r241019 - PR23942: a pure-specifier's integer literal must be spelled '0'

[Filipe Cabecinhas]

Hi Richard,

It looks like this rev caused all our bots (Mac, Windows, and Linux. Clean
Release+Asserts builds) to fail:

FAIL: Clang :: SemaCXX/override-in-system-header.cpp (6652 of 23274)
******************** TEST 'Clang :: SemaCXX/override-in-system-header.cpp'
FAILED ********************
Script:
--
/home/buildbot/Buildbot/Slave/builds/1.LLVM.Linux.Phase.2/llvm.obj/./bin/clang
-cc1 -internal-isystem
/home/buildbot/Buildbot/Slave/builds/1.LLVM.Linux.Phase.2/llvm.obj/bin/../lib/clang/3.7.0/include
-nostdsysteminc -std=c++11 -isystem
/home/buildbot/Buildbot/Slave/builds/1.LLVM.Linux.Phase.2/clang.src/test/SemaCXX/Inputs
/home/buildbot/Buildbot/Slave/builds/1.LLVM.Linux.Phase.2/clang.src/test/SemaCXX/override-in-system-header.cpp
-verify
--
Exit Code: 1

Command Output (stderr):
--
error: 'error' diagnostics seen but not expected:
  Line 10: initializer on function does not look like a pure-specifier
  Line 17: initializer on function does not look like a pure-specifier
2 errors generated.

--

********************
Testing: 0 .. 10.. 20.. 30.. 40.. 50.. 60.. 70.. 80.. 90..
Testing Time: 30.62s
********************
Failing Tests (1):
    Clang :: SemaCXX/override-in-system-header.cpp


Could you take a look?

Thanks,

  Filipe

On Mon, Jun 29, 2015 at 4:19 PM, Richard Smith <richard-llvm at metafoo.co.uk>
wrote:

> Author: rsmith
> Date: Mon Jun 29 18:19:23 2015
> New Revision: 241019
>
> URL: http://llvm.org/viewvc/llvm-project?rev=241019&view=rev
> Log:
> PR23942: a pure-specifier's integer literal must be spelled '0'
>
> Modified:
>     cfe/trunk/lib/Sema/SemaDecl.cpp
>     cfe/trunk/test/Parser/cxx-class.cpp
>
> Modified: cfe/trunk/lib/Sema/SemaDecl.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Sema/SemaDecl.cpp?rev=241019&r1=241018&r2=241019&view=diff
>
> ==============================================================================
> --- cfe/trunk/lib/Sema/SemaDecl.cpp (original)
> +++ cfe/trunk/lib/Sema/SemaDecl.cpp Mon Jun 29 18:19:23 2015
> @@ -8783,6 +8783,19 @@ namespace {
>    }
>  }
>
> +/// Determine whether the given expression was formed from the token '0'.
> This
> +/// test is necessary to determine whether an initializer is really a
> +/// pure-specifier.
> +static bool isZeroToken(Sema &S, Expr *E) {
> +  auto *IL = dyn_cast<IntegerLiteral>(E);
> +  if (!IL || !!IL->getValue() ||
> +      !IL->getType()->isSpecificBuiltinType(BuiltinType::Int))
> +    return false;
> +
> +  SmallString<8> Buffer;
> +  return S.PP.getSpelling(E->getLocStart(), Buffer) == "0";
> +}
> +
>  /// AddInitializerToDecl - Adds the initializer Init to the
>  /// declaration dcl. If DirectInit is true, this is C++ direct
>  /// initialization rather than copy initialization.
> @@ -8799,9 +8812,11 @@ void Sema::AddInitializerToDecl(Decl *Re
>      // With declarators parsed the way they are, the parser cannot
>      // distinguish between a normal initializer and a pure-specifier.
>      // Thus this grotesque test.
> -    IntegerLiteral *IL;
> -    if ((IL = dyn_cast<IntegerLiteral>(Init)) && IL->getValue() == 0 &&
> -        Context.getCanonicalType(IL->getType()) == Context.IntTy)
> +    //
> +    // FIXME: The parser should instead treat anything that looks like a
> +    // pure-specifier as a pure-specifier, and Sema should convert it to
> an
> +    // initializer when necessary, rather than doing things this way
> around.
> +    if (!DirectInit && isZeroToken(*this, Init))
>        CheckPureMethod(Method, Init->getSourceRange());
>      else {
>        Diag(Method->getLocation(),
> diag::err_member_function_initialization)
>
> Modified: cfe/trunk/test/Parser/cxx-class.cpp
> URL:
> http://llvm.org/viewvc/llvm-project/cfe/trunk/test/Parser/cxx-class.cpp?rev=241019&r1=241018&r2=241019&view=diff
>
> ==============================================================================
> --- cfe/trunk/test/Parser/cxx-class.cpp (original)
> +++ cfe/trunk/test/Parser/cxx-class.cpp Mon Jun 29 18:19:23 2015
> @@ -24,6 +24,16 @@ public:
>    ; // expected-warning{{extra ';' inside a class}}
>
>    virtual int vf() const volatile = 0;
> +
> +  virtual int vf0() = 0l; // expected-error {{does not look like a
> pure-specifier}}
> +  virtual int vf1() = 1; // expected-error {{does not look like a
> pure-specifier}}
> +  virtual int vf2() = 00; // expected-error {{does not look like a
> pure-specifier}}
> +  virtual int vf3() = 0x0; // expected-error {{does not look like a
> pure-specifier}}
> +  virtual int vf4() = 0.0; // expected-error {{does not look like a
> pure-specifier}}
> +  virtual int vf5(){0}; // expected-error +{{}} expected-warning
> {{unused}}
> +  virtual int vf5a(){0;}; // function definition, expected-warning
> {{unused}}
> +  virtual int vf6()(0); // expected-error +{{}} expected-note +{{}}
> +  virtual int vf7() = { 0 }; // expected-error {{does not look like a
> pure-specifier}}
>
>  private:
>    int x,f(),y,g();
>
>
> _______________________________________________
> cfe-commits mailing list
> cfe-commits at cs.uiuc.edu
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20150629/7d616ee9/attachment.html>