FileUtilities.cpp revision 599ddf9a7693a315a03fc77cf612eec53dee5f67 
1//===- Support/FileUtilities.cpp - File System Utilities ------------------===//
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// This file implements a family of utility functions which are useful for doing
11// various things with files.
12//
13//===----------------------------------------------------------------------===//
14
15#include "llvm/Support/FileUtilities.h"
16#include "llvm/System/Path.h"
17#include "llvm/System/MappedFile.h"
18#include "llvm/ADT/StringExtras.h"
19#include <cstring>
20#include <cctype>
21using namespace llvm;
22
23static bool isSignedChar(char C) {
24  if (C == '+' || C == '-')
25    return true;
26  else
27    return false;
28}
29
30static bool isExpoentChar(char C) {
31  switch (C) {
32  case 'D':  // Strange exponential notation.
33  case 'd':  // Strange exponential notation.
34  case 'e':
35  case 'E': return true;
36  default: return false;
37  }
38}
39
40static bool isNumberChar(char C) {
41  switch (C) {
42  case '0': case '1': case '2': case '3': case '4':
43  case '5': case '6': case '7': case '8': case '9':
44  case '.': return true;
45  default: return isSignedChar(C) || isExpoentChar(C);
46  }
47}
48
49static char *BackupNumber(char *Pos, char *FirstChar) {
50  // If we didn't stop in the middle of a number, don't backup.
51  if (!isNumberChar(*Pos)) return Pos;
52
53  // Otherwise, return to the start of the number.
54  while (Pos > FirstChar && isNumberChar(Pos[-1])) {
55    --Pos;
56    if (Pos > FirstChar && isSignedChar(Pos[0]) && !isExpoentChar(Pos[-1]))
57      break;
58  }
59  return Pos;
60}
61
62/// CompareNumbers - compare two numbers, returning true if they are different.
63static bool CompareNumbers(char *&F1P, char *&F2P, char *F1End, char *F2End,
64                           double AbsTolerance, double RelTolerance,
65                           std::string *ErrorMsg) {
66  char *F1NumEnd, *F2NumEnd;
67  double V1 = 0.0, V2 = 0.0;
68
69  // If one of the positions is at a space and the other isn't, chomp up 'til
70  // the end of the space.
71  while (isspace(*F1P) && F1P != F1End)
72    ++F1P;
73  while (isspace(*F2P) && F2P != F2End)
74    ++F2P;
75
76  // If we stop on numbers, compare their difference.  Note that some ugliness
77  // is built into this to permit support for numbers that use "D" or "d" as
78  // their exponential marker, e.g. "1.234D45".  This occurs in 200.sixtrack in
79  // spec2k.
80  if (isNumberChar(*F1P) && isNumberChar(*F2P)) {
81    bool isDNotation;
82    do {
83      isDNotation = false;
84      V1 = strtod(F1P, &F1NumEnd);
85      V2 = strtod(F2P, &F2NumEnd);
86
87      if (*F1NumEnd == 'D' || *F1NumEnd == 'd') {
88        *F1NumEnd = 'e';  // Strange exponential notation!
89        isDNotation = true;
90      }
91      if (*F2NumEnd == 'D' || *F2NumEnd == 'd') {
92        *F2NumEnd = 'e';  // Strange exponential notation!
93        isDNotation = true;
94      }
95    } while (isDNotation);
96  } else {
97    // Otherwise, the diff failed.
98    F1NumEnd = F1P;
99    F2NumEnd = F2P;
100  }
101
102  if (F1NumEnd == F1P || F2NumEnd == F2P) {
103    if (ErrorMsg) {
104      *ErrorMsg = "FP Comparison failed, not a numeric difference between '";
105      *ErrorMsg += F1P[0];
106      *ErrorMsg += "' and '";
107      *ErrorMsg += F2P[0];
108      *ErrorMsg += "'";
109    }
110    return true;
111  }
112
113  // Check to see if these are inside the absolute tolerance
114  if (AbsTolerance < std::abs(V1-V2)) {
115    // Nope, check the relative tolerance...
116    double Diff;
117    if (V2)
118      Diff = std::abs(V1/V2 - 1.0);
119    else if (V1)
120      Diff = std::abs(V2/V1 - 1.0);
121    else
122      Diff = 0;  // Both zero.
123    if (Diff > RelTolerance) {
124      if (ErrorMsg) {
125        *ErrorMsg = "Compared: " + ftostr(V1) + " and " + ftostr(V2) + "\n";
126        *ErrorMsg += "abs. diff = " + ftostr(std::abs(V1-V2)) +
127                     " rel.diff = " + ftostr(Diff) + "\n";
128        *ErrorMsg += "Out of tolerance: rel/abs: " + ftostr(RelTolerance) +
129                     "/" + ftostr(AbsTolerance);
130      }
131      return true;
132    }
133  }
134
135  // Otherwise, advance our read pointers to the end of the numbers.
136  F1P = F1NumEnd;  F2P = F2NumEnd;
137  return false;
138}
139
140// PadFileIfNeeded - If the files are not identical, we will have to be doing
141// numeric comparisons in here.  There are bad cases involved where we (i.e.,
142// strtod) might run off the beginning or end of the file if it starts or ends
143// with a number.  Because of this, if needed, we pad the file so that it starts
144// and ends with a null character.
