//
// The LLVM Compiler Infrastructure
//
-// This file was developed by the LLVM research group and is distributed under
-// the University of Illinois Open Source License. See LICENSE.TXT for details.
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
//
//===----------------------------------------------------------------------===//
#include "llvm/Support/FileUtilities.h"
-#include "llvm/System/Path.h"
-#include "llvm/System/MappedFile.h"
-#include "llvm/ADT/StringExtras.h"
-#include <cstring>
+#include "llvm/ADT/SmallString.h"
+#include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/Path.h"
+#include "llvm/Support/raw_ostream.h"
#include <cctype>
+#include <cstdlib>
+#include <cstring>
+#include <system_error>
using namespace llvm;
-static bool isNumberChar(char C) {
+static bool isSignedChar(char C) {
+ return (C == '+' || C == '-');
+}
+
+static bool isExponentChar(char C) {
switch (C) {
- case '0': case '1': case '2': case '3': case '4':
- case '5': case '6': case '7': case '8': case '9':
- case '.': case '+': case '-':
case 'D': // Strange exponential notation.
case 'd': // Strange exponential notation.
case 'e':
}
}
-static char *BackupNumber(char *Pos, char *FirstChar) {
+static bool isNumberChar(char C) {
+ switch (C) {
+ case '0': case '1': case '2': case '3': case '4':
+ case '5': case '6': case '7': case '8': case '9':
+ case '.': return true;
+ default: return isSignedChar(C) || isExponentChar(C);
+ }
+}
+
+static const char *BackupNumber(const char *Pos, const char *FirstChar) {
// If we didn't stop in the middle of a number, don't backup.
if (!isNumberChar(*Pos)) return Pos;
// Otherwise, return to the start of the number.
- while (Pos > FirstChar && isNumberChar(Pos[-1]))
+ bool HasPeriod = false;
+ while (Pos > FirstChar && isNumberChar(Pos[-1])) {
+ // Backup over at most one period.
+ if (Pos[-1] == '.') {
+ if (HasPeriod)
+ break;
+ HasPeriod = true;
+ }
+
--Pos;
+ if (Pos > FirstChar && isSignedChar(Pos[0]) && !isExponentChar(Pos[-1]))
+ break;
+ }
+ return Pos;
+}
+
+/// EndOfNumber - Return the first character that is not part of the specified
+/// number. This assumes that the buffer is null terminated, so it won't fall
+/// off the end.
+static const char *EndOfNumber(const char *Pos) {
+ while (isNumberChar(*Pos))
+ ++Pos;
return Pos;
}
/// CompareNumbers - compare two numbers, returning true if they are different.
-static bool CompareNumbers(char *&F1P, char *&F2P, char *F1End, char *F2End,
+static bool CompareNumbers(const char *&F1P, const char *&F2P,
+ const char *F1End, const char *F2End,
double AbsTolerance, double RelTolerance,
std::string *ErrorMsg) {
- char *F1NumEnd, *F2NumEnd;
+ const char *F1NumEnd, *F2NumEnd;
double V1 = 0.0, V2 = 0.0;
// If one of the positions is at a space and the other isn't, chomp up 'til
// the end of the space.
- while (isspace(*F1P) && F1P != F1End)
+ while (isspace(static_cast<unsigned char>(*F1P)) && F1P != F1End)
++F1P;
- while (isspace(*F2P) && F2P != F2End)
+ while (isspace(static_cast<unsigned char>(*F2P)) && F2P != F2End)
++F2P;
- // If we stop on numbers, compare their difference. Note that some ugliness
- // is built into this to permit support for numbers that use "D" or "d" as
- // their exponential marker, e.g. "1.234D45". This occurs in 200.sixtrack in
- // spec2k.
- if (isNumberChar(*F1P) && isNumberChar(*F2P)) {
- bool isDNotation;
- do {
- isDNotation = false;
- V1 = strtod(F1P, &F1NumEnd);
- V2 = strtod(F2P, &F2NumEnd);
-
- if (*F1NumEnd == 'D' || *F1NumEnd == 'd') {
- *F1NumEnd = 'e'; // Strange exponential notation!
