August 2010 - h323plus - lists.packetizer.com

Configure problem?
by Francisco Olarte (M) 11 Aug '10

11 Aug '10

With a fresh CVS h323plus checkout configure emits ERROR: you must first run 'aclocal' in /home/folarte/tmp/h323plus/plugins It's trivial to solve, just '(cd plugins && aclocal)' solves it, but I wonder why I have this message. The log of what I did is at end. Regards. Francisco Olarte. --- LOG, PTlib 2.8.2 compiled & installed with prefix=$HOME, h323plus checked out from cvs ( OS is folarte@gkaps ~/tmp/h323plus $ export PTLIBDIR=~/src/ptlib-2.8.2 folarte@gkaps ~/tmp/h323plus $ export PWLIBDIR=$PTLIBDIR folarte@gkaps ~/tmp/h323plus $ ./configure --prefix=$HOME \ > --disable-h46018 \ > --disable-h46023 \ > --disable-mysql \ > --disable-pgsql \ > --disable-firebird \ > --disable-unixodbc \ > --disable-sqlite \ > --disable-radius checking for g++... g++ checking for C++ compiler default output file name... a.out checking whether the C++ compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... checking for suffix of object files... o checking whether we are using the GNU C++ compiler... yes checking whether g++ accepts -g... yes checking for gcc... gcc checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking for gcc option to accept ISO C89... none needed checking how to run the C preprocessor... gcc -E checking for grep that handles long lines and -e... /bin/grep checking for egrep... /bin/grep -E checking for ANSI C header files... yes OpenH323 version is 1.22.0 checking for /home/folarte/src/ptlib-2.8.2/version.h... yes PTLib prefix set to.... /home/folarte/src/ptlib-2.8.2 checking Checking PTLib version 2.8.2... ok configure: Disabling H.350 configure: Disabling H.460.18/.19 configure: Disabling H.460.23/.24 configure: Enabling T.38 configure: Disabling GnuGk checking for a BSD-compatible install... /usr/bin/install -c configure: creating ./config.status config.status: creating openh323u.mak config.status: creating Makefile config.status: creating include/openh323buildopts.h === configuring in plugins (/home/folarte/tmp/h323plus/plugins) configure: running /bin/sh ./configure --disable-option-checking '--prefix=/home/folarte' '--disable-h46018' '--disable-h46023' '--disable-mysql' '--disable-pgsql' '--disable-firebird' '--disable-unixodbc' '--disable-sqlite' '--disable-radius' --cache-file=/dev/null --srcdir=. ERROR: you must first run 'aclocal' in /home/folarte/tmp/h323plus/plugins folarte@gkaps ~/tmp/h323plus $ uname -a Linux gkaps 2.6.34-gentoo-r1 #1 SMP Wed Jul 21 13:57:48 CEST 2010 x86_64 Intel(R) Xeon(R) CPU E5530 @ 2.40GHz GenuineIntel GNU/Linux folarte@gkaps ~/tmp/h323plus $ gcc -v Using built-in specs. Target: x86_64-pc-linux-gnu Configured with: /var/tmp/portage/sys-devel/gcc-4.4.3-r2/work/gcc-4.4.3/configure --prefix=/usr --bindir=/usr/x86_64-pc-linux-gnu/gcc-bin/4.4.3 --includedir=/usr/lib/gcc/x86_64-pc-linux-gnu/4.4.3/include --datadir=/usr/share/gcc-data/x86_64-pc-linux-gnu/4.4.3 --mandir=/usr/share/gcc-data/x86_64-pc-linux-gnu/4.4.3/man --infodir=/usr/share/gcc-data/x86_64-pc-linux-gnu/4.4.3/info --with-gxx-include-dir=/usr/lib/gcc/x86_64-pc-linux-gnu/4.4.3/include/g++-v4 --host=x86_64-pc-linux-gnu --build=x86_64-pc-linux-gnu --disable-altivec --disable-fixed-point --without-ppl --without-cloog --enable-nls --without-included-gettext --with-system-zlib --disable-checking --disable-werror --enable-secureplt --enable-multilib --disable-libmudflap --disable-libssp --disable-libgomp --with-python-dir=/share/gcc-data/x86_64-pc-linux-gnu/4.4.3/python --disable-libgcj --enable-languages=c,c++ --enable-shared --enable-threads=posix --enable-__cxa_atexit --enable-clocale=gnu --with-bugurl=http://bugs.gentoo.org/ --with-pkgversion='Gentoo 4.4.3-r2 p1.2' Thread model: posix gcc version 4.4.3 (Gentoo 4.4.3-r2 p1.2)

