Date: Wed, 05 Jan 2000 12:45:11 -0800 From: "Ronald F. Guilmette" <rfg@monkeys.com> To: Thomas David Rivers <rivers@dignus.com> Cc: cracauer@cons.org, hackers@FreeBSD.ORG Subject: Re: [OFFTOPIC] alt. C compiler Message-ID: <1814.947105111@monkeys.com> In-Reply-To: Your message of Wed, 05 Jan 2000 15:27:28 -0500. <200001052027.PAA32580@lakes.dignus.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message <200001052027.PAA32580@lakes.dignus.com>, Thomas David Rivers <rivers@dignus.com> wrote: >> But pragmatically, it sure would be nice if I (or you) as a programmer >> developing stuff on FreeBSD could include various of the FreeBSD include >> files into any program that we happen to be working on, and then fire up >> the compiler (with our own personal favorite set of command line options) >> and then _not_ be plagued by a whole bunch of spurious warnings and/or >> errors that have noting actually to do with _our_ code. >> >> This isn't about standards conformance. It is about providing a top >> quality _complete_ software developement system/environment. >> > > Ok - i just wanted to be clear what exactly we're talking about. > > We're talking about making it easier to use the FreeBSD library code > in other programs (i.e. `port' it somewhere) - presumably with a > strict ANSI C compiler other than gcc. > > While that is certainly a laudable goal - I question the term "pragmatic". > Just how often is this going to happen? Is it worth the effort? > [These are questions that likely should be asked... - I'm just the > Devil's advocate here..., personally - I agree... it would be nice > to have "ANSI-clean" header files.] It would be more than `nice'. I consider it critical. And this has nothing to do with ``porting'' FreeBSD's library code... well... very little anyway. It mostly has to do with the portability of _my_ application code, and your's, and his, and everybody's. Anyone who is developing code that they *really* want to be ultra-portable and who is using gcc during the development *will* use the -pedantic-errors gcc compiler option for all of his/her compiles. I do this routinely. Often however, when doing that, I get spurious errors (relating to some system include file or another) that has nothing to do with _my_ code whatsoever. And that is most annoying when it happens. Keep in mind also that we are only talking now about ANSI conformance problems with the system include files. But as the list of system include file problems I just posted shows, some of the FreeBSD 3.3 system include files have a some even more glaring errors that go well beyond being mere ANSI conformance issues... like all those cases where header X tries to include header Y, and where the file Y doesn't even exist. Those are outright bugs in the relevant include files that will bite you no matter what compilation options you are using. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1814.947105111>