ASDenysPetrov added a comment.
I'll update tommorow.
================
Comment at: clang/unittests/StaticAnalyzer/RangeSetTest.cpp:120
+
+ template <typename T> const llvm::APSInt &from(T X) {
+ static llvm::APSInt Int = APSIntTy<T>.getZeroValue();
----------------
vsavchenko wrote:
> Default to `BaseType`?
It's implicitly deduced. I think it is not necessary.
================
Comment at: clang/unittests/StaticAnalyzer/RangeSetTest.cpp:131
+ template <typename T>
+ RangeSet from(RawRangeSetT<T> Init, APSIntType Ty = APSIntTy<BaseType>) {
RangeSet RangeSet = F.getEmptySet();
----------------
vsavchenko wrote:
> Unused parameter?
Oh... :-)
================
Comment at: clang/unittests/StaticAnalyzer/RangeSetTest.cpp:721-799
+TYPED_TEST(RangeSetCastToNoopTest, RangeSetCastToNoopTest) {
+ // Just to reduce the verbosity.
+ using F = typename TypeParam::FromType; // From
+ using T = typename TypeParam::ToType; // To
+
+ using TV = TestValues<T>;
+ constexpr auto MIN = TV::MIN;
----------------
vsavchenko wrote:
> If loop and promotion share the same test case, why should we split them into
> two groups?
Yes, the tests are identical but they use different `testing::Types` in the
suites. Also I don't want to mix them to make problem localizing easier if
occurs. But still I haven't strong preferences on that.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D103094/new/
https://reviews.llvm.org/D103094
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
