ELFObjectWriter: deduplicate suffices in strtab

We already do this for shstrtab, so might as well do it for strtab. This
extracts the string table building code into a separate class. The idea
is to use it for other object formats too.

I mostly wanted to do this for the general principle, but it does save a
little bit on object file size. I tried this on a clang bootstrap and
saved 0.54% on the sum of object file sizes (1.14 MB out of 212 MB for
a release build).

Differential Revision: http://reviews.llvm.org/D3533

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@207670 91177308-0d34-0410-b5e6-96231b3b80d8

include/llvm/Object/StringTableBuilder.h

@@ -0,0 +1,59 @@
+//===-- StringTableBuilder.h - String table building utility ------*- C++ -*-=//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_OBJECT_STRINGTABLE_BUILDER_H
+#define LLVM_OBJECT_STRINGTABLE_BUILDER_H
+
+#include "llvm/ADT/SmallString.h"
+#include "llvm/ADT/StringMap.h"
+#include <cassert>
+
+namespace llvm {
+
+/// \brief Utility for building string tables with deduplicated suffixes.
+class StringTableBuilder {
+  SmallString<256> StringTable;
+  StringMap<size_t> StringIndexMap;
+
+public:
+  /// \brief Add a string to the builder. Returns a StringRef to the internal
+  /// copy of s. Can only be used before the table is finalized.
+  StringRef add(StringRef s) {
+    assert(!isFinalized());
+    return StringIndexMap.GetOrCreateValue(s, 0).getKey();
+  }
+
+  /// \brief Analyze the strings and build the final table. No more strings can
+  /// be added after this point.
+  void finalize();
+
+  /// \brief Retrieve the string table data. Can only be used after the table
+  /// is finalized.
+  StringRef data() {
+    assert(isFinalized());
+    return StringTable;
+  }
+
+  /// \brief Get the offest of a string in the string table. Can only be used
+  /// after the table is finalized.
+  size_t getOffset(StringRef s) {
+    assert(isFinalized());
+    assert(StringIndexMap.count(s) && "String is not in table!");
+    return StringIndexMap[s];
+  }
+
+private:
+  bool isFinalized() {
+    return !StringTable.empty();
+  }
+};
+
+} // end llvm namespace
+
+#endif

lib/MC/ELFObjectWriter.cpp

@@ -28,6 +28,7 @@
 #include "llvm/MC/MCObjectWriter.h"
 #include "llvm/MC/MCSectionELF.h"
 #include "llvm/MC/MCValue.h"
+#include "llvm/Object/StringTableBuilder.h"
 #include "llvm/Support/Compression.h"
 #include "llvm/Support/Debug.h"
 #include "llvm/Support/Endian.h"
@@ -132,11 +133,11 @@ class ELFObjectWriter : public MCObjectWriter {
       MCSymbolData *SymbolData;
       uint64_t StringIndex;
       uint32_t SectionIndex;
+      StringRef Name;
 
       // Support lexicographic sorting.
       bool operator<(const ELFSymbolData &RHS) const {
-        return SymbolData->getSymbol().getName() <
+        return Name < RHS.Name;
-               RHS.SymbolData->getSymbol().getName();
       }
     };
 
@@ -149,13 +150,13 @@ class ELFObjectWriter : public MCObjectWriter {
 
     llvm::DenseMap<const MCSectionData *, std::vector<ELFRelocationEntry>>
     Relocations;
-    DenseMap<const MCSection*, uint64_t> SectionStringTableIndex;
+    StringTableBuilder ShStrTabBuilder;
 
     /// @}
     /// @name Symbol Table Data
     /// @{
 
-    SmallString<256> StringTable;
+    StringTableBuilder StrTabBuilder;
     std::vector<uint64_t> FileSymbolData;
     std::vector<ELFSymbolData> LocalSymbolData;
     std::vector<ELFSymbolData> ExternalSymbolData;
@@ -676,7 +677,6 @@ void ELFObjectWriter::WriteSymbolTable(MCDataFragment *SymtabF,
                                        SectionIndexMapTy &SectionIndexMap) {
   // The string table must be emitted first because we need the index
   // into the string table for all the symbol names.
-  assert(StringTable.size() && "Missing string table");
 
   // FIXME: Make sure the start of the symbol table is aligned.
 
