emacsclient breaks build on OS X 10.3.9

emacsclient breaks build on OS X 10.3.9

[David Reitter]

[-- Attachment #1.1: Type: text/plain, Size: 1448 bytes --]

It appears that a recent change (on or after Nov 2) has broken the  
build on my OS X 10.3.9 system:

gcc -fpascal-strings -DMAC_OSX -I../mac/src -DHAVE_CONFIG_H -I. -I../ 
src -I/Users/dr/Aquamacs/emacs/lib-src -I/Users/dr/Aquamacs/emacs/lib- 
src/../src  -fpascal-strings -DMAC_OSX -I../mac/src   -g -O2  /Users/ 
dr/Aquamacs/emacs/lib-src/emacsclient.c getopt.o getopt1.o - 
DVERSION="\"22.0.90\""  -o emacsclient
In file included from /Users/dr/Aquamacs/emacs/lib-src/emacsclient.c:45:
/usr/include/netinet/in.h:255: error: parse error before "in_addr_t"

full log here: http://homepages.inf.ed.ac.uk/dreitter/Aquamacs/emacs- 
build.log

It appears that this doesn't occur on more current OS X versions  
because they include this in their netinet/in.h:

----
#include <sys/_types.h>

#ifndef _IN_ADDR_T
#define _IN_ADDR_T
typedef	__uint32_t	in_addr_t;	/* base type for internet address */
#endif

#ifndef _IN_PORT_T
#define _IN_PORT_T
typedef	__uint16_t	in_port_t;
#endif
-----

which is not present in the 10.3.9 variant.

A recent revision is likely to be responsible:

revision 1.83
date: 2006-11-02 09:54:09 +0000;  author: lektu;  state: Exp;  lines:  
+7 -22
[WINDOWSNT]: Define HAVE_INET_SOCKETS.
[!WINDOWSNT]: Include <netinet/in.h> if available.
[HAVE_SOCKETS]: Also require HAVE_INET_SOCKETS.
(IOCTL, IOCTL_BOOL_ARG): Remove.
(set_tcp_socket): Don't set the socket in blocking mode.  Remove c_arg.


Hope that is specific enough.




[-- Attachment #1.2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 2454 bytes --]

[-- Attachment #2: Type: text/plain, Size: 142 bytes --]

_______________________________________________
Emacs-devel mailing list
Emacs-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/emacs-devel

Re: emacsclient breaks build on OS X 10.3.9

[Juanma Barranquero]

On 11/10/06, David Reitter <david.reitter@gmail.com> wrote:

> It appears that this doesn't occur on more current OS X versions
> because they include this in their netinet/in.h:
>
> ----
> #include <sys/_types.h>
>
> #ifndef _IN_ADDR_T
> #define _IN_ADDR_T
> typedef __uint32_t      in_addr_t;      /* base type for internet address */
> #endif

Could you please determine what is the right .h file to include in your setup?

Thanks,
                    /L/e/k/t/u

Re: emacsclient breaks build on OS X 10.3.9

[David Reitter]

On 10 Nov 2006, at 12:17, Juanma Barranquero wrote:

> On 11/10/06, David Reitter <david.reitter@gmail.com> wrote:
>
>> It appears that this doesn't occur on more current OS X versions
>> because they include this in their netinet/in.h:
>>
>> ----
>> #include <sys/_types.h>
>>
>> #ifndef _IN_ADDR_T
>> #define _IN_ADDR_T
>> typedef __uint32_t      in_addr_t;      /* base type for internet  
>> address */
>> #endif
>
> Could you please determine what is the right .h file to include in  
> your setup?

#include <sys/types.h>

does the job for me. You may want to enclose it in #ifndef in_addr_t  
to avoid side-effects on other systems at this point.

in_addr_t is mentioned in reference 2 below as "Not present on many  
pre-POSIX systems. Functions that otherwise use in_addr_t usually use  
int on those."
I guess that applies to OS X 10.3.9.


References:

[1] http://openvpn.net/archive/openvpn-devel/2002-05/msg00004.html
[2] http://daniel.haxx.se/projects/portability/


--
David Reitter
Blog: http://www.davids-world.com   Homepage: http://www.david- 
reitter.com

Re: emacsclient breaks build on OS X 10.3.9

[Juanma Barranquero]

On 11/10/06, David Reitter <david.reitter@gmail.com> wrote:

> #include <sys/types.h>
>
> does the job for me. You may want to enclose it in #ifndef in_addr_t
> to avoid side-effects on other systems at this point.

Is <sys/types.h> safe to include unconditionally? Quite a few src/*.c
modules do include it without protection. (src/buffer.c even includes
it twice ;)

                    /L/e/k/t/u

Re: emacsclient breaks build on OS X 10.3.9

[Kim F. Storm]

"Juanma Barranquero" <lekktu@gmail.com> writes:

> Is <sys/types.h> safe to include unconditionally?

Yes.

-- 
Kim F. Storm <storm@cua.dk> http://www.cua.dk

Re: emacsclient breaks build on OS X 10.3.9

[Juanma Barranquero]

On 11/10/06, Kim F. Storm <storm@cua.dk> wrote:

> > Is <sys/types.h> safe to include unconditionally?
>
> Yes.

Thanks.

OK, David, could you try again from the CVS?

                    /L/e/k/t/u

Re: emacsclient breaks build on OS X 10.3.9

[David Reitter]

On 10 Nov 2006, at 15:43, Juanma Barranquero wrote:

> On 11/10/06, Kim F. Storm <storm@cua.dk> wrote:
>
>> > Is <sys/types.h> safe to include unconditionally?
>>
>> Yes.
>
> Thanks.
>
> OK, David, could you try again from the CVS?

It compiled fine last night.

Re: emacsclient breaks build on OS X 10.3.9

[Juanma Barranquero]

On 11/11/06, David Reitter <david.reitter@gmail.com> wrote:

> It compiled fine last night.

OK, thanks.

                    /L/e/k/t/u