No return statement does NOT result in a compile error!!!

Jos wrote:
> I have 1 question. If I create a XCode default C++ Command Line application
> and change the default main.cpp (remove the return statement) to:
>
> int main(int argc,char** arg) {
> }
>
> I do NOT get a compile time error on the missing return statement.

somebody else pointed out that main() is a special case. however, even in
regular functions, gcc by default does not give me errors or warnings for
missing return statements. i'm still using macosx 10.3.9 though, so that may
be part of it. what i did was add some text to the "other warning flags"
field of my projects' build styles, like this:

-Wreturn-type

that fixed it. it really should be turned on be default, though. lord only
knows why it isn't.

Allen Brunson <brunsona DeleteThis @takethisout.newsguy.com> writes:

> what i did was add some text to the "other warning flags"
> field of my projects' build styles, like this:
>
> -Wreturn-type

Even better:

-Wall

sherm--

--
Cocoa programming in Perl: http://camelbones.sourceforge.net
Hire me! My resume: http://www.dot-app.org

Sherm Pendley <sherm RemoveThis @dot-app.org> wrote:
> Allen Brunson <brunsona RemoveThis @takethisout.newsguy.com> writes:
>
>> what i did was add some text to the "other warning flags"
>> field of my projects' build styles, like this:
>>
>> -Wreturn-type
>
> Even better:
>
> -Wall

Better yet (IMO):

-W -Wall -Wno-unused-parameter

This turns on insane levels of warnings, and then disables
unused-parameter warnings because Cocoa really encourages them, so they
tend to get deeply annoying. Other than some altivec initializers, these
warnings have saved my goose more than once, and never been a major
hassle.

--
Michael Ash
Rogue Amoeba Software

In article <1129370497.774641.DeleteThis@nfs-db1.segnet.com>,
Michael Ash <mike.DeleteThis@mikeash.com> wrote:

> Sherm Pendley <sherm.DeleteThis@dot-app.org> wrote:
> > Allen Brunson <brunsona.DeleteThis@takethisout.newsguy.com> writes:
> >
> >> what i did was add some text to the "other warning flags"
> >> field of my projects' build styles, like this:
> >>
> >> -Wreturn-type
> >
> > Even better:
> >
> > -Wall
>
> Better yet (IMO):
>
> -W -Wall -Wno-unused-parameter
>
> This turns on insane levels of warnings, and then disables
> unused-parameter warnings because Cocoa really encourages them, so they
> tend to get deeply annoying. Other than some altivec initializers, these
> warnings have saved my goose more than once, and never been a major
> hassle.

You can avoid the unused parameter warnings also by putting
#pragma unused (parameter_name)
in the function body. It's not universal (although it is widespread) but
it shouldn't break any conformat compiler and provides an additional
level of documentation.

--
Goal 2005: Convincing James Hetfield to cover the Strawberry Shortcake
"Are You Berry Berry Happy?" song.

Michael Ash <mike.RemoveThis@mikeash.com> wrote:

> Sherm Pendley <sherm.RemoveThis@dot-app.org> wrote:
> > Allen Brunson <brunsona.RemoveThis@takethisout.newsguy.com> writes:
> >
> >> what i did was add some text to the "other warning flags"
> >> field of my projects' build styles, like this:
> >>
> >> -Wreturn-type
> >
> > Even better:
> >
> > -Wall
>
> Better yet (IMO):
>
> -W -Wall -Wno-unused-parameter
>
> This turns on insane levels of warnings, and then disables
> unused-parameter warnings because Cocoa really encourages them, so they
> tend to get deeply annoying. Other than some altivec initializers, these
> warnings have saved my goose more than once, and never been a major
> hassle.

And for gcc4.0 I think it's
-Wextra -Wall -Wno-unused-parameter

--
Frédéric Testuz
<mailto:ftestuz@SANSPUBmac.com.invalid>

Gregory Weston <uce RemoveThis @splook.com> writes:

> You can avoid the unused parameter warnings also by putting
> #pragma unused (parameter_name)
> in the function body.

