[llvm-dev] [RFC] New diagnostic handler for llc

[Reid Kleckner via llvm-dev]

On Thu, May 12, 2016 at 10:16 AM, Diana Picus via llvm-dev
<llvm-dev at lists.llvm.org> wrote:
> On 12 May 2016 at 16:19, Krzysztof Parzyszek via llvm-dev
> <llvm-dev at lists.llvm.org> wrote:
>> The problem in PR24071 seemed to be that clang proceeded with compilation
>> even though the inline asm was not valid.  I'm not sure that there is value
>> in trying to make the backend continue compiling code that most likely has
>> no meaning.
>
> I'm not 100% convinced that is the case here. The inline asm errors
> are reported via LLVMContext::emitError, which has this documentation:
>
>   /// emitError - Emit an error message to the currently installed error handler
>   /// with optional location information.  This function returns, so code should
>   /// be prepared to drop the erroneous construct on the floor and "not crash".
>   /// The generated code need not be correct.
>
> Since the backend is the one calling emitError, the backend is the one
> that shouldn't crash, regardless of what diagnostic handler is
> installed. I think llc should help us test that this is the case. At
> least that's my interpretation of things.

I agree with your interpretation. We should give llc a diagnostic
handler so that 'emitError' doesn't exit immediately when we're
testing with llc.