[PATCH] Change llvm-objdump to print (bad) for invalid instructions

[Jim Grosbach]

On Oct 16, 2013, at 11:49 AM, Stephen Checkoway <s at pahtak.org> wrote:

> 
> On Oct 16, 2013, at 2:25 PM, Stephen Checkoway <s at pahtak.org> wrote:
> 
>> 
>> On Oct 16, 2013, at 2:05 PM, Jim Grosbach <grosbach at apple.com> wrote:
>> 
>>> I don’t think you want to always set the size to ‘1’ here. Won’t that result in mis-aligned instruction decoding for fixed-width instruction sets?
>> 
>> 
>> That's a good point and there's a similar issue in MCObjectDisassemble.cpp. I wasn't sure if DisAsm::getInstruction() would set Size if it returns false but, looking closer at the code, I think it does.
>> 
>> I'll fix this and also the MCObjectDisassembler patch I sent a little while ago that Rafael said he was going to look at.
> 
> Attached.
> 

LGTM. Perhaps also adjust the documentation header comment for getInstruction() to reflect that size must always return non-zero?

> If you get a moment to look at the MCInstrAnalysis::evaluateBranch() patch and suggest the correct way to fix it, that'd be great.
> 
> -- 
> Stephen Checkoway
> 
> 
> <llvm-objdump-invalid-opcode.diff>
> 
>