =
299    MemoryBuffer::getFileOrSTDIN(InputFilename.c_str(), &ErrorStr);
300  if (F == 0) {
301    errs() << "Could not open input file '" << InputFilename << "': "
302           << ErrorStr << '\n';
303    return true;
304  }
305
306  // Remove duplicate spaces in the input file if requested.
307  if (!NoCanonicalizeWhiteSpace)
308    F = CanonicalizeInputFile(F);
309
310  SM.AddNewSourceBuffer(F, SMLoc());
311
312  // Check that we have all of the expected strings, in order, in the input
313  // file.
314  StringRef Buffer = F->getBuffer();
315
316  const char *LastMatch = Buffer.data();
317
318  for (unsigned StrNo = 0, e = CheckStrings.size(); StrNo != e; ++StrNo) {
319    const CheckString &CheckStr = CheckStrings[StrNo];
320
321    StringRef SearchFrom = Buffer;
322
323    // Find StrNo in the file.
324    size_t MatchLen = 0;
325    Buffer = Buffer.substr(CheckStr.Pat.Match(Buffer, MatchLen));
326
327    // If we didn't find a match, reject the input.
328    if (Buffer.empty()) {
329      PrintCheckFailed(SM, CheckStr, SearchFrom);
330      return 1;
331    }
332
333    StringRef SkippedRegion(LastMatch, Buffer.data()-LastMatch);
334
335    // If this check is a "CHECK-NEXT", verify that the previous match was on
336    // the previous line (i.e. that there is one newline between them).
337    if (CheckStr.IsCheckNext) {
338      // Count the number of newlines between the previous match and this one.
339      assert(LastMatch != F->getBufferStart() &&
340             "CHECK-NEXT can't be the first check in a file");
341
342      unsigned NumNewLines = CountNumNewlinesBetween(SkippedRegion);
343      if (NumNewLines == 0) {
344        SM.PrintMessage(CheckStr.Loc,
345                    CheckPrefix+"-NEXT: is on the same line as previous match",
346                        "error");
347        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()),
348                        "'next' match was here", "note");
349        SM.PrintMessage(SMLoc::getFromPointer(LastMatch),
350                        "previous match was here", "note");
351        return 1;
352      }
353
354      if (NumNewLines != 1) {
355        SM.PrintMessage(CheckStr.Loc,
356                        CheckPrefix+
357                        "-NEXT: is not on the line after the previous match",
358                        "error");
359        SM.PrintMessage(SMLoc::getFromPointer(Buffer.data()),
360                        "'next' match was here", "note");
361        SM.PrintMessage(SMLoc::getFromPointer(LastMatch),
362                        "previous match was here", "note");
363        return 1;
364      }
365    }
366
367    // If this match had "not strings", verify that they don't exist in the
368    // skipped region.
369    for (unsigned i = 0, e = CheckStr.NotStrings.size(); i != e; ++i) {
370      size_t Pos = SkippedRegion.find(CheckStr.NotStrings[i].second);
371      if (Pos == StringRef::npos) continue;
372
373      SM.PrintMessage(SMLoc::getFromPointer(LastMatch+Pos),
374                      CheckPrefix+"-NOT: string occurred!", "error");
375      SM.PrintMessage(CheckStr.NotStrings[i].first,
376                      CheckPrefix+"-NOT: pattern specified here", "note");
377      return 1;
378    }
379
380
381    // Otherwise, everything is good.  Step over the matched text and remember
382    // the position after the match as the end of the last match.
383    Buffer = Buffer.substr(MatchLen);
384    LastMatch = Buffer.data();
385  }
386
387  return 0;
388}
389