llvm-profdata: Update to use the naive text format with function hash

This also uses line_iterator to simplify the parsing logic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204210 91177308-0d34-0410-b5e6-96231b3b80d8

1 files changed, 53 insertions, 101 deletions

diff --git a/tools/llvm-profdata/llvm-profdata.cpp b/tools/llvm-profdata/llvm-profdata.cpp
index b338efd27e..b4dd55c833 100644
--- a/tools/llvm-profdata/llvm-profdata.cpp
+++ b/tools/llvm-profdata/llvm-profdata.cpp
@@ -13,6 +13,7 @@
 
 #include "llvm/ADT/StringRef.h"
 #include "llvm/Support/CommandLine.h"
+#include "llvm/Support/LineIterator.h"
 #include "llvm/Support/ManagedStatic.h"
 #include "llvm/Support/MemoryBuffer.h"
 #include "llvm/Support/PrettyStackTrace.h"
@@ -32,56 +33,6 @@ static cl::opt<std::string> OutputFilename("output", cl::value_desc("output"),
 static cl::alias OutputFilenameA("o", cl::desc("Alias for --output"),
                                  cl::aliasopt(OutputFilename));
 
-static bool readLine(const char *&Start, const char *End, StringRef &S) {
-  if (Start == End)
-    return false;
-
-  for (const char *I = Start; I != End; ++I) {
-    assert(*I && "unexpected binary data");
-    if (*I == '\n') {
-      S = StringRef(Start, I - Start);
-      Start = I + 1;
-      return true;
-    }
-  }
-
-  S = StringRef(Start, End - Start);
-  Start = End;
-  return true;
-}
-
-static StringRef getWord(const char *&Start, const char *End) {
-  for (const char *I = Start; I != End; ++I)
-    if (*I == ' ') {
-      StringRef S(Start, I - Start);
-      Start = I + 1;
-      return S;
-    }
-  StringRef S(Start, End - Start);
-  Start = End;
-  return S;
-}
-
-static size_t splitWords(const StringRef &Line, std::vector<StringRef> &Words) {
-  const char *Start = Line.data();
-  const char *End = Line.data() + Line.size();
-  Words.clear();
-  while (Start != End)
-    Words.push_back(getWord(Start, End));
-  return Words.size();
-}
-
-static bool getNumber(const StringRef &S, uint64_t &N) {
-  N = 0;
-  for (StringRef::iterator I = S.begin(), E = S.end(); I != E; ++I)
-    if (*I >= '0' && *I <= '9')
-      N = N * 10 + (*I - '0');
-    else
-      return false;
-
-  return true;
-}
-
 static void exitWithError(const std::string &Message,
                           const std::string &Filename, int64_t Line = -1) {
   errs() << "error: " << Filename;
@@ -115,63 +66,64 @@ int main(int argc, char **argv) {
   if (!ErrorInfo.empty())
     exitWithError(ErrorInfo, OutputFilename);
 
-  const char *Start1 = File1->getBufferStart();
-  const char *Start2 = File2->getBufferStart();
-  const char *End1 = File1->getBufferEnd();
-  const char *End2 = File2->getBufferEnd();
-  const char *P1 = Start1;
-  const char *P2 = Start2;
-
-  StringRef Line1, Line2;
-  int64_t Num = 0;
-  while (readLine(P1, End1, Line1)) {
-    ++Num;
-    if (!readLine(P2, End2, Line2))
-      exitWithError("truncated file", Filename2, Num);
-
-    std::vector<StringRef> Words1, Words2;
-    if (splitWords(Line1, Words1) != splitWords(Line2, Words2))
-      exitWithError("data mismatch", Filename2, Num);
-
-    if (Words1.size() > 2)
-      exitWithError("invalid data", Filename1, Num);
-
-    if (Words1.empty()) {
+  enum {ReadName, ReadHash, ReadCount, ReadCounters} State = ReadName;
+  uint64_t N1, N2, NumCounters;
+  line_iterator I1(*File1, '#'), I2(*File2, '#');
+  for (; !I1.is_at_end() && !I2.is_at_end(); ++I1, ++I2) {
+    if (I1->empty()) {
+      if (!I2->empty())
+        exitWithError("data mismatch", Filename2, I2.line_number());
       Output << "\n";
       continue;
     }
-
-    if (Words1.size() == 2) {
-      if (Words1[0] != Words2[0])
-        exitWithError("function name mismatch", Filename2, Num);
-
-      uint64_t N1, N2;
-      if (!getNumber(Words1[1], N1))
-        exitWithError("bad function count", Filename1, Num);
-      if (!getNumber(Words2[1], N2))
-        exitWithError("bad function count", Filename2, Num);
-
+    switch (State) {
+    case ReadName:
+      if (*I1 != *I2)
+        exitWithError("function name mismatch", Filename2, I2.line_number());
+      Output << *I1 << "\n";
+      State = ReadHash;
+      break;
+    case ReadHash:
+      if (I1->getAsInteger(10, N1))
+        exitWithError("bad function hash", Filename1, I1.line_number());
+      if (I2->getAsInteger(10, N2))
+        exitWithError("bad function hash", Filename2, I2.line_number());
       if (N1 != N2)
-        exitWithError("function count mismatch", Filename2, Num);
-
-      Output << Line1 << "\n";
-      continue;
+        exitWithError("function hash mismatch", Filename2, I2.line_number());
+      Output << N1 << "\n";
+      State = ReadCount;
+      break;
+    case ReadCount:
+      if (I1->getAsInteger(10, N1))
+        exitWithError("bad function count", Filename1, I1.line_number());
+      if (I2->getAsInteger(10, N2))
+        exitWithError("bad function count", Filename2, I2.line_number());
+      if (N1 != N2)
+        exitWithError("function count mismatch", Filename2, I2.line_number());
+      Output << N1 << "\n";
+      NumCounters = N1;
+      State = ReadCounters;
+      break;
+    case ReadCounters:
+      if (I1->getAsInteger(10, N1))
+        exitWithError("invalid counter", Filename1, I1.line_number());
+      if (I2->getAsInteger(10, N2))
+        exitWithError("invalid counter", Filename2, I2.line_number());
+      uint64_t Sum = N1 + N2;
+      if (Sum < N1)
+        exitWithError("counter overflow", Filename2, I2.line_number());
+      Output << N1 + N2 << "\n";
+      if (--NumCounters == 0)
+        State = ReadName;
+      break;
     }
-
-    uint64_t N1, N2;
-    if (!getNumber(Words1[0], N1))
-      exitWithError("invalid counter", Filename1, Num);
-    if (!getNumber(Words2[0], N2))
-      exitWithError("invalid counter", Filename2, Num);
-
-    uint64_t Sum = N1 + N2;
-    if (Sum < N1)
-      exitWithError("counter overflow", Filename2, Num);
-
-    Output << N1 + N2 << "\n";
   }
-  if (readLine(P2, End2, Line2))
-    exitWithError("truncated file", Filename1, Num + 1);
+  if (!I1.is_at_end())
+    exitWithError("truncated file", Filename1, I1.line_number());
+  if (!I2.is_at_end())
+    exitWithError("truncated file", Filename2, I2.line_number());
+  if (State != ReadName)
+    exitWithError("truncated file", Filename1, I1.line_number());
 
   return 0;
 }