Object: Use offset+size as the irsymtab string representation.

[pcc]

Summary

This is consistent with the bitcode string table.

[tejohnson]

It isn't clear to me why this is needed now but wasn't before. Was it the case that before the names were assumed to be null-terminated, and they will no longer be?

llvm/lib/Object/IRSymtab.cpp

33

How does this change related to the rest of the patch?

[pcc]

In D31922#728259, @tejohnson wrote:

It isn't clear to me why this is needed now but wasn't before. Was it the case that before the names were assumed to be null-terminated, and they will no longer be?

Correct. In D32061 we will start sharing the string table between the module and the irsymtab, and the string table I am adding in D31838 uses non-null-terminated strings.

llvm/lib/Object/IRSymtab.cpp

33

Strings in ELF string tables are null terminated, RAW string tables aren't.

[tejohnson]

LGTM

llvm/include/llvm/Object/IRSymtab.h

 // The data structures in this namespace define the low-level serialization
 // format. Clients that just want to read a symbol table should use the
 // irsymtab::Reader class.
 
 typedef support::ulittle32_t Word;
 
 /// A reference to a string in the string table.
 struct Str {
-  Word Offset;
+  Word Offset, Size;
   StringRef get(StringRef Strtab) const {
-    return Strtab.data() + Offset;
+    return {Strtab.data() + Offset, Size};
   }
 };
 
 /// A reference to a range of objects in the symbol table.
 template <typename T> struct Range {
   Word Offset, Size;
   ArrayRef<T> get(StringRef Symtab) const {
     return {reinterpret_cast<const T *>(Symtab.data() + Offset), Size};

llvm/lib/Object/IRSymtab.cpp

 
 /// Stores the temporary state that is required to build an IR symbol table.
 struct Builder {
   SmallVector<char, 0> &Symtab;
   SmallVector<char, 0> &Strtab;
   Builder(SmallVector<char, 0> &Symtab, SmallVector<char, 0> &Strtab)
       : Symtab(Symtab), Strtab(Strtab) {}
 
-  StringTableBuilder StrtabBuilder{StringTableBuilder::ELF};
+  StringTableBuilder StrtabBuilder{StringTableBuilder::RAW};
 
   BumpPtrAllocator Alloc;
   StringSaver Saver{Alloc};
 
   DenseMap<const Comdat *, unsigned> ComdatMap;
   ModuleSymbolTable Msymtab;
   SmallPtrSet<GlobalValue *, 8> Used;
   Mangler Mang;
   Triple TT;
 
   std::vector<storage::Comdat> Comdats;
   std::vector<storage::Module> Mods;
   std::vector<storage::Symbol> Syms;
   std::vector<storage::Uncommon> Uncommons;
 
   std::string COFFLinkerOpts;
   raw_string_ostream COFFLinkerOptsOS{COFFLinkerOpts};
 
   void setStr(storage::Str &S, StringRef Value) {
     S.Offset = StrtabBuilder.add(Value);
+    S.Size = Value.size();
   }
   template <typename T>
   void writeRange(storage::Range<T> &R, const std::vector<T> &Objs) {
     R.Offset = Symtab.size();
     R.Size = Objs.size();
     Symtab.insert(Symtab.end(), reinterpret_cast<const char *>(Objs.data()),
                   reinterpret_cast<const char *>(Objs.data() + Objs.size()));
   }