Displaying 3 results from an estimated 3 matches for "invalid_".
Did you mean:
invalid
2013 Jul 23
2
[LLVMdev] Analysis of polly-detect overhead in oggenc
..._ostream OS(Message);
> +
> + if (PollyViewMode || ::llvm::DebugFlag) {
This is a little unsatisfying. We now have two conditions that need to
be in sync. I think you can avoid this, if you rename the function to
std::string ScopDetection::formatInvalidAlias(AliasSet &AS)
and keep the INVALID_ macro at the place where the error happens.
> diff --git a/lib/RegisterPasses.cpp b/lib/RegisterPasses.cpp
> index 7fc0960..2e25e4d 100644
> --- a/lib/RegisterPasses.cpp
> +++ b/lib/RegisterPasses.cpp
> @@ -125,28 +125,34 @@ static cl::opt<bool> DeadCodeElim("polly-run-dc...
2013 Jul 23
0
[LLVMdev] Analysis of polly-detect overhead in oggenc
...lls:
>> if (!Context.Verifying) \
>> ++Bad##NAME##ForScop;
>> So, I do not think it is necessary to translate the INVALID macro into function calls.
>> Do you still think we should translate INVALID macro into a serial of functions like "invalid_CFG, invalid_IndVar, invalid_IndEdge, ... ? In that case, I could provide a small patch file for this purpose -:)
>
>I think it would still be nice to get rid of this macro. We could
>probably have a default function that takes an enum to report different
>errors in the reportInvalid(e...
2013 Jul 22
2
[LLVMdev] Analysis of polly-detect overhead in oggenc
...n in function calls:
> if (!Context.Verifying) \
> ++Bad##NAME##ForScop;
> So, I do not think it is necessary to translate the INVALID macro into function calls.
> Do you still think we should translate INVALID macro into a serial of functions like "invalid_CFG, invalid_IndVar, invalid_IndEdge, ... ? In that case, I could provide a small patch file for this purpose -:)
I think it would still be nice to get rid of this macro. We could
probably have a default function that takes an enum to report different
errors in the reportInvalid(enum errorKind) s...