Well, CVS trunk is work in progress, so it is generally not recommended for non-developers. You should at least subscribe to serdev as it during substantial changes may not even compile or be consistent. People like you who are willing to test are very welcome :-) See http://www.iptel.org/ser/development/documentation.

If you have found issues with the rel_2_0_0 branch of the CVS, the tracker is the right place to make sure that the right developer picks it up. This holds for suspicions as well. It's better to register one bug too much...
g-)

Mike Trest - Personal wrote:
To All,
Thanks for the suggestions.

I have complete a working build from CVS on RH environment.

I had to change my hat which said "Enterprise Tester" and put on the hat
that says  "Software Developer" and read thru the code in CVS tree.  I
was able to chase down the issues with  "db_key_t"  and a few other
db things.  I discovered that the CVS tree was NOT (as I had expected)
in a state where "everything is in order".  I discovered a lot of the DB
material in a state of change between previous generations to the new
2.0 / 2.1 generation.

I took Jiri's suggestion and built a config based on the ser-oof.cfg .
I only got into issues when I added other modules not on that specific
loadmodule list.

I am still waiting for input from the more experienced developers as to
exactly what to include in The Bundle.

I am still hopeful.

Regards,  ..mike..





Date: Fri, 22 Feb 2008 10:59:15 -0500
To: serusers@lists.iptel.org
From: Mike Trest - Personal <Mike@Trest.COM>
Subject: [Serusers]  SIP Express Bundle ready-to-go idea, BUILD FROM  CVS on RH ?

Hi,
Followup on prior message:

When doing  make .... modules I found a similar error that did not involve a system include.

The offending line in avpops_db.h is:

int db_store_avp( db_key_t *keys, db_val_t *vals, int n, char *table);

Both examples below and this example are related to forward declarations
with pointers to structures.

Regards,  ..mike..





make[1]: Entering directory `/usr/local/src/ser-2.0.x-CVS/modules/avpops'
gcc -fPIC -DPIC -g -O9 -funroll-loops  -Wcast-align  -minline-all-stringops -falign-loops -ftree-vectorize -mtune=athlon64   -Wall     -DNAME='"ser"' -DVERSION='"2.1.0-dev17-tcp"' -DARCH='"i386"' -DOS='linux_' -DOS_QUOTED='"linux"' -DCOMPILER='"gcc 4.1.1"' -D__CPU_i386 -D__OS_linux -DSER_VER=2001000 -DCFG_DIR='"/usr/local/etc/ser/"' -DMODS_DIR='"/usr/local/lib/ser/modules/"' -DPKG_MALLOC -DSHM_MEM  -DSHM_MMAP -DDNS_IP_HACK -DUSE_IPV6 -DUSE_MCAST -DUSE_TCP -DDISABLE_NAGLE -DHAVE_RESOLV_RES -DUSE_DNS_CACHE -DUSE_DNS_FAILOVER -DUSE_DST_BLACKLIST -DUSE_NAPTR  -DUSE_TLS -DTLS_HOOKS -DFAST_LOCK -DADAPTIVE_WAIT -DADAPTIVE_WAIT_LOOPS=1024  -DCC_GCC_LIKE_ASM -DHAVE_GETHOSTBYNAME2 -DHAVE_UNION_SEMUN -DHAVE_SCHED_YIELD -DHAVE_MSG_NOSIGNAL -DHAVE_MSGHDR_MSG_CONTROL -DHAVE_ALLOCA_H -DHAVE_TIMEGM -DHAVE_SCHED_SETSCHEDULER -DHAVE_EPOLL -DHAVE_SIGIO_RT -DSIGINFO64_WORKARROUND -DUSE_FUTEX -DHAVE_SELECT -c avpops.c -o avpops.o
In file included from avpops_impl.h:41,
                 from avpops_parse.h:38,
                 from avpops.c:50:
avpops_db.h:66: error: expected â)â before â*â token
make[1]: *** [avpops.o] Error 1
make[1]: Leaving directory `/usr/local/src/ser-2.0.x-CVS/modules/avpops'
make: *** [modules] Error 1


========================================================================
Hi,
I am doing a ser build from CVS check-out on a RH based linux with gcc 4.1.1.
The build commands are this.
make clean
make proper
make group_include="standard standard-dep stable mysql"
When I get to any core or module section code which included futexlock.h header, I get the following:

gcc -g -O9 -funroll-loops -Wcast-align -minline-all-stringops -falign-loops -ftree-vectorize -mtune=athlon64 -Wall     -DNAME='"ser"' -DVERSION='"2.1.0-dev17-tcp"' -DARCH='"i386"' -DOS='linux_' -DOS_QUOTED='"linux"' -DCOMPILER='"gcc 4.1.1"' -D__CPU_i386 -D__OS_linux -DSER_VER=2001000 -DCFG_DIR='"/usr/local/etc/ser/"' -DMODS_DIR='"/usr/local/lib/ser/modules/"' -DPKG_MALLOC -DSHM_MEM  -DSHM_MMAP -DDNS_IP_HACK -DUSE_IPV6 -DUSE_MCAST -DUSE_TCP -DDISABLE_NAGLE -DHAVE_RESOLV_RES -DUSE_DNS_CACHE -DUSE_DNS_FAILOVER -DUSE_DST_BLACKLIST -DUSE_NAPTR  -DUSE_TLS -DTLS_HOOKS -DFAST_LOCK -DADAPTIVE_WAIT -DADAPTIVE_WAIT_LOOPS=1024  -DCC_GCC_LIKE_ASM -DHAVE_GETHOSTBYNAME2 -DHAVE_UNION_SEMUN -DHAVE_SCHED_YIELD -DHAVE_MSG_NOSIGNAL -DHAVE_MSGHDR_MSG_CONTROL -DHAVE_ALLOCA_H -DHAVE_TIMEGM -DHAVE_SCHED_SETSCHEDULER -DHAVE_EPOLL -DHAVE_SIGIO_RT -DSIGINFO64_WORKARROUND -DUSE_FUTEX -DHAVE_SELECT -c core_cmd.c -o core_cmd.o
In file included from mem/../futexlock.h:48,
                 from mem/../lock_ops.h:85,
                 from mem/shm_mem.h:76,
                 from core_cmd.c:32:
/usr/include/linux/futex.h:96: error: expected â)â before â*â token
/usr/include/linux/futex.h:98: error: expected â)â before â*â token
make: *** [core_cmd.o] Error 1
=========================================================================

The offending lines came from the  system header  <futex.h>.  The offending lines are:

long do_futex(u32 *uaddr, int op, u32 val, unsigned long timeout, u32 *uaddr2, u32 val2, u32 val3);

extern int handle_futex_death(u32 *uaddr, struct task_struct *curr, int pi);

By pointing to an alternate version of <futex.h> where these lines are commented out
I can compile all the way thru with non-fatal warnings.

My questions are:
1) can anyone tell me what are the equivalent lines in the DEB version where
this code/compile was most recently developed?
2) has anyone got a fix for this error?

..mike..




_______________________________________________
Serusers mailing list
Serusers@lists.iptel.org
http://lists.iptel.org/mailman/listinfo/serusers

_______________________________________________ Serusers mailing list Serusers@lists.iptel.org http://lists.iptel.org/mailman/listinfo/serusers