Re: problem with def of inet_ntop() in git-compat-util.h as well as other places

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Wed, Aug 27, 2014 at 03:15:05PM -0400, dev wrote:

> This causes a problem on things like Solaris :
> 
>  * new build flags
>     CC credential-store.o
> "git-compat-util.h", line 516: error: identifier redeclared: inet_ntop
>         current : function(int, pointer to const void, pointer to char,
> unsigned long) returning pointer to const char
>         previous: function(int, pointer to const void, pointer to char,
> unsigned int) returning pointer to const char :
> "/usr/include/arpa/inet.h", line 68
> cc: acomp failed for credential-store.c
> gmake: *** [credential-store.o] Error 2

That declaration should only be used if you have NO_INET_NTOP defined by
the Makefile. That is not defined by default for Solaris. Have you
specified it yourself, or are you using the autoconf script? If the
latter, I suspect its test for inet_ntop needs to be fixed.

> However ran into a problem, again, with compat/inet_ntop.c which seems
> to be not needed at all since inet_ntop() handles both IPv6 and IPv4
> just fine.   Really I don't see why this file gets carted around so much
> as it is even in the Apache svn codebase as well.

Again, that should not be compiled at all unless you have NO_INET_NTOP
set. Fixing that should solve both of your problems.

> Also the Makefile's generated are all borked full of GCCism "CFLAGS = -g
> -O2 -Wall"  which means very little on some OS wherein the gcc compiler
> is not the default.

Yes, this is a potential portability problem we've discussed before, but
literally nobody has ever complained. I suspect it's a combination of
many compilers accepting those arguments (e.g., clang is fine with it)
and people on exotic compilers accepting that "make CFLAGS=" is a
reasonable starting point (you can also put "CFLAGS = " into your
config.mak if you do not want to remember to include it on each make
invocation).

> is Solaris and am using ORacle Studio 12.3 compilers and therefore the
> CFLAGS in the Makefiles are just silly.  Lastly, the linkage to libintl
> should look in /usr/local/lib if the LD_LIBRARY_PATH and other env vars
> are setup correctly. However the Makefile's seem to miss this fact and
> -lintl needs to be manually hacked into place.

I do not usually see Makefiles peeking at LD_LIBRARY_PATH, which is for
runtime resolution. Do you need to set LDFLAGS in your environment (or
in config.mak)?

-Peff
--
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




[Index of Archives]     [Linux Kernel Development]     [Gcc Help]     [IETF Annouce]     [DCCP]     [Netdev]     [Networking]     [Security]     [V4L]     [Bugtraq]     [Yosemite]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux RAID]     [Linux SCSI]     [Fedora Users]