Re: lint warning about extern (was: Re: CVS commit: src)

2023-03-28 Thread Valery Ushakov
On Tue, Mar 28, 2023 at 22:13:25 +0200, Roland Illig wrote:

> Am 28.03.2023 um 19:52 schrieb Valery Ushakov:
> > On Tue, Mar 28, 2023 at 14:44:36 +, Roland Illig wrote:
> > 
> > > lint: warn about extern declarations outside headers
> > 
> > vs.
> > 
> > > https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html
> > 
> > which is nested extern declaration inside a function body.
> > 
> > I.e. the two lines of this log message contradict each other.  So
> > which one of those does this change actually address?
> 
> Oops, I implemented an entirely unrelated warning in message 351, which
> is about a function or object declaration that is outside a header and
> thus may lead to inconsistent definitions when it is declared
> differently in several translation units. This message helps finding
> cases where identifiers are needlessly exported instead of being
> declared 'static'.
> 
> I now added another message 352 that covers the case from the
> discussion, which is about declarations inside function bodies with
> storage class 'extern'.
> 
> Sorry for the confusion.

Thanks for the clarification and a quick follow up!

-uwe


lint warning about extern (was: Re: CVS commit: src)

2023-03-28 Thread Roland Illig

Am 28.03.2023 um 19:52 schrieb Valery Ushakov:

On Tue, Mar 28, 2023 at 14:44:36 +, Roland Illig wrote:


lint: warn about extern declarations outside headers


vs.


https://mail-index.netbsd.org/tech-userlevel/2023/03/15/msg013727.html


which is nested extern declaration inside a function body.

I.e. the two lines of this log message contradict each other.  So
which one of those does this change actually address?


Oops, I implemented an entirely unrelated warning in message 351, which
is about a function or object declaration that is outside a header and
thus may lead to inconsistent definitions when it is declared
differently in several translation units. This message helps finding
cases where identifiers are needlessly exported instead of being
declared 'static'.

I now added another message 352 that covers the case from the
discussion, which is about declarations inside function bodies with
storage class 'extern'.

Sorry for the confusion.

Roland