Thanks for the tip - I'm much more comfortable with flagging specific cases
than with disabling the warning globally.

sherm--

--
Cocoa programming in Perl: http://camelbones.sourceforge.net
Hire me! My resume: http://www.dot-app.org

In article <uce-53BC31.09172015102005 RemoveThis @comcast.dca.giganews.com>,
Gregory Weston <uce RemoveThis @splook.com> wrote:

> In article <1129370497.774641 RemoveThis @nfs-db1.segnet.com>,
> Michael Ash <mike RemoveThis @mikeash.com> wrote:
>
> > Better yet (IMO):
> >
> > -W -Wall -Wno-unused-parameter
> >
> > This turns on insane levels of warnings, and then disables
> > unused-parameter warnings because Cocoa really encourages them, so they
> > tend to get deeply annoying.
>
> You can avoid the unused parameter warnings also by putting
> #pragma unused (parameter_name)
> in the function body. It's not universal (although it is widespread) but
> it shouldn't break any conformat compiler and provides an additional
> level of documentation.

Another way to do that is by commenting or leaving out the variable name
in the implementation:

void aFunction( int x, char * /* anAnusedVariable */)
{
...
}

I haven't seen a compiler that does not understand this.

Reinder

In article <reinder-210D5C.15442815102005.DeleteThis@reader28.wxs.nl>,
Reinder Verlinde <reinder.DeleteThis@verlinde.invalid> wrote:

> Another way to do that is by commenting or leaving out the variable name
> in the implementation:
>
> void aFunction( int x, char * /* anAnusedVariable */)
> {
> ...
> }
>
> I haven't seen a compiler that does not understand this.

I believe this is legal in C++ but not in C; don't know about Obj-C/C++

Ben

--
If this message helped you, consider buying an item
from my wish list: <http://artins.org/ben/wishlist>

I changed my name: <http://periodic-kingdom.org/People/NameChange.php>

in article reinder-210D5C.15442815102005.RemoveThis@reader28.wxs.nl, Reinder Verlinde
at reinder.RemoveThis@verlinde.invalid wrote on 10/15/05 6:44 AM:

> Another way to do that is by commenting or leaving out the variable name
> in the implementation:
>
> void aFunction( int x, char * /* anAnusedVariable */)
> {
> ...
> }
>
> I haven't seen a compiler that does not understand this.

Oddly enough, I just ran into one two days ago that can't handle exactly
that, when an I-thought-innocent checkin broke the other guys' compile, and
much angst ensued. You might have heard of their compiler, too, even if you
haven't seen it; it's called "Visual C++" and it's from a company called
"Microsoft".

.... only in .c files, mind you. But still.

In article <BF76BAE2.19457%alex@alexcurylo.com>,
Alex Curylo <alex RemoveThis @alexcurylo.com> wrote:

> in article reinder-210D5C.15442815102005 RemoveThis @reader28.wxs.nl, Reinder Verlinde
> at reinder RemoveThis @verlinde.invalid wrote on 10/15/05 6:44 AM:
>
> > Another way to do that is by commenting or leaving out the variable name
> > in the implementation:
> >
> > void aFunction( int x, char * /* anAnusedVariable */)
> > {
> > ...
> > }
> >
> > I haven't seen a compiler that does not understand this.
>
> Oddly enough, I just ran into one two days ago that can't handle exactly
> that, when an I-thought-innocent checkin broke the other guys' compile, and
> much angst ensued. You might have heard of their compiler, too, even if you
> haven't seen it; it's called "Visual C++" and it's from a company called
> "Microsoft".
>
> ... only in .c files, mind you. But still.

Well, you are never old enough to learn. Apparently, I have seen a
compiler tat does not understand that, but I haven't looked close enough
at it.

In article <macdev-AF4269.13280015102005 RemoveThis @pbdl.news.astraweb.com>, Ben
Artin <macdev RemoveThis @artins.org> wrote:

> I believe this is legal in C++ but not in C; don't know about Obj-C/C++

My guesstimate would be that it will be OK in Obj-C++, but not in Obj-C,
since, for all x, Obj-x is just x with a few additions.

Reinder

Hi,

Gregory Weston writes:
> You can avoid the unused parameter warnings also by putting
> #pragma unused (parameter_name)
> in the function body. It's not universal (although it is widespread) but

I use

(void) parameter_name;

Although I have seen compilers that give a (different) warning about
this, it's not quite as compiler specific as a #pragma.

benny