================
@@ -2305,7 +2335,45 @@ TEST(BlockEndHints, PointerToMemberFunction) {
$ptrmem[[}]]
} // suppress
)cpp",
- ExpectedHint{" // if", "ptrmem"});
+ ExpectedHint{" // if ()", "ptrmem"});
+}
+
+TEST(BlockEndHints, MinLineLimit) {
+ assertBlockEndHintsWithOpts(
+ R"cpp(
+ namespace ns {
+ int Var;
+ int func1();
+ int func2(int, int);
+ struct S {
+ int Field;
+ int method1() const;
+ int method2(int, int) const;
+ $struct[[}]];
----------------
HighCommander4 wrote:
It's not necessary.
The core assertion the tests make is [this
one](https://searchfox.org/llvm/rev/bb21a6819b3fb9d689de776f7ee768030dfbacea/clang-tools-extra/clangd/unittests/InlayHintTests.cpp#101-102):
```
EXPECT_THAT(hintsOfKind(AST, Kind),
ElementsAre(HintMatcher(Expected, Source)...));
```
`hintsOfKind(AST, Kind)` returns **all** inlay hints of the specified kind in
the code; its output does not depend on the presence or absence of annotations
in the code.
`ElementsAre(...)` then assert that these returned hints match the provided
`ExpectedHint` objects, including how many there are. If in the test case you
have two `ExpectedHint` objects, and the `hintsOfKind()` starts returning three
hints, the test will start failing.
The only purpose of the annotations is so that you can name a range passed in
to an `ExpectedHint`; if you never need to name the range, you don't need the
annotation.
https://github.com/llvm/llvm-project/pull/136106
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