2 2

Incorrect include directive in h323caps.h
by Francisco Olarte (M) 10 Aug '10

10 Aug '10

Trying to compile an app (yate3) with h323plus 2010-05-25 cvs ( from gnugk site ) I've got the following error: >>> In file included from /home/folarte/include/openh323/h323con.h:555, from /home/folarte/include/openh323/h323.h:492, from h323chan.cpp:32: /home/folarte/include/openh323/h323caps.h:289:31: error: ../include/codecs.h: No such file or directory <<< My setup is ptlib 2.8.2, 'configure --prefix=$HOME', make opt, make install. h323plus 2010-05-25 cvs, 'configure ..... --prefix=$HOME', make opt, make install. I end up having the following include directory: folarte@gkaps ~ $ ls -l ~/include total 44 drwxr-xr-x 8 folarte folarte 4096 Aug 10 18:12 openh323 -r--r--r-- 1 folarte folarte 23668 Aug 10 18:08 ptbuildopts.h drwxr-xr-x 2 folarte folarte 4096 Aug 10 18:08 ptclib drwxr-xr-x 3 folarte folarte 4096 Aug 10 18:08 ptlib -r--r--r-- 1 folarte folarte 4794 Aug 10 18:08 ptlib.h folarte@gkaps ~ $ find include -name h323caps.h -ls 1056359 104 -r--r--r-- 1 folarte folarte 105274 Aug 10 18:12 include/openh323/h323caps.h folarte@gkaps ~ $ find include -name codecs.h -ls 1056338 44 -r--r--r-- 1 folarte folarte 44804 Aug 10 18:12 include/openh323/codecs.h And the compilation command issued by yate seems correct: make[1]: Entering directory `/home/folarte/src/yate/modules' g++ -Wall -I.. -I.. -O2 -fno-check-new -fno-exceptions -fPIC -DHAVE_GCC_FORMAT_CHECK -finline -Winline -export-dynamic -shared -Wl,--unresolved-symbols=ignore-in-shared-libs -Wl,--retain-symbols-file,/dev/null -o h323chan.yate -DPHAS_TEMPLATES -D_REENTRANT -DP_HAS_SEMAPHORES -I/home/folarte/include/ptlib -I/home/folarte/include/openh323 h323chan.cpp -L/home/folarte/lib -lopenh323 -L/home/folarte/lib -lpt -L.. -lyate In file included from /home/folarte/include/openh323/h323con.h:555, from /home/folarte/include/openh323/h323.h:492, from h323chan.cpp:32: /home/folarte/include/openh323/h323caps.h:289:31: error: ../include/codecs.h: No such file or directory In file included from /home/folarte/include/openh323/channels.h:190, from /home/folarte/include/openh323/h323caps.h:290, from /home/folarte/include/openh323/h323con.h:555, from /home/folarte/include/openh323/h323.h:492, from h323chan.cpp:32: Browsing cvs for h323caps.c I've found: >>> Revision 1.15 - (view) (download) (as text) (annotate) - [select for diffs] Fri Aug 21 04:35:47 2009 UTC (11 months, 2 weeks ago) by shorne Branch: MAIN Changes since 1.14: +4 -1 lines Diff to previous 1.14 Expressly