home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.std.c
- Path: sparky!uunet!spool.mu.edu!umn.edu!csus.edu!netcom.com!rfg
- From: rfg@netcom.com (Ronald F. Guilmette)
- Subject: Re: Standard conformance and GCC 2.3.3
- Message-ID: <1993Jan1.115804.23071@netcom.com>
- Organization: Netcom Online Communications Services (408-241-9760 login: guest)
- References: <1992Dec30.173634.15487@crd.ge.com> <1992Dec30.203555.3615@netcom.com> <1992Dec31.150047.22321@crd.ge.com>
- Date: Fri, 1 Jan 1993 11:58:04 GMT
- Lines: 45
-
- In article <1992Dec31.150047.22321@crd.ge.com> volpe@ausable.crd.ge.com writes:
- +In article <1992Dec30.203555.3615@netcom.com>, rfg@netcom.com (Ronald F. Guilmette) writes:
- +|> In article <1992Dec30.173634.15487@crd.ge.com> volpe@ausable.crd.ge.com writes:
- +|> +These diagnostics are prefectly friendly and very helpful. Since the
- +|> +declarations allocate no storage (there's no object being defined), a
- +|> +storage class specifier is useless and superfluous.
- +|>
- +|> I disagree violently with the assertion that the issuance of errors for
- +|> perfecly "standard conformant" code is "perfectly friendly and very helpful".
- +|> Anyone who says such a thing obviously does not port code for a living!
- +
- +I disagree violently with the assertion that any errors one might make
- +would necessarily result in non-standard-conformant code. Just because
- +it's conformant doesn't mean it's intended. Anyone who says such a thing
- +obviously never writes original code for a living, or otherwise is a
- +complete programming God who has never made a mistake and never found
- +lint to be the least bit useful.
-
- There is a big difference between using an optional tool like lint, and/or
- having warnings which you can *ignore* at your option, and getting *hard*
- errors from your compiler for perfectly standard conformant code.
-
- Lint is good. Warnings are good. Hard errors for perfectly standard
- conformant code are BAD! I do not wish to have my compiler give me s**t
- when I'm trying to port some perfectly standard conforming code. Nor do
- I think that it should be forcing me to change such code, just to get
- it to compile and link.
-
- +I would *never* write the above piece of code intentionally, no matter
- +how conformant it is, and I would appreciate some kind of warning that
- +I had obviously omitted some variable names in those declarations.
-
- As I say, warnings are one thing, and hard errors are another. Please
- do not confuse the two.
-
- --
-
- // Ron ("Loose Cannon") Guilmette uucp: ...uunet!lupine!segfault!rfg
- //
- // "On the one hand I knew that programs could have a compelling
- // and deep logical beauty, on the other hand I was forced to
- // admit that most programs are presented in a way fit for
- // mechanical execution, but even if of any beauty at all,
- // totally unfit for human appreciation."
- // -- Edsger W. Dijkstra
-