- isDNotation = true;
- }
- if (*F2NumEnd == 'D' || *F2NumEnd == 'd') {
- *F2NumEnd = 'e'; // Strange exponential notation!
- isDNotation = true;
- }
- } while (isDNotation);
- } else {
- // Otherwise, the diff failed.
+ // If we stop on numbers, compare their difference.
+ if (!isNumberChar(*F1P) || !isNumberChar(*F2P)) {
+ // The diff failed.
F1NumEnd = F1P;
F2NumEnd = F2P;
+ } else {
+ // Note that some ugliness is built into this to permit support for numbers
+ // that use "D" or "d" as their exponential marker, e.g. "1.234D45". This
+ // occurs in 200.sixtrack in spec2k.
+ V1 = strtod(F1P, const_cast<char**>(&F1NumEnd));
+ V2 = strtod(F2P, const_cast<char**>(&F2NumEnd));
+
+ if (*F1NumEnd == 'D' || *F1NumEnd == 'd') {
+ // Copy string into tmp buffer to replace the 'D' with an 'e'.
+ SmallString<200> StrTmp(F1P, EndOfNumber(F1NumEnd)+1);
+ // Strange exponential notation!
+ StrTmp[static_cast<unsigned>(F1NumEnd-F1P)] = 'e';
+
+ V1 = strtod(&StrTmp[0], const_cast<char**>(&F1NumEnd));
+ F1NumEnd = F1P + (F1NumEnd-&StrTmp[0]);
+ }
+
+ if (*F2NumEnd == 'D' || *F2NumEnd == 'd') {
+ // Copy string into tmp buffer to replace the 'D' with an 'e'.
+ SmallString<200> StrTmp(F2P, EndOfNumber(F2NumEnd)+1);
+ // Strange exponential notation!
+ StrTmp[static_cast<unsigned>(F2NumEnd-F2P)] = 'e';
+
+ V2 = strtod(&StrTmp[0], const_cast<char**>(&F2NumEnd));
+ F2NumEnd = F2P + (F2NumEnd-&StrTmp[0]);
+ }
}
if (F1NumEnd == F1P || F2NumEnd == F2P) {
Diff = 0; // Both zero.
if (Diff > RelTolerance) {
if (ErrorMsg) {
- *ErrorMsg = "Compared: " + ftostr(V1) + " and " + ftostr(V2) + "\n";
- *ErrorMsg += "abs. diff = " + ftostr(std::abs(V1-V2)) +
- " rel.diff = " + ftostr(Diff) + "\n";
- *ErrorMsg += "Out of tolerance: rel/abs: " + ftostr(RelTolerance) +
- "/" + ftostr(AbsTolerance);
+ raw_string_ostream(*ErrorMsg)
+ << "Compared: " << V1 << " and " << V2 << '\n'
+ << "abs. diff = " << std::abs(V1-V2) << " rel.diff = " << Diff << '\n'
+ << "Out of tolerance: rel/abs: " << RelTolerance << '/'
+ << AbsTolerance;
}
return true;
}
return false;
}
-// PadFileIfNeeded - If the files are not identical, we will have to be doing
-// numeric comparisons in here. There are bad cases involved where we (i.e.,
-// strtod) might run off the beginning or end of the file if it starts or ends
-// with a number. Because of this, if needed, we pad the file so that it starts
-// and ends with a null character.
-static void PadFileIfNeeded(char *&FileStart, char *&FileEnd, char *&FP) {
- if (FileStart-FileEnd < 2 ||
- isNumberChar(FileStart[0]) || isNumberChar(FileEnd[-1])) {
- unsigned FileLen = FileEnd-FileStart;
- char *NewFile = new char[FileLen+2];
- NewFile[0] = 0; // Add null padding
- NewFile[FileLen+1] = 0; // Add null padding
- memcpy(NewFile+1, FileStart, FileLen);
- FP = NewFile+(FP-FileStart)+1;
- FileStart = NewFile+1;
- FileEnd = FileStart+FileLen;
- }
-}
-
/// DiffFilesWithTolerance - Compare the two files specified, returning 0 if the
/// files match, 1 if they are different, and 2 if there is a file error. This
/// function differs from DiffFiles in that you can specify an absolete and
/// error occurs, allowing the caller to distinguish between a failed diff and a
/// file system error.