@@ -1031,27 +1031,6 @@ ELFObjectWriter::computeSymbolTable(MCAssembler &Asm, const MCAsmLayout &Layout,
     MCELF::SetBinding(Data, ELF::STB_GLOBAL);
   }
 
-  // Index 0 is always the empty string.
-  StringMap<uint64_t> StringIndexMap;
-  StringTable += '\x00';
-
-  // FIXME: We could optimize suffixes in strtab in the same way we
-  // optimize them in shstrtab.
-
-  for (MCAssembler::const_file_name_iterator it = Asm.file_names_begin(),
-                                            ie = Asm.file_names_end();
-                                            it != ie;
-                                            ++it) {
-    StringRef Name = *it;
-    uint64_t &Entry = StringIndexMap[Name];
-    if (!Entry) {
-      Entry = StringTable.size();
-      StringTable += Name;
-      StringTable += '\x00';
-    }
-    FileSymbolData.push_back(Entry);
-  }
-
   // Add the data for the symbols.
   for (MCSymbolData &SD : Asm.symbols()) {
     const MCSymbol &Symbol = SD.getSymbol();
@@ -1102,7 +1081,6 @@ ELFObjectWriter::computeSymbolTable(MCAssembler &Asm, const MCAsmLayout &Layout,
     // @@ in defined ones.
     StringRef Name = Symbol.getName();
     SmallString<32> Buf;
-
     size_t Pos = Name.find("@@@");
     if (Pos != StringRef::npos) {
       Buf += Name.substr(0, Pos);
@@ -1110,14 +1088,8 @@ ELFObjectWriter::computeSymbolTable(MCAssembler &Asm, const MCAsmLayout &Layout,
       Buf += Name.substr(Pos + Skip);
       Name = Buf;
     }
+    MSD.Name = StrTabBuilder.add(Name);
 
-    uint64_t &Entry = StringIndexMap[Name];
-    if (!Entry) {
-      Entry = StringTable.size();
-      StringTable += Name;
-      StringTable += '\x00';
-    }
-    MSD.StringIndex = Entry;
     if (MSD.SectionIndex == ELF::SHN_UNDEF)
       UndefinedSymbolData.push_back(MSD);
     else if (Local)
@@ -1126,6 +1098,21 @@ ELFObjectWriter::computeSymbolTable(MCAssembler &Asm, const MCAsmLayout &Layout,
       ExternalSymbolData.push_back(MSD);
   }
 
+  for (auto i = Asm.file_names_begin(), e = Asm.file_names_end(); i != e; ++i)
+    StrTabBuilder.add(*i);
+
+  StrTabBuilder.finalize();
+
+  for (auto i = Asm.file_names_begin(), e = Asm.file_names_end(); i != e; ++i)
+    FileSymbolData.push_back(StrTabBuilder.getOffset(*i));
+
+  for (ELFSymbolData& MSD : LocalSymbolData)
+    MSD.StringIndex = StrTabBuilder.getOffset(MSD.Name);
+  for (ELFSymbolData& MSD : ExternalSymbolData)
+    MSD.StringIndex = StrTabBuilder.getOffset(MSD.Name);
+  for (ELFSymbolData& MSD : UndefinedSymbolData)
+    MSD.StringIndex = StrTabBuilder.getOffset(MSD.Name);
+
   // Symbols are required to be in lexicographic order.
   array_pod_sort(LocalSymbolData.begin(), LocalSymbolData.end());
   array_pod_sort(ExternalSymbolData.begin(), ExternalSymbolData.end());
@@ -1436,23 +1423,6 @@ void ELFObjectWriter::WriteRelocationsFragment(const MCAssembler &Asm,
   }
 }
 
