Fix test problem.

Thanks Alexey. This to fix test format.

Thanks Alexey! Address the comment.

In D127803#3693660, @ABataev wrote:

In D127803#3693655, @jyu2 wrote:

In D127803#3693301, @mgabka wrote:

I noticed that this patch is causing now an assertion failure for cases like :

class A{

void a() {
  #pragma omp parallel
  a(b);
}

};

The failed assertion is: "const clang::ValueDecl* getCanonicalDecl(const clang::ValueDecl*): Assertion `FD' failed."

while before it clang was correctly reporting error:
error: use of undeclared identifier 'b'

is it the same assertion you were trying to fix here?

No, the assert I am fixing is when default(firstprivate) is used inside member function.

Could you please investigate and fix it?

In D127803#3693301, @mgabka wrote:

I noticed that this patch is causing now an assertion failure for cases like :

class A{

void a() {
  #pragma omp parallel
  a(b);
}

};

The failed assertion is: "const clang::ValueDecl* getCanonicalDecl(const clang::ValueDecl*): Assertion `FD' failed."

while before it clang was correctly reporting error:
error: use of undeclared identifier 'b'

is it the same assertion you were trying to fix here?

In D129872#3671847, @ABataev wrote:

Why does it get truncated if the type must be integer? Probably, something incorrect in sema.

Ping!

Thank Alexey for the review. This is remove check for omp region just use hasDSA.

Thanks Alexey for the review. This patch to address his comment.

Thanks Alexey for the review.
This is to address Alexey's comments,

Thanks Alexey for the review.
Add change in TreeTransform.h to rebuilt member expression.

Remove change in TreeTransform.h. Will deal that in other patch. Thanks.

Address Alexey's comments! Thanks.

Thanks Alexey's review. Address Alexey's review.

Thanks Alexey for the review. This is address Alexey's comments.

Thanks Alexey's comment. Address his comment. Instead save info into DSA, add new field in ImplicitDefaultFirstprivateFDs in SharingMapTy.

Thanks Alexey for the review. This is fix the nit.

Add test in cfg-openmp.cpp

Thanks Alexey, for the review! This patch addresses his comments.

In D122338#3405948, @ABataev wrote:

LG

Address Alexey comment.

Address Alexey's comment.

In D122338#3403414, @ABataev wrote:

If ParamPositions grows, it means that the item is not found. Is it correct behavior if the compiler inserts new item in the ParamPositions map?

This is okay with me.

Fix format problem.

Looks good to me. Thanks.

Thanks Alexey for the code review.

Address Alexey's comments.

Good point. I did not find that. At this point, I don't know if we need this change. I checked gnu, gnu allow this.... :-(

fix format problem.

In D112292#3080945, @lebedev.ri wrote:

In which standard version was that rule introduced, is this missing some checks for the OpenMP version?

Hi Alexey, Thanks for review.

Adding check for map type in test. Thanks Alexey.

Thanks Alexey! Fix missing Modifiers.

Address Alexey's comments.

Thanks Alexey's review. This patch is address Alexe's comments by adding new field LambdasMap.

Thank you so much for Alex's review!!!

Address Alex's comment.