aaron.ballman added inline comments.
================
Comment at: clang-tidy/modernize/UseBoolLiteralsCheck.cpp:51-52
@@ -34,4 +50,4 @@
void UseBoolLiteralsCheck::check(const MatchFinder::MatchResult &Result) {
- const auto *Literal = Result.Nodes.getNodeAs<IntegerLiteral>("literal");
- const auto *Cast = Result.Nodes.getNodeAs<Expr>("cast");
- bool LiteralBooleanValue = Literal->getValue().getBoolValue();
+ for (const auto &BindingName :
+ {"literal", "trueBranchLiteral", "falseBranchLiteral"}) {
+ const auto *Literal = Result.Nodes.getNodeAs<IntegerLiteral>(BindingName);
----------------
omtcyfz wrote:
> aaron.ballman wrote:
> > omtcyfz wrote:
> > > aaron.ballman wrote:
> > > > Any reason not to name the bind "literal" in all three cases? That
> > > > eliminates the need for the loop entirely, since `check()` will trigger
> > > > for each instance of a match.
> > > It doesn't make sense to try binding both `TrueExpression` and
> > > `FalseExpression` literals to a single value.
> > Why? In all three cases, you don't care what matched, just that *some* case
> > is matched. None of the logic in `check()` relies on which part of the
> > expression is matched.
> Well, in case of second matcher I may have **two** literals matched upon
> triggering. I don't understand how I could possibly get **two** literals
> bound to **one** value after **one** matcher got triggered.
>
> Am I missing something?
One matcher isn't what's getting triggered then, is it? I could be wrong on
this point, but I thought that in that case, `check()` would be called twice,
once for each literal. Is that not the case?
https://reviews.llvm.org/D23243
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