-static int compareBySuffix(const MCSectionELF *const *a,
-                           const MCSectionELF *const *b) {
-  const StringRef &NameA = (*a)->getSectionName();
-  const StringRef &NameB = (*b)->getSectionName();
-  const unsigned sizeA = NameA.size();
-  const unsigned sizeB = NameB.size();
-  const unsigned len = std::min(sizeA, sizeB);
-  for (unsigned int i = 0; i < len; ++i) {
-    char ca = NameA[sizeA - i - 1];
-    char cb = NameB[sizeB - i - 1];
-    if (ca != cb)
-      return cb - ca;
-  }
-
-  return sizeB - sizeA;
-}
-
 void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm,
                                              MCAsmLayout &Layout,
                                              SectionIndexMapTy &SectionIndexMap,
@@ -1493,45 +1463,20 @@ void ELFObjectWriter::CreateMetadataSections(MCAssembler &Asm,
   WriteSymbolTable(F, Asm, Layout, SectionIndexMap);
 
   F = new MCDataFragment(&StrtabSD);
-  F->getContents().append(StringTable.begin(), StringTable.end());
+  F->getContents().append(StrTabBuilder.data().begin(),
+                          StrTabBuilder.data().end());
 
   F = new MCDataFragment(&ShstrtabSD);
 
-  std::vector<const MCSectionELF*> Sections;
+  // Section header string table.
-  for (MCAssembler::const_iterator it = Asm.begin(),
+  for (auto it = Asm.begin(), ie = Asm.end(); it != ie; ++it) {
-         ie = Asm.end(); it != ie; ++it) {
     const MCSectionELF &Section =
       static_cast<const MCSectionELF&>(it->getSection());
-    Sections.push_back(&Section);
+    ShStrTabBuilder.add(Section.getSectionName());
-  }
-  array_pod_sort(Sections.begin(), Sections.end(), compareBySuffix);
-
-  // Section header string table.
-  //
-  // The first entry of a string table holds a null character so skip
-  // section 0.
-  uint64_t Index = 1;
-  F->getContents().push_back('\x00');
-
-  for (unsigned int I = 0, E = Sections.size(); I != E; ++I) {
-    const MCSectionELF &Section = *Sections[I];
-
-    StringRef Name = Section.getSectionName();
-    if (I != 0) {
-      StringRef PreviousName = Sections[I - 1]->getSectionName();
-      if (PreviousName.endswith(Name)) {
-        SectionStringTableIndex[&Section] = Index - Name.size() - 1;
-        continue;
-      }
-    }
-    // Remember the index into the string table so we can write it
-    // into the sh_name field of the section header table.
-    SectionStringTableIndex[&Section] = Index;
-
-    Index += Name.size() + 1;
-    F->getContents().append(Name.begin(), Name.end());
-    F->getContents().push_back('\x00');
   }
+  ShStrTabBuilder.finalize();
+  F->getContents().append(ShStrTabBuilder.data().begin(),
+                          ShStrTabBuilder.data().end());
 }
 
 void ELFObjectWriter::CreateIndexedSections(MCAssembler &Asm,
@@ -1599,7 +1544,7 @@ void ELFObjectWriter::WriteSection(MCAssembler &Asm,
 
   switch(Section.getType()) {
   case ELF::SHT_DYNAMIC:
-    sh_link = SectionStringTableIndex[&Section];
+    sh_link = ShStrTabBuilder.getOffset(Section.getSectionName());
     sh_info = 0;
     break;
 
@@ -1680,7 +1625,8 @@ void ELFObjectWriter::WriteSection(MCAssembler &Asm,
     }
   }
 