reference codecs.h to avoid problems with windows SDK codecs.h *** h323caps.h 2009/07/09 15:11:12 1.14 --- h323caps.h 2009/08/21 04:35:47 1.15 *************** *** 27,32 **** --- 27,35 ---- * Contributor(s): ______________________________________. * * $Log: h323caps.h,v $ + * Revision 1.15 2009/08/21 04:35:47 shorne + * Expressly reference codecs.h to avoid problems with windows SDK codecs.h + * * Revision 1.14 2009/07/09 15:11:12 shorne * Simplfied and standardised compiler directives * *************** *** 274,280 **** #endif ! #include "codecs.h" #include "channels.h" #include "mediafmt.h" --- 277,283 ---- #endif ! #include "../include/codecs.h" #include "channels.h" #include "mediafmt.h" <<<<<<<<< This may fix up problems in windows, but leads to a problem when using the installed version. During make phase in h323 headers are $BASE/h323plus/include, but they get installed to $PREFIX/include/openh323, and become incorrect. I don't know how to fix up these in windows, apart of renaming codecs.h to h323_codecs.h, but I thing reverting the fix should make it work again. Making a symbolic link: folarte@gkaps ~/include $ ln -s openh323 include folarte@gkaps ~/include $ ls -l include lrwxrwxrwx 1 folarte folarte 8 Aug 10 19:32 include -> openh323 Makes it work again ( note it includes properly, warns later on): make[1]: Entering directory `/home/folarte/src/yate/modules' g++ -Wall -I.. -I.. -O2 -fno-check-new -fno-exceptions -fPIC -DHAVE_GCC_FORMAT_CHECK -finline -Winline -export-dynamic -shared -Wl,--unresolved-symbols=ignore-in-shared-libs -Wl,--retain-symbols-file,/dev/null -o h323chan.yate -DPHAS_TEMPLATES -D_REENTRANT -DP_HAS_SEMAPHORES -I/home/folarte/include/ptlib -I/home/folarte/include/openh323 h323chan.cpp -L/home/folarte/lib -lopenh323 -L/home/folarte/lib -lpt -L.. -lyate In file included from /home/folarte/include/openh323/../include/codecs.h:345, from /home/folarte/include/openh323/h323caps.h:289, from /home/folarte/include/openh323/h323con.h:555, from /home/folarte/include/openh323/h323.h:492, from h323chan.cpp:32: /home/folarte/include/ptlib/ptlib/object.h: In member function 'virtual PBoolean<unnamed>::YateH323_ExternalRTPChannel::InternalIsDescendant(const char*) const': /home/folarte/include/ptlib/ptlib/object.h:1107: warning: inlining failed in call to 'virtual PBoolean PObject::InternalIsDescendant(const char*) const': call is unlikely and code size would grow /home/folarte/include/openh323/channels.h:248: warning: called from here but it's an horrible hack. Regards. Francisco Olarte.

