[PATCH] Fix warning about parentheses in boolean expression.

[gonsolo]

Summary

Fix a warning from gcc.

[• dberlin]

I'm not sure why it's complaining since clang doesn't, but you want the parentheses around everything before the &&

[gonsolo]

Better?

[fhahn]

Sorry for the noise, I'm not sure what I did for Phabricator to think I updated the revision. But it looks like the review still shows the most recent revision.

[davide]

lib/Transforms/Scalar/NewGVN.cpp

1080–1088

This needs to be clang-formatt'ed

[gonsolo]

Clang-formatted. I think it was more readable before that.

[davide]

In D28731#646483, @gonsolo wrote:

Clang-formatted. I think it was more readable before that.

In D28731#646483, @gonsolo wrote:

Clang-formatted. I think it was more readable before that.

You formatted unrelated chunks. Revert that and LGTM. Do you need somebody to commit this on your behalf?

[gonsolo]

I thought clang-formatting a few unrelated changes is ok but anyways, here is the new patch.

Please commit, I can't.

[davide]

In D28731#646533, @gonsolo wrote:

I thought clang-formatting a few unrelated changes is ok but anyways, here is the new patch.

It's generally good practice not mixing functional changes and stylistic changes.

[davide]

In D28731#646533, @gonsolo wrote:

I thought clang-formatting a few unrelated changes is ok but anyways, here is the new patch.

Please commit, I can't.

Also, please update a patch with context next time.

lib/Transforms/Scalar/NewGVN.cpp

   assert(
       !isa<Instruction>(NewClass->RepLeader) || !NewClass->RepLeader ||
       I == NewClass->RepLeader ||
-      !DT->properlyDominates(
+      (!DT->properlyDominates(
           I->getParent(),
           cast<Instruction>(NewClass->RepLeader)->getParent()) &&
-          "New class for instruction should not be dominated by instruction");
+          "New class for instruction should not be dominated by instruction"));
 
   if (NewClass->RepLeader != I) {
     auto DFSNum = InstrDFS.lookup(I);