Hi Dave, Dave Korn wrote 16 April 2007 14:28:- > > On 16 April 2007 13:14, Adye, TJ (Tim) wrote: > > > I'm a bit surprised this isn't flagged by the compiler (it > > warns you about lots of other non-standard usages), though I guess it > > would be a bit ugly to implement (the compiler would have to > > distinguish between usage in a standard header and other code). > > Actually, the compiler already has a good deal of knowlege > about what constitute system headers and what constitute user > headers: see '-Wsystem-headers' in the docs. So I think that > it wouldn't actually be hard to implement, and I may even > take a look at doing so myself.
That would certainly be an interesting improvement. An alternative might be to distinguish between <> and "" includes. My concern was more that the determination of system headers is done in the preprocessor, while the type of identifier is determined later by the compiler. But I guess the latter knows about the former (eg. for error messages). It's a long time since I delved into the GCC sources... Tim. =========================== cut here ============================ Tim Adye [EMAIL PROTECTED] http://hepunx.rl.ac.uk/~adye BaBar/Atlas Groups, Particle Physics Dept, Rutherford Appleton Lab -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Problem reports: http://cygwin.com/problems.html Documentation: http://cygwin.com/docs.html FAQ: http://cygwin.com/faq/