-  WriteSecHdrEntry(SectionStringTableIndex[&Section], Section.getType(),
+  WriteSecHdrEntry(ShStrTabBuilder.getOffset(Section.getSectionName()),
+                   Section.getType(),
                    Section.getFlags(), 0, Offset, Size, sh_link, sh_info,
                    Alignment, Section.getEntrySize());
 }

lib/Object/CMakeLists.txt

@@ -12,6 +12,7 @@ add_llvm_library(LLVMObject
   MachOUniversal.cpp
   Object.cpp
   ObjectFile.cpp
+  StringTableBuilder.cpp
   SymbolicFile.cpp
   YAML.cpp
   )

lib/Object/StringTableBuilder.cpp

@@ -0,0 +1,51 @@
+//===-- StringTableBuilder.cpp - String table building utility ------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/ADT/SmallVector.h"
+#include "llvm/Object/StringTableBuilder.h"
+
+using namespace llvm;
+
+static bool compareBySuffix(StringRef a, StringRef b) {
+  size_t sizeA = a.size();
+  size_t sizeB = b.size();
+  size_t len = std::min(sizeA, sizeB);
+  for (size_t i = 0; i < len; ++i) {
+    char ca = a[sizeA - i - 1];
+    char cb = b[sizeB - i - 1];
+    if (ca != cb)
+      return ca > cb;
+  }
+  return sizeA > sizeB;
+}
+
+void StringTableBuilder::finalize() {
+  SmallVector<StringRef, 8> Strings;
+  for (auto i = StringIndexMap.begin(), e = StringIndexMap.end(); i != e; ++i)
+    Strings.push_back(i->getKey());
+
+  std::sort(Strings.begin(), Strings.end(), compareBySuffix);
+
+  // FIXME: Starting with a null byte is ELF specific. Generalize this so we
+  // can use the class with other object formats.
+  StringTable += '\x00';
+
+  StringRef Previous;
+  for (StringRef s : Strings) {
+    if (Previous.endswith(s)) {
+      StringIndexMap[s] = StringTable.size() - 1 - s.size();
+      continue;
+    }
+
+    StringIndexMap[s] = StringTable.size();
+    StringTable += s;
+    StringTable += '\x00';
+    Previous = s;
+  }
+}

test/MC/AArch64/tls-relocs.s

@@ -543,7 +543,7 @@
 
 // CHECK-ELF:      Symbols [
 // CHECK-ELF:        Symbol {
-// CHECK-ELF:          Name: var (6)
+// CHECK-ELF:          Name: var
 // CHECK-ELF-NEXT:     Value:
 // CHECK-ELF-NEXT:     Size:
 // CHECK-ELF-NEXT:     Binding: Global

test/MC/ARM64/tls-relocs.s

@@ -313,7 +313,7 @@
 
 // CHECK-ELF:      Symbols [
 // CHECK-ELF:        Symbol {
-// CHECK-ELF:          Name: var (6)
+// CHECK-ELF:          Name: var
 // CHECK-ELF-NEXT:     Value:
 // CHECK-ELF-NEXT:     Size:
 // CHECK-ELF-NEXT:     Binding: Global

test/MC/ELF/comdat.s

@@ -49,7 +49,7 @@
 // Test that g1 and g2 are local, but g3 is an undefined global.
 
 // CHECK:        Symbol {
-// CHECK:          Name: g1 (1)
+// CHECK:          Name: g1
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -58,7 +58,7 @@
 // CHECK-NEXT:     Section: .foo (0x7)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: g2 (4)
+// CHECK-NEXT:     Name: g2
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -68,7 +68,7 @@
 // CHECK-NEXT:   }
 
 // CHECK:        Symbol {
-// CHECK:          Name: g3 (7)
+// CHECK:          Name: g3
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/common.s

@@ -9,7 +9,7 @@
 	.comm	common1,1,1
 
 // CHECK:        Symbol {
-// CHECK:          Name: common1 (1)
+// CHECK:          Name: common1
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 1
 // CHECK-NEXT:     Binding: Local
@@ -25,7 +25,7 @@
 	.comm	common2,1,1
 
