Re: -Wignored-returns?

Re: -Wignored-returns?

[dewar]

<That's fair. Should the issue be taken up again, then? It would be a very
useful feature for us.
>

So why not submit a proposed patch?

-Wignored-returns?

[David E. Weekly]

GCC Team,

Hello! I saw in the GCC Bug mail archives a post that a certain Gray Watson
had proposed in 1992 for warning on ignored return values:
http://www.geocrawler.com/mail/msg.php3?msg_id=1968564&list=356. Such a
warning flag would prove invaluable for myself and for others seeking to
wholly replace lint with GCC warning flags and seems trivial to implement.
Is this feature already in GCC or has it yet to be added?

Yours,
 -david

Re: -Wignored-returns?

[Dale Johannesen]

On Thursday, December 6, 2001, at 04:11 PM, David E. Weekly wrote:

> GCC Team,
>
> Hello! I saw in the GCC Bug mail archives a post that a certain Gray 
> Watson
> had proposed in 1992 for warning on ignored return values:
> http://www.geocrawler.com/mail/msg.php3?msg_id=1968564&list=356. Such a
> warning flag would prove invaluable for myself and for others seeking to
> wholly replace lint with GCC warning flags and seems trivial to implement.
> Is this feature already in GCC or has it yet to be added?

See "Certain Changes We Don't Want To Make" in trouble.texi.

Re: -Wignored-returns?

[David E. Weekly]

Dale,

Thanks for your prompt response, even if it's one that makes me sad.

I don't really understand why GCC refuses to allow its users to specify this
option. I perfectly understand not including it in -W or -Wall, but not even
permitting this option to be made seems to be more about enforcing the
particular tastes of the GCC developers than it does actually providing a
useful compiler for users. ("Coming as I do from a Lisp background, I balk
at the idea that there is something dangerous about discarding a value" was
the excuse given.) It's rather patronizing to have the GCC developer team
tell us what coding style we want to use.

With the addition of this singular warning, we could completely eliminate
lint and lclint from our development environment, saving us a whole lot of
money. [sigh] It may be trickier without this simple feature.

All the same, I have the deepest of gratitude towards the GCC development
team for making an immensely powerful toolchain for nearly all of the
platforms out there.

 -david


PS: Dale, if GDB on OS/X 10.1.1 is itself segfaulting and getting "internal
error"s right and left, are you the man with whom to speak?


----- Original Message -----
From: "Dale Johannesen" <dalej@apple.com>
To: "David E. Weekly" <dweekly@legato.com>
Cc: "Dale Johannesen" <dalej@apple.com>; <gcc@gnu.org>
Sent: Thursday, December 06, 2001 4:22 PM
Subject: Re: -Wignored-returns?



On Thursday, December 6, 2001, at 04:11 PM, David E. Weekly wrote:

> GCC Team,
>
> Hello! I saw in the GCC Bug mail archives a post that a certain Gray
> Watson
> had proposed in 1992 for warning on ignored return values:
> http://www.geocrawler.com/mail/msg.php3?msg_id=1968564&list=356. Such a
> warning flag would prove invaluable for myself and for others seeking to
> wholly replace lint with GCC warning flags and seems trivial to implement.
> Is this feature already in GCC or has it yet to be added?

See "Certain Changes We Don't Want To Make" in trouble.texi.

Re: -Wignored-returns?

[Dale Johannesen]

On Thursday, December 6, 2001, at 04:46 PM, David E. Weekly wrote:

> Dale,
>
> Thanks for your prompt response, even if it's one that makes me sad.
>
> I don't really understand why GCC refuses to allow its users to specify 
> this
> option. I perfectly understand not including it in -W or -Wall, but not 
> even
> permitting this option to be made seems to be more about enforcing the
> particular tastes of the GCC developers than it does actually providing a
> useful compiler for users. ("Coming as I do from a Lisp background, I balk
> at the idea that there is something dangerous about discarding a value" 
> was
> the excuse given.) It's rather patronizing to have the GCC developer team
> tell us what coding style we want to use.

Personally I agree with you; I didn't write it.

> PS: Dale, if GDB on OS/X 10.1.1 is itself segfaulting and getting 
> "internal
> error"s right and left, are you the man with whom to speak?

No!  Try gdbhelp@group.apple.com

Re: -Wignored-returns?

[David E. Weekly]

Dale,

> Personally I agree with you; I didn't write it.

That's fair. Should the issue be taken up again, then? It would be a very
useful feature for us.

> Try gdbhelp@group.apple.com

Thanks for the pointer!

Yours,
 David

Re: -Wignored-returns?

[Per Bothner]

I think the text in trouble.texi is that
"we don't think the compiler should warn
about unused return values", but I don't
read it as a refusal to provide such an
option, as long as it is not the default,
nor implicitly enabled by other flags.

So I say go ahead and submit a patch.  If
the implementation is clean, I'd be in favor.
-- 
	--Per Bothner
per@bothner.com   http://www.bothner.com/per/