On Mon, Sep 13, 2010 at 01:32:38PM +1000, matthew green wrote: >>>> Should we make perror() produce a warning so people will stop using it? >>> >>> No, as it is not a NetBSD specific API. >> >> Neither is gets() > > using perror() doesn't lead to fatal security flaws. > > i don't see any good reason to warn on it.
The man page should, however, at least mention warn() and perhaps note that warn() is preferred. I will attend to this. And, btw, the man page also claims that perror first appeared in 4.4BSD. Is this really correct? -- David A. Holland dholl...@netbsd.org
