Re: c++/10527: [3.3/3.4 regression] confused error message with "new int()" parameter initializer

Re: c++/10527: [3.3/3.4 regression] confused error message with "new int()" parameter initializer

[giovannibajo]

Old Synopsis: [3.3/3.4 regression] ICE with illegal template: error reporting functions reentered
New Synopsis: [3.3/3.4 regression] confused error message with "new int()" parameter initializer

Responsible-Changed-From-To: unassigned->mmitchel
Responsible-Changed-By: bajo
Responsible-Changed-When: Wed Apr 30 13:52:42 2003
Responsible-Changed-Why:
    He fixed the ICE here, probably it's something related
State-Changed-From-To: closed->analyzed
State-Changed-By: bajo
State-Changed-When: Wed Apr 30 13:52:42 2003
State-Changed-Why:
    The problem with the wrong error message is still present:
    
    -----------------------------------------------
    struct Foo
    {};
    
    void Foo::NON_EXISTENT(int* val = new int())
    {}
    -----------------------------------------------
    pr10527.cpp:7: error: no `void Foo::NON_EXISTENT(int* = (operator new(unsigned
       int)(4), ((true, (((*<anonymous>) <unknown operator> 0), (<anonymous>
       <unknown operator> false))), <anonymous>)))' member function declared in
       class `Foo'
    
    
    This is a regression since 2.95, where we got:
    
    pr10527.cpp:7: no `void Foo::NON_EXISTENT(int * = new int)' member function 
    declared in class `Foo'
    
    
    Since it's simply a regression on error message formatting, I lower the priority.

http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view%20audit-trail&database=gcc&pr=10527