 // CHECK:        Symbol {
-// CHECK:          Name: common2 (9)
+// CHECK:          Name: common2
 // CHECK-NEXT:     Value: 0x1
 // CHECK-NEXT:     Size: 1
 // CHECK-NEXT:     Binding: Local
@@ -39,7 +39,7 @@
         .comm	common6,8,16
 
 // CHECK:        Symbol {
-// CHECK:          Name: common6 (17)
+// CHECK:          Name: common6
 // CHECK-NEXT:     Value: 0x10
 // CHECK-NEXT:     Size: 8
 // CHECK-NEXT:     Binding: Local
@@ -54,7 +54,7 @@
 	.comm	common3,4,4
 
 // CHECK:        Symbol {
-// CHECK:          Name: common3 (25)
+// CHECK:          Name: common3
 // CHECK-NEXT:     Value: 0x4
 // CHECK-NEXT:     Size: 4
 // CHECK-NEXT:     Binding: Global
@@ -76,7 +76,7 @@ foo:
 	.comm	common4,40,16
 
 // CHECK:        Symbol {
-// CHECK:          Name: common4 (37)
+// CHECK:          Name: common4
 // CHECK-NEXT:     Value: 0x10
 // CHECK-NEXT:     Size: 40
 // CHECK-NEXT:     Binding: Global
@@ -89,7 +89,7 @@ foo:
         .comm	common5,4,4
 
 // CHECK:        Symbol {
-// CHECK:          Name: common5 (45)
+// CHECK:          Name: common5
 // CHECK-NEXT:     Value: 0x4
 // CHECK-NEXT:     Size: 4
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/file-double.s

@@ -11,7 +11,7 @@ foo.c:
 bar.c:
 
 // CHECK:        Symbol {
-// CHECK:          Name: foo.c (1)
+// CHECK:          Name: foo.c
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -19,7 +19,7 @@ bar.c:
 // CHECK-NEXT:     Other: 0
 // CHECK-NEXT:     Section: Absolute (0xFFF1)
 // CHECK-NEXT:   }
-// CHECK:          Name: bar.c (7)
+// CHECK:          Name: bar.c
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -28,7 +28,7 @@ bar.c:
 // CHECK-NEXT:     Section: Absolute (0xFFF1)
 // CHECK-NEXT:   }
 // CHECK:        Symbol {
-// CHECK:        Name: bar.c (7)
+// CHECK:        Name: bar.c
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -37,7 +37,7 @@ bar.c:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK:        Symbol {
-// CHECK:        Name: foo.c (1)
+// CHECK:        Name: foo.c
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/lcomm.s

@@ -4,7 +4,7 @@
 .lcomm B, 32 << 20
 
 // CHECK:        Symbol {
-// CHECK:          Name: A (1)
+// CHECK:          Name: A
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 5
 // CHECK-NEXT:     Binding: Local
@@ -13,7 +13,7 @@
 // CHECK-NEXT:     Section: .bss (0x3)
 // CHECK-NEXT:   }
 // CHECK:        Symbol {
-// CHECK:          Name: B (3)
+// CHECK:          Name: B
 // CHECK-NEXT:     Value: 0x5
 // CHECK-NEXT:     Size: 33554432
 // CHECK-NEXT:     Binding: Local

test/MC/ELF/many-sections-2.s

@@ -12,7 +12,7 @@
 
 
 // Test that both a and b show up in the correct section.
-// SYMBOLS:         Name: a (1)
+// SYMBOLS:         Name: a
 // SYMBOLS-NEXT:    Value: 0x0
 // SYMBOLS-NEXT:    Size: 0
 // SYMBOLS-NEXT:    Binding: Local (0x0)
@@ -21,7 +21,7 @@
 // SYMBOLS-NEXT:    Section: last (0xFF00)
 // SYMBOLS-NEXT:  }
 // SYMBOLS-NEXT:  Symbol {
-// SYMBOLS-NEXT:    Name: b (3)
+// SYMBOLS-NEXT:    Name: b
 // SYMBOLS-NEXT:    Value: 0x1
 // SYMBOLS-NEXT:    Size: 0
 // SYMBOLS-NEXT:    Binding: Local (0x0)
@@ -32,7 +32,7 @@
 
