Nut upsdev - Mar 2013 - libconf branch build failure: sigemptyset()

Vaclav,

NUT is failing to build on OS X 10.6 and 10.7

With GCC/g++:

http://buildbot.networkupstools.org/public/nut/builders/MacOSX-10.6-x86/builds/340/steps/compile/logs/stdio

The same branch with clang/clang++:

/bin/sh ../libtool --tag=CXX   --mode=compile clang++ -DHAVE_CONFIG_H -I.
-I../include    -I../include -g -O2 -MT nutipc.lo -MD -MP -MF .deps/nutipc.Tpo
-c -o nutipc.lo nutipc.cpp
libtool: compile:  clang++ -DHAVE_CONFIG_H -I. -I../include -I../include -g -O2
-MT nutipc.lo -MD -MP -MF .deps/nutipc.Tpo -c nutipc.cpp  -fno-common -DPIC -o
.libs/nutipc.o
In file included from nutipc.cpp:22:
../include/nutipc.hpp:646:5: error: expected unqualified-id
                ::sigemptyset(&action.sa_mask);
                  ^
/usr/include/signal.h:127:26: note: expanded from macro 'sigemptyset'
#define sigemptyset(set)        (*(set) = 0, 0)
                                ^

sigemptyset() probably should have been implemented as an inline function, but
that's what we have to work with.

This particular error was masked by a problem with Makefile.am; see
https://github.com/networkupstools/nut/commit/4290a0d0955b8437c57941cc64e2c941ab2d2094

I'd recommend dropping the :: for C library functions - if there is a method
in another namespace whose name collides with the C library, there's bound
to be confusion when someone is reading through the code.

-- 
Charles Lepple
clepple at gmail

?Hello Charles,

thanks for the note, I'll take a look at it ASAP.
I have one virtual OS X myself; I guess I'll install devel.
packages and check for more such things...

I've qualified the C functions on purpose because
of the possibility to be confused with a local routine;
however if it's implemented as a macro, there's no
way aroud that :-(

Do you by any chance have a list of package names
for the OS X (mine is just a fresh clean installation)?
I'd also appreciate an advice on the packages source and
an overview of OS X packaging system (I suppose it has some
but I've never worked with it, before).
Am I correct in thinking that it's you who owns the Darwin
machine used as NUT buildslave?

Thanks again,

Regards,

vasek





-----------------------------
Eaton Elektrotechnika s.r.o. ~ S?dlo spolecnosti, jak je zaps?no v rejstr?ku:
Kom?rovsk? 2406, Praha 9 - Horn? Pocernice, 193 00, Cesk? Republika ~ Jm?no,
m?sto, kde byla spolecnost zaregistrov?na: Praha ~ Identifikacn? c?slo (ICO):
498 11 894

-----------------------------

-----Original Message-----
From: Charles Lepple [mailto:clepple at gmail.com] 
Sent: Sunday, March 10, 2013 9:43 PM
To: Krpec, Vaclav
Cc: NUT Developers
Subject: libconf branch build failure: sigemptyset()

Vaclav,

NUT is failing to build on OS X 10.6 and 10.7

With GCC/g++:

http://buildbot.networkupstools.org/public/nut/builders/MacOSX-10.6-x86/builds/340/steps/compile/logs/stdio

The same branch with clang/clang++:

/bin/sh ../libtool --tag=CXX   --mode=compile clang++ -DHAVE_CONFIG_H -I.
-I../include    -I../include -g -O2 -MT nutipc.lo -MD -MP -MF .deps/nutipc.Tpo
-c -o nutipc.lo nutipc.cpp
libtool: compile:  clang++ -DHAVE_CONFIG_H -I. -I../include -I../include -g -O2
-MT nutipc.lo -MD -MP -MF .deps/nutipc.Tpo -c nutipc.cpp  -fno-common -DPIC -o
.libs/nutipc.o In file included from nutipc.cpp:22:
../include/nutipc.hpp:646:5: error: expected unqualified-id
                ::sigemptyset(&action.sa_mask);
                  ^
/usr/include/signal.h:127:26: note: expanded from macro 'sigemptyset'
#define sigemptyset(set)        (*(set) = 0, 0)
                                ^

sigemptyset() probably should have been implemented as an inline function, but
that's what we have to work with.

This particular error was masked by a problem with Makefile.am; see
https://github.com/networkupstools/nut/commit/4290a0d0955b8437c57941cc64e2c941ab2d2094

I'd recommend dropping the :: for C library functions - if there is a method
in another namespace whose name collides with the C library, there's bound
to be confusion when someone is reading through the code.

--
Charles Lepple
clepple at gmail