This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] Fix -Wdisallowed-function-list=* (PR c++/39554)


Mark Mitchell <mark@codesourcery.com> writes:

> Ian Lance Taylor wrote:
>
>> I don't think it's very interesting to disallow one version of an
>> overloaded function while permitting other versions.
>
> Why not?  If the point is to avoid using the function, it seems just as
> possible to me that one of the set of overloads has been declared evil
> (perhaps, for example, it's API doesn't permit safe usage), while
> another replacement function is valid.

Well, that seems sort of theoretical to me.  But I agree that in theory
it is possible.

Of course, we then would require a way to name either a complete set of
overloads or a specific overload.  I suppose the easiest way to do that
would be to permit a regexp, and compare it against the full demangled
name.  Is the full demangled name readily available?  cc1 doesn't
currently use regexes, but the gcc driver does, and the libiberty
support is not large.

Or I'm willing to let this option die, and be replaced by a plugin.  I
wanted to object to letting it die because there is a different way to
do something different.  That doesn't seem like an appropriate way to do
things.

Ian


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]