3 3

Strange library names due to change in PTLIB macros.
by Francisco Olarte (M) 10 Aug '10

10 Aug '10

Due to some configure scripts hacking I've been trying to do I've noticed that both H323 plus and old openh323 generate library names (OH323_BASE, OH323_FILE) using from OBJ_SUFFIX via OH323_SUFFIX: >>>H323plus, cvs 20100525, openh323u.mak.in 248-endif # OH323_LIBDIR 249- 250-ifeq ($(NOTRACE), 1) 251-STDCCFLAGS += -DPASN_NOPRINTON -DPASN_LEANANDMEAN 252:OH323_SUFFIX = n 253-else 254-STDCCFLAGS += -DPTRACING 255-RCFLAGS += -DPTRACING 256:OH323_SUFFIX = $(OBJ_SUFFIX) 257-endif # NOTRACE 258- 259:OH323_BASE = h323_$(PLATFORM_TYPE)_$(OH323_SUFFIX)$(LIB_TYPE) 260-OH323_FILE = lib$(OH323_BASE).$(LIB_SUFFIX) 261- 262-LDFLAGS += -L$(OH323_LIBDIR) 263-LDLIBS := -l$(OH323_BASE) $(LDLIBS) <<< >>>/openh323_v1_19_0_1/openh323u.mak.in 212-endif # OH323_LIBDIR 213- 214-ifeq ($(NOTRACE), 1) 215-STDCCFLAGS += -DPASN_NOPRINTON -DPASN_LEANANDMEAN 216:OH323_SUFFIX = n 217-else 218-STDCCFLAGS += -DPTRACING 219-RCFLAGS += -DPTRACING 220:OH323_SUFFIX = $(OBJ_SUFFIX) 221-endif # NOTRACE 222- 223:OH323_BASE = h323_$(PLATFORM_TYPE)_$(OH323_SUFFIX)$(LIB_TYPE) 224-OH323_FILE = lib$(OH323_BASE).$(LIB_SUFFIX) 225- 226-LDFLAGS += -L$(OH323_LIBDIR) 227-LDLIBS := -l$(OH323_BASE) $(LDLIBS) <<< But the problem is while old pwlibs always had OBJ_SUFFIX as 'd' 'r' ( for debug/release ): >>>pwlib_v1_11_0/make/unix.mak 882- 883:ifndef OBJ_SUFFIX 884-ifdef DEBUG 885:OBJ_SUFFIX := d 886-else 887:OBJ_SUFFIX := r 888-endif # DEBUG 889:endif # OBJ_SUFFIX 890- 891-ifndef OBJDIR_SUFFIX 892:OBJDIR_SUFFIX = $(OBJ_SUFFIX) 893-endif <<< ptlib changed that in revision 20877 http://www.opalvoip.org/websvn/listing.php?repname=Opal&path=%2Fptlib%2Ftru… http://www.opalvoip.org/websvn/comp.php?repname=Opal&path=%2F&compare[]=%2F… from 'd'/'r' to '_d'/'' and ajusted it's paths acordingly, but h233plus did not, so the current paths end up being lrwxrwxrwx 1 gnugk gnugk 31 Aug 10 16:46 libh323_linux_x86_64_.so -> libh323_linux_x86_64_.so.1.22.0 lrwxrwxrwx 1 gnugk gnugk 31 Aug 10 16:46 libh323_linux_x86_64_.so.1 -> libh323_linux_x86_64_.so.1.22.0 lrwxrwxrwx 1 gnugk gnugk 31 Aug 10 16:46 libh323_linux_x86_64_.so.1.22 -> libh323_linux_x86_64_.so.1.22.0 -r--r--r-- 1 gnugk gnugk 10059909 Aug 10 16:46 libh323_linux_x86_64_.so.1.22.0 lrwxrwxrwx 1 gnugk gnugk 24 Aug 10 16:46 libopenh323.so -> libh323_linux_x86_64_.so This is only cosmetic ( it affects some makefiles I have which try to locate the _r names, but that's another story ), but maybe changing the names so the names are libh323_linux_x86_64.so / libh323_linux_x86_64_d.so instead of libh323_linux_x86_64_.so and libh323_linux_x86_64__d.so (double underscore in second ) is worthwile This could be done just by changing 259:OH323_BASE = h323_$(PLATFORM_TYPE)_$(OH323_SUFFIX)$(LIB_TYPE) to 259:OH323_BASE = h323_$(PLATFORM_TYPE)$(OH323_SUFFIX)$(LIB_TYPE) ( underscore zapped between the two macros ) and will possibly need changing 250-ifeq ($(NOTRACE), 1) 251-STDCCFLAGS += -DPASN_NOPRINTON -DPASN_LEANANDMEAN 252:OH323_SUFFIX = n 253-else to 250-ifeq ($(NOTRACE), 1) 251-STDCCFLAGS += -DPASN_NOPRINTON -DPASN_LEANANDMEAN 252:OH323_SUFFIX = _n 253-else (underscore added @ 252 ) to keep the notrace suffix pretty. Regards. Francisco Olarte.

1 0

Memory leak in gkserver.cxx
by Francisco Olarte (M) 05 Aug '10

05 Aug '10

We've been working with and old version of openh233 which had a memory leak in H323GatekeeperServer::OnAdmission ( it was tested, we had call counters and noticed a call was leaked whenever we rejected it ). We are now trying to switch to h323plus, and I've noticed the memory leak still is in place, so I've subscribed to report it. The method in question is H323GatekeeperServer::OnAdmission. I've checked the issue in the last CVS version and it still persists. Below is a copy of the method with the ( working in many millions calls across last months ) fix we are using in our code in the middle. If it's neccessary I can try to make a patch, but I prefer to give this in context. >>>>>>>>>>>>>>>>> H323GatekeeperRequest::Response H323GatekeeperServer::OnAdmission(H323GatekeeperARQ & info) { PTRACE_BLOCK("H323GatekeeperServer::OnAdