[llvm-commits] [llvm] r51468 - in /llvm/trunk: include/llvm/Bitcode/LLVMBitCodes.h include/llvm/Type.h lib/AsmParser/LLLexer.cpp lib/AsmParser/llvmAsmParser.cpp.cvs lib/AsmParser/llvmAsmParser.h.cvs lib/AsmParser/llvmAsmParser.y lib/AsmParser/llvmAsmParser.y.cvs lib/Bitcode/Reader/BitcodeReader.cpp lib/Bitcode/Writer/BitcodeWriter.cpp lib/Bitcode/Writer/ValueEnumerator.cpp test/Assembler/insertextractvalue.ll test/Verifier/2006-07-11-StoreStruct.ll utils/llvm.grm

[Chris Lattner]

On May 23, 2008, at 3:24 PM, Dan Gohman wrote:

>>> +  // Partition the Type ID's so that the single-value types occur
>>> before the
>>> // aggregate types.  This allows the aggregate types to be dropped
>>> from the
>>> // type table after parsing the global variable initializers.
>>> -  std::partition(Types.begin(), Types.end(), isFirstClassType);
>>> +  std::partition(Types.begin(), Types.end(), isSingleValueType);
>>
>> Is this required for correctness?  This could cause a significant BC
>> file size increase for real world apps.  I'd try kimwitu and 176.gcc
>> as examples.
>
> It is required for correctness, given some other stuff that the
> BC writer does. But it shouldn't change the BC file size;
> isSingleValueType does exactly what isFirstClassType used to do;
> that's why it exists even.

Oh, right!  Thanks Dan,

-Chris