 
 // Test that this file has one section too many.
-// SYMBOLS:         Name: last (0)
+// SYMBOLS:         Name: last
 // SYMBOLS-NEXT:    Value: 0x0
 // SYMBOLS-NEXT:    Size: 0
 // SYMBOLS-NEXT:    Binding: Local (0x0)

test/MC/ELF/pic-diff.s

@@ -7,7 +7,7 @@
 // CHECK-NEXT: ]
 
 // CHECK:        Symbol {
-// CHECK:          Name: baz (5)
+// CHECK:          Name: baz
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/pr9292.s

@@ -8,7 +8,7 @@ mov %eax,bar
 
 
 // CHECK:        Symbol {
-// CHECK:          Name: bar (5)
+// CHECK:          Name: bar
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -17,7 +17,7 @@ mov %eax,bar
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo (1)
+// CHECK-NEXT:     Name: foo
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/set.s

@@ -5,7 +5,7 @@
 .set kernbase,0xffffffff80000000
 
 // CHECK:        Symbol {
-// CHECK:          Name: kernbase (1)
+// CHECK:          Name: kernbase
 // CHECK-NEXT:     Value: 0xFFFFFFFF80000000
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -26,7 +26,7 @@
 
 // Test that there is an undefined reference to bar
 // CHECK:        Symbol {
-// CHECK:          Name: bar (10)
+// CHECK:          Name: bar
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/strtab-suffix-opt.s

@@ -0,0 +1,21 @@
+// RUN: llvm-mc -filetype=obj -triple i686-pc-linux-gnu %s -o - | llvm-readobj -symbols | FileCheck %s
+
+	.text
+	.globl	foobar
+	.align	16, 0x90
+	.type	foobar,@function
+foobar:
+	pushl	%ebp
+	movl	%esp, %ebp
+	subl	$8, %esp
+	calll	foo
+	calll	bar
+	addl	$8, %esp
+	popl	%ebp
+	retl
+.Ltmp3:
+	.size	foobar, .Ltmp3-foobar
+
+// CHECK:     Name: foobar (1)
+// CHECK:     Name: bar (4)
+// CHECK:     Name: foo (8)

test/MC/ELF/tls-i386.s

@@ -18,7 +18,7 @@
         .long   fooE@INDNTPOFF
 
 // CHECK:        Symbol {
-// CHECK:          Name: foo1 (1)
+// CHECK:          Name: foo1
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -27,7 +27,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo2 (6)
+// CHECK-NEXT:     Name: foo2
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -36,7 +36,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo3 (11)
+// CHECK-NEXT:     Name: foo3
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -45,7 +45,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo4 (16)
+// CHECK-NEXT:     Name: foo4
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -54,7 +54,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo5 (21)
+// CHECK-NEXT:     Name: foo5
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -63,7 +63,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo6 (26)
+// CHECK-NEXT:     Name: foo6
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -72,7 +72,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo7 (31)
+// CHECK-NEXT:     Name: foo7
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -81,7 +81,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo8 (36)
+// CHECK-NEXT:     Name: foo8
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -90,7 +90,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo9 (41)
+// CHECK-NEXT:     Name: foo9
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -99,7 +99,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: fooA (46)
+// CHECK-NEXT:     Name: fooA
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -108,7 +108,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: fooB (51)
+// CHECK-NEXT:     Name: fooB
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -117,7 +117,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: fooC (56)
+// CHECK-NEXT:     Name: fooC
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -126,7 +126,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: fooD (61)
+// CHECK-NEXT:     Name: fooD
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -135,7 +135,7 @@
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: fooE (66)
+// CHECK-NEXT:     Name: fooE
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/tls.s