///
-int llvm::DiffFilesWithTolerance(const sys::Path &FileA,
- const sys::Path &FileB,
+int llvm::DiffFilesWithTolerance(StringRef NameA,
+ StringRef NameB,
double AbsTol, double RelTol,
std::string *Error) {
- const sys::FileStatus *FileAStat = FileA.getFileStatus(false, Error);
- if (!FileAStat)
- return 2;
- const sys::FileStatus *FileBStat = FileB.getFileStatus(false, Error);
- if (!FileBStat)
- return 2;
-
- // Check for zero length files because some systems croak when you try to
- // mmap an empty file.
- size_t A_size = FileAStat->getSize();
- size_t B_size = FileBStat->getSize();
-
- // If they are both zero sized then they're the same
- if (A_size == 0 && B_size == 0)
- return 0;
-
- // If only one of them is zero sized then they can't be the same
- if ((A_size == 0 || B_size == 0)) {
+ // Now its safe to mmap the files into memory because both files
+ // have a non-zero size.
+ ErrorOr<std::unique_ptr<MemoryBuffer>> F1OrErr = MemoryBuffer::getFile(NameA);
+ if (std::error_code EC = F1OrErr.getError()) {
if (Error)
- *Error = "Files differ: one is zero-sized, the other isn't";
- return 1;
+ *Error = EC.message();
+ return 2;
}
+ MemoryBuffer &F1 = *F1OrErr.get();
- // Now its safe to mmap the files into memory becasue both files
- // have a non-zero size.
- sys::MappedFile F1;
- if (F1.open(FileA, sys::MappedFile::READ_ACCESS, Error))
- return 2;
- sys::MappedFile F2;
- if (F2.open(FileB, sys::MappedFile::READ_ACCESS, Error))
- return 2;
- if (!F1.map(Error))
- return 2;
- if (!F2.map(Error))
+ ErrorOr<std::unique_ptr<MemoryBuffer>> F2OrErr = MemoryBuffer::getFile(NameB);
+ if (std::error_code EC = F2OrErr.getError()) {
+ if (Error)
+ *Error = EC.message();
return 2;
+ }
+ MemoryBuffer &F2 = *F2OrErr.get();
// Okay, now that we opened the files, scan them for the first difference.
- char *File1Start = F1.charBase();
- char *File2Start = F2.charBase();
- char *File1End = File1Start+A_size;
- char *File2End = File2Start+B_size;
- char *F1P = File1Start;
- char *F2P = File2Start;
+ const char *File1Start = F1.getBufferStart();
+ const char *File2Start = F2.getBufferStart();
+ const char *File1End = F1.getBufferEnd();
+ const char *File2End = F2.getBufferEnd();
+ const char *F1P = File1Start;
+ const char *F2P = File2Start;
+ uint64_t A_size = F1.getBufferSize();
+ uint64_t B_size = F2.getBufferSize();
- if (A_size == B_size) {
- // Are the buffers identical?
- if (std::memcmp(File1Start, File2Start, A_size) == 0)
- return 0;
+ // Are the buffers identical? Common case: Handle this efficiently.
+ if (A_size == B_size &&
+ std::memcmp(File1Start, File2Start, A_size) == 0)
+ return 0;
- if (AbsTol == 0 && RelTol == 0) {
- if (Error)
- *Error = "Files differ without tolerance allowance";
- return 1; // Files different!
- }
+ // Otherwise, we are done a tolerances are set.
+ if (AbsTol == 0 && RelTol == 0) {
+ if (Error)
+ *Error = "Files differ without tolerance allowance";
+ return 1; // Files different!
}
- char *OrigFile1Start = File1Start;
- char *OrigFile2Start = File2Start;
-
- // If the files need padding, do so now.
- PadFileIfNeeded(File1Start, File1End, F1P);
- PadFileIfNeeded(File2Start, File2End, F2P);
-
bool CompareFailed = false;
while (1) {
// Scan for the end of file or next difference.
CompareFailed = true;
}
- if (OrigFile1Start != File1Start)
- delete[] (File1Start-1); // Back up past null byte
- if (OrigFile2Start != File2Start)
- delete[] (File2Start-1); // Back up past null byte
return CompareFailed;
}