differ in signdness warnings.

differ in signdness warnings.

[Jan D.]

Hello.

With gcc >= 4.0 I get a lot of these when compiling Emacs:

coding.c:3471: varning: pointer targets in passing argument 1 of 
"safe_bcopy" differ in signedness

coding.c:3471: varning: pointer targets in passing argument 2 of 
"safe_bcopy" differ in signedness

frame.c:2107: varning: pointer targets in passing argument 1 of 
"strncmp" differ in signedness

and so on.

Is there anything said about what to do about these?  Should we ignore 
them or add casts where appropriate?  Introducing a SSDATA, which is 
(char *)SDATA, would be a good idea.

    Jan D.

Re: differ in signdness warnings.

[Romain Francoise]

"Jan D." <jan.h.d@swipnet.se> writes:

> Is there anything said about what to do about these?  Should we ignore
> them or add casts where appropriate?

I just compile with -Wno-pointer-sign.

-- 
Romain Francoise <romain@orebokech.com> | The sea! the sea! the open
it's a miracle -- http://orebokech.com/ | sea! The blue, the fresh, the
                                        | ever free! --Bryan W. Procter

Re: differ in signdness warnings.

[Eli Zaretskii]

> Date: Fri, 02 Dec 2005 16:03:40 +0100
> From: "Jan D." <jan.h.d@swipnet.se>
> 
> With gcc >= 4.0 I get a lot of these when compiling Emacs:
> 
> coding.c:3471: varning: pointer targets in passing argument 1 of 
> "safe_bcopy" differ in signedness
> 
> coding.c:3471: varning: pointer targets in passing argument 2 of 
> "safe_bcopy" differ in signedness
> 
> frame.c:2107: varning: pointer targets in passing argument 1 of 
> "strncmp" differ in signedness
> 
> and so on.
> 
> Is there anything said about what to do about these?

How about if we complain loudly to the GCC maintainers about these
silly warnings, and ask them to modify GCC so that it issues them only
under -Wall?

Re: differ in signdness warnings.

[Richard M. Stallman]

    I just compile with -Wno-pointer-sign.

How about making this the default?

Re: differ in signdness warnings.

[Eli Zaretskii]

> From: "Richard M. Stallman" <rms@gnu.org>
> Date: Fri, 02 Dec 2005 20:20:38 -0500
> Cc: emacs-devel@gnu.org
> 
>     I just compile with -Wno-pointer-sign.
> 
> How about making this the default?

That would add complexity that we should try to avoid: we will need to
add tests to the configury that probe whether GCC used for the build
supports this option.

I really think we should try to convince GCC maintainers to make
"-Wno-pointer-sign" be the default.  Then we need do nothing in Emacs.

Re: differ in signdness warnings.

[Jan Djärv]

Eli Zaretskii wrote:
>>Date: Fri, 02 Dec 2005 16:03:40 +0100
>>From: "Jan D." <jan.h.d@swipnet.se>
>>
>>With gcc >= 4.0 I get a lot of these when compiling Emacs:
>>
>>coding.c:3471: varning: pointer targets in passing argument 1 of 
>>"safe_bcopy" differ in signedness
>>
>>coding.c:3471: varning: pointer targets in passing argument 2 of 
>>"safe_bcopy" differ in signedness
>>
>>frame.c:2107: varning: pointer targets in passing argument 1 of 
>>"strncmp" differ in signedness
>>
>>and so on.
>>
>>Is there anything said about what to do about these?
> 
> 
> How about if we complain loudly to the GCC maintainers about these
> silly warnings, and ask them to modify GCC so that it issues them only
> under -Wall?

I am all for that!

	Jan D.

Re: differ in signdness warnings.

[Richard M. Stallman]

    How about if we complain loudly to the GCC maintainers about these
    silly warnings, and ask them to modify GCC so that it issues them only
    under -Wall?

I would support asking them to show a justification for this change.

Re: differ in signdness warnings.

[Eli Zaretskii]

> From: "Richard M. Stallman" <rms@gnu.org>
> CC: jan.h.d@swipnet.se, emacs-devel@gnu.org
> Date: Sat, 03 Dec 2005 10:58:30 -0500
> 
>     How about if we complain loudly to the GCC maintainers about these
>     silly warnings, and ask them to modify GCC so that it issues them only
>     under -Wall?
> 
> I would support asking them to show a justification for this change.

Could you perhaps actually write to them?  Based on my past (and
bitter) experience of talking to them about similar issues, it will
probably take your authority to get them to give any serious and
technical answers.

Re: differ in signdness warnings.

[Richard M. Stallman]

    That would add complexity that we should try to avoid: we will need to
    add tests to the configury that probe whether GCC used for the build
    supports this option.

That should not be hard using autoconf.

    I really think we should try to convince GCC maintainers to make
    "-Wno-pointer-sign" be the default.  Then we need do nothing in Emacs.

I will start by asking them what their reason was.