@@ -13,7 +13,7 @@ foobar:
 	.long	43
 
 // CHECK:        Symbol {
-// CHECK:          Name: foobar (31)
+// CHECK:          Name: foobar
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -23,7 +23,7 @@ foobar:
 // CHECK-NEXT:   }
 
 // CHECK:        Symbol {
-// CHECK:          Name: foo1 (1)
+// CHECK:          Name: foo1
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -32,7 +32,7 @@ foobar:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo2 (6)
+// CHECK-NEXT:     Name: foo2
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -41,7 +41,7 @@ foobar:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo3 (11)
+// CHECK-NEXT:     Name: foo3
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -50,7 +50,7 @@ foobar:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo4 (16)
+// CHECK-NEXT:     Name: foo4
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -59,7 +59,7 @@ foobar:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo5 (21)
+// CHECK-NEXT:     Name: foo5
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -68,7 +68,7 @@ foobar:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: foo6 (26)
+// CHECK-NEXT:     Name: foo6
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

test/MC/ELF/type.s

@@ -176,7 +176,7 @@ alias12:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym1 (54)
+// CHECK-NEXT:    Name: sym1
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -185,7 +185,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym10 (162)
+// CHECK-NEXT:    Name: sym10
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -194,7 +194,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym11 (176)
+// CHECK-NEXT:    Name: sym11
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -203,7 +203,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym12 (190)
+// CHECK-NEXT:    Name: sym12
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -212,7 +212,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym2 (66)
+// CHECK-NEXT:    Name: sym2
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -221,7 +221,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym3 (78)
+// CHECK-NEXT:    Name: sym3
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -230,7 +230,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym4 (90)
+// CHECK-NEXT:    Name: sym4
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -239,7 +239,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym5 (102)
+// CHECK-NEXT:    Name: sym5
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -248,7 +248,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym6 (114)
+// CHECK-NEXT:    Name: sym6
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -257,7 +257,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym7 (126)
+// CHECK-NEXT:    Name: sym7
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -266,7 +266,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym8 (138)
+// CHECK-NEXT:    Name: sym8
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)
@@ -275,7 +275,7 @@ alias12:
 // CHECK-NEXT:    Section: .text (0x1)
 // CHECK-NEXT:  }
 // CHECK-NEXT:  Symbol {
-// CHECK-NEXT:    Name: sym9 (150)
+// CHECK-NEXT:    Name: sym9
 // CHECK-NEXT:    Value: 0x0
 // CHECK-NEXT:    Size: 0
 // CHECK-NEXT:    Binding: Global (0x1)

test/MC/ELF/weakref.s

@@ -80,7 +80,7 @@ bar15:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar6 (21)
+// CHECK-NEXT:     Name: bar6
 // CHECK-NEXT:     Value: 0x18
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -89,7 +89,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar7 (26)
+// CHECK-NEXT:     Name: bar7
 // CHECK-NEXT:     Value: 0x18
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -98,7 +98,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar8 (31)
+// CHECK-NEXT:     Name: bar8
 // CHECK-NEXT:     Value: 0x1C
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -107,7 +107,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar9 (36)
+// CHECK-NEXT:     Name: bar9
 // CHECK-NEXT:     Value: 0x20
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -116,7 +116,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: .text (0)
+// CHECK-NEXT:     Name: .text
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -125,7 +125,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: .data (0)
+// CHECK-NEXT:     Name: .data
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -134,7 +134,7 @@ bar15:
 // CHECK-NEXT:     Section: .data (0x3)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: .bss (0)
+// CHECK-NEXT:     Name: .bss
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Local
@@ -143,7 +143,7 @@ bar15:
 // CHECK-NEXT:     Section: .bss (0x4)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar10 (41)
+// CHECK-NEXT:     Name: bar10
 // CHECK-NEXT:     Value: 0x28
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -152,7 +152,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar11 (47)
+// CHECK-NEXT:     Name: bar11
 // CHECK-NEXT:     Value: 0x30
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -161,7 +161,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar12 (53)
+// CHECK-NEXT:     Name: bar12
 // CHECK-NEXT:     Value: 0x30
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -170,7 +170,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar13 (59)
+// CHECK-NEXT:     Name: bar13
 // CHECK-NEXT:     Value: 0x34
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -179,7 +179,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar14 (65)
+// CHECK-NEXT:     Name: bar14
 // CHECK-NEXT:     Value: 0x38
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -188,7 +188,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar15 (71)
+// CHECK-NEXT:     Name: bar15
 // CHECK-NEXT:     Value: 0x40
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -197,7 +197,7 @@ bar15:
 // CHECK-NEXT:     Section: .text (0x1)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar2 (1)
+// CHECK-NEXT:     Name: bar2
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -206,7 +206,7 @@ bar15:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar3 (6)
+// CHECK-NEXT:     Name: bar3
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Weak
@@ -215,7 +215,7 @@ bar15:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar4 (11)
+// CHECK-NEXT:     Name: bar4
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global
@@ -224,7 +224,7 @@ bar15:
 // CHECK-NEXT:     Section: Undefined (0x0)
 // CHECK-NEXT:   }
 // CHECK-NEXT:   Symbol {
-// CHECK-NEXT:     Name: bar5 (16)
+// CHECK-NEXT:     Name: bar5
 // CHECK-NEXT:     Value: 0x0
 // CHECK-NEXT:     Size: 0
 // CHECK-NEXT:     Binding: Global

tools/yaml2obj/yaml2elf.cpp

@@ -30,10 +30,8 @@ using namespace llvm;
 // This class has a deliberately small interface, since a lot of
 // implementation variation is possible.
 //
-// TODO: Use an ordered container with a suffix-based comparison in order
+// TODO: Use the StringTable builder from lib/Object instead, since it
-// to deduplicate suffixes. std::map<> with a custom comparator is likely
+// will deduplicate suffixes.
-// to be the simplest implementation, but a suffix trie could be more
-// suitable for the job.
 namespace {
 class StringTableBuilder {
   /// \brief Indices of strings currently present in `Buf`.

unittests/Object/CMakeLists.txt

@@ -4,5 +4,6 @@ set(LLVM_LINK_COMPONENTS
   )
 
 add_llvm_unittest(ObjectTests
+  StringTableBuilderTest.cpp
   YAMLTest.cpp
   )

unittests/Object/StringTableBuilderTest.cpp

@@ -0,0 +1,40 @@
+//===----------- StringTableBuilderTest.cpp -------------------------------===//
+//
+//                     The LLVM Compiler Infrastructure
+//
+// This file is distributed under the University of Illinois Open Source
+// License. See LICENSE.TXT for details.
+//
+//===----------------------------------------------------------------------===//
+
+#include "gtest/gtest.h"
+#include "llvm/Object/StringTableBuilder.h"
+#include <string>
+
+using namespace llvm;
+
+namespace {
+
+TEST(StringTableBuilderTest, Basic) {
+  StringTableBuilder B;
+
+  B.add("foo");
+  B.add("bar");
+  B.add("foobar");
+
+  B.finalize();
+
+  std::string Expected;
+  Expected += '\x00';
+  Expected += "foobar";
+  Expected += '\x00';
+  Expected += "foo";
+  Expected += '\x00';
+
+  EXPECT_EQ(Expected, B.data());
+  EXPECT_EQ(1U, B.getOffset("foobar"));
+  EXPECT_EQ(4U, B.getOffset("bar"));
+  EXPECT_EQ(8U, B.getOffset("foo"));
+}
+
+}