[llvm-bugs] [Bug 48196] New: improve diagnostic for "new func()"

[via llvm-bugs]

https://bugs.llvm.org/show_bug.cgi?id=48196

            Bug ID: 48196
           Summary: improve diagnostic for "new func()"
           Product: clang
           Version: unspecified
          Hardware: PC
                OS: Linux
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: -New Bugs
          Assignee: unassignedclangbugs at nondot.org
          Reporter: nick at wasmer.io
                CC: htmldeveloper at gmail.com, llvm-bugs at lists.llvm.org,
                    neeilans at live.com, richard-llvm at metafoo.co.uk

Testcase:

class Foo {
  Foo() = default;
public:
  static Foo make();
};

Foo *test() {
  return new Foo::make();
}

Clang emits:

<source>:8:19: error: no type named 'make' in 'Foo'
  return new Foo::make();
             ~~~~~^
1 error generated.
Compiler returned: 1

Which was quite confusing at first. There is a 'make' in 'Foo'. Once I
confirmed that, I read the message again. Why does clang demand a type, it's
not like I said typename? What is going on. The error doesn't point at the real
bug, the extra 'new'.

In general when emitting a diagnostics with "there is no X named Y in Z" please
point out that there is a Y in Z that isn't an X. Also when possible please
point to the reason why we're expecting a type (in this case, the 'new'). In
this specific case, it would be magical if clang tried to typo-correct removing
'new' to see if the function call expression would parse and produce an
expression of the correct type.

-- 
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20201116/391fff8d/attachment-0001.html>