145static void PadFileIfNeeded(char *&FileStart, char *&FileEnd, char *&FP) {
146  if (FileStart-FileEnd < 2 ||
147      isNumberChar(FileStart[0]) || isNumberChar(FileEnd[-1])) {
148    unsigned FileLen = FileEnd-FileStart;
149    char *NewFile = new char[FileLen+2];
150    NewFile[0] = 0;              // Add null padding
151    NewFile[FileLen+1] = 0;      // Add null padding
152    memcpy(NewFile+1, FileStart, FileLen);
153    FP = NewFile+(FP-FileStart)+1;
154    FileStart = NewFile+1;
155    FileEnd = FileStart+FileLen;
156  }
157}
158
159/// DiffFilesWithTolerance - Compare the two files specified, returning 0 if the
160/// files match, 1 if they are different, and 2 if there is a file error.  This
161/// function differs from DiffFiles in that you can specify an absolete and
162/// relative FP error that is allowed to exist.  If you specify a string to fill
163/// in for the error option, it will set the string to an error message if an
164/// error occurs, allowing the caller to distinguish between a failed diff and a
165/// file system error.
166///
167int llvm::DiffFilesWithTolerance(const sys::PathWithStatus &FileA,
168                                 const sys::PathWithStatus &FileB,
169                                 double AbsTol, double RelTol,
170                                 std::string *Error) {
171  const sys::FileStatus *FileAStat = FileA.getFileStatus(false, Error);
172  if (!FileAStat)
173    return 2;
174  const sys::FileStatus *FileBStat = FileB.getFileStatus(false, Error);
175  if (!FileBStat)
176    return 2;
177
178  // Check for zero length files because some systems croak when you try to
179  // mmap an empty file.
180  size_t A_size = FileAStat->getSize();
181  size_t B_size = FileBStat->getSize();
182
183  // If they are both zero sized then they're the same
184  if (A_size == 0 && B_size == 0)
185    return 0;
186
187  // If only one of them is zero sized then they can't be the same
188  if ((A_size == 0 || B_size == 0)) {
189    if (Error)
190      *Error = "Files differ: one is zero-sized, the other isn't";
191    return 1;
192  }
193
194  // Now its safe to mmap the files into memory becasue both files
195  // have a non-zero size.
196  sys::MappedFile F1;
197  if (F1.open(FileA, sys::MappedFile::READ_ACCESS, Error))
198    return 2;
199  sys::MappedFile F2;
200  if (F2.open(FileB, sys::MappedFile::READ_ACCESS, Error))
201    return 2;
202  if (!F1.map(Error))
203    return 2;
204  if (!F2.map(Error))
205    return 2;
206
207  // Okay, now that we opened the files, scan them for the first difference.
208  char *File1Start = F1.charBase();
209  char *File2Start = F2.charBase();
210  char *File1End = File1Start+A_size;
211  char *File2End = File2Start+B_size;
212  char *F1P = File1Start;
213  char *F2P = File2Start;
214
215  if (A_size == B_size) {
216    // Are the buffers identical?
217    if (std::memcmp(File1Start, File2Start, A_size) == 0)
218      return 0;
219
220    if (AbsTol == 0 && RelTol == 0) {
221      if (Error)
222        *Error = "Files differ without tolerance allowance";
223      return 1;   // Files different!
224    }
225  }
226
227  char *OrigFile1Start = File1Start;
228  char *OrigFile2Start = File2Start;
229
230  // If the files need padding, do so now.
231  PadFileIfNeeded(File1Start, File1End, F1P);
232  PadFileIfNeeded(File2Start, File2End, F2P);
233
234  bool CompareFailed = false;
235  while (1) {
236    // Scan for the end of file or next difference.
237    while (F1P < File1End && F2P < File2End && *F1P == *F2P)
238      ++F1P, ++F2P;
239
240    if (F1P >= File1End || F2P >= File2End) break;
241
242    // Okay, we must have found a difference.  Backup to the start of the
243    // current number each stream is at so that we can compare from the
244    // beginning.
245    F1P = BackupNumber(F1P, File1Start);
246    F2P = BackupNumber(F2P, File2Start);
247
248    // Now that we are at the start of the numbers, compare them, exiting if
249    // they don't match.
250    if (CompareNumbers(F1P, F2P, File1End, File2End, AbsTol, RelTol, Error)) {
251      CompareFailed = true;
252      break;
253    }
254  }
255
256  // Okay, we reached the end of file.  If both files are at the end, we
257  // succeeded.
258  bool F1AtEnd = F1P >= File1End;
259  bool F2AtEnd = F2P >= File2End;
260  if (!CompareFailed && (!F1AtEnd || !F2AtEnd)) {
261    // Else, we might have run off the end due to a number: backup and retry.
262    if (F1AtEnd && isNumberChar(F1P[-1])) --F1P;
263    if (F2AtEnd && isNumberChar(F2P[-1])) --F2P;
264    F1P = BackupNumber(F1P, File1Start);
265    F2P = BackupNumber(F2P, File2Start);
266
267    // Now that we are at the start of the numbers, compare them, exiting if
268    // they don't match.
269    if (CompareNumbers(F1P, F2P, File1End, File2End, AbsTol, RelTol, Error))
270      CompareFailed = true;
271
272    // If we found the end, we succeeded.
273    if (F1P < File1End || F2P < File2End)
274      CompareFailed = true;
275  }
276
277  if (OrigFile1Start != File1Start)
278    delete[] (File1Start-1);   // Back up past null byte
279  if (OrigFile2Start != File2Start)
280    delete[] (File2Start-1);   // Back up past null byte
281  return CompareFailed;
282}
283