------- Additional Comments From joseph at codesourcery dot com 2005-03-11 20:01 ------- Subject: Re: Mismatch in pointer indirection level should give specialized warning
On Fri, 11 Mar 2005, pinskia at gcc dot gnu dot org wrote: > Really in my mind, we should be rejecting this invalid code instead of just > warning. I should point out that some cases of this are common in real code: various functions declared in system headers take different pointer argument types on different systems and software may not always get the configuration right on all systems. It could arguably be specialised into different cases: if the only differences are qualifiers, signedness and differences between integer types (e.g. int versus long) of the same width or between character types and void then it is likely less dangerous in practice than if there are bigger differences, so one could give a pedwarn and the other an error. But equally such distinctions between when this is a warning and when it is an error might just serve to confuse users. -- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20422
