Samisa:
The latest, and final, artifacts are uploaded here:
http://people.apache.org/~samisa/release/1.0.0/final/
Please test. I plan to send the release note in about 12 hours form
now.
I have tested all the sources and binaries successfully.
Thanks for all the quick fixes! The configure --help notes and
the INSTALL files are much clearer. I successfully built and
installed and tested with --enable-tests, and with the samples
directory.
Two things, one small, one maybe not:
The small thing is that is seems like the
test/core/clientapi/test_clientapi program fails (with a message now,
thank you!) unless AXIS2C_HOME points to the installed location.
(I'm guessing it's looking for axis2.xml.) This means you really
should run "make && make install", then set
AXIS2C_HOME=/path/to/install,
and then "make check". That's a little different than the usual
"make && make check && make install" process.
We should perhaps make that clean in the main INSTALL file, or else
alter test_clientapi.c somehow so that it will work if you set
AXIS2C_HOME=/path/to/source or something. Then the more
typical "check then install" process would function, so long as you
had AXIS2C_HOME set to the location of the Axis2/C sources.
The maybe not-so-small thing is that for me, at least, on an
older RedHat Linux box, util/test/util/test_thread fails sometimes
and not other times. That feels like a race condition problem to me.
(This is regardless of what AXIS2C_HOME is set to, or even if it's
unset; that was a red herring in my previous email.)
Here's the output from an unsuccessful run:
success - thread_init - axutil_thread_once_init success - thread_init
- axutil_thread_mutex_create success - test_axutil_thread_create -
axutil_thread_create success - test_axutil_thread_create -
axutil_thread_create thread data = 5 x = 1 thread data = 25 x = 2
success - test_axutil_thread_create - axutil_thread_join Segmentation
fault
Here's another, different one:
success - thread_init - axutil_thread_once_init success - thread_init
- axutil_thread_mutex_create thread data = 5 x = 1 success -
test_axutil_thread_create - axutil_thread_create success -
test_axutil_thread_create - axutil_thread_create Segmentation fault
Here's a backtrace from a failure like the first one above:
(gdb) bt
#0 0x003854bd in axutil_thread_join (thd=0x80552b0) at
thread_unix.c:158
#1 0x08048d27 in test_axutil_thread_detach (env=0x80551e0)
at test_thread.c:132
#2 0x08048fdb in create_env_with_error_log () at test_thread.c:260
#3 0x0804915d in __libc_csu_fini ()
from
/tmp/chrisd_build/axis2c-src-1.0.0/util/test/util/.libs/test_thread
#4 0x00125e23 in __libc_start_main () from /lib/tls/libc.so.6
#5 0x080489e1 in __do_global_dtors_aux ()
from
/tmp/chrisd_build/axis2c-src-1.0.0/util/test/util/.libs/test_thread
(gdb) frame
#0 0x003854bd in axutil_thread_join (thd=0x80552b0) at
thread_unix.c:158
158 if (0 == pthread_join(*(thd->td), (void *)(&thread_stat)))
(gdb) print *(thd->td)
Cannot access memory at address 0x62696c2f
I'm not sure if that helps, but it looks to me like something
must be getting deallocated at the wrong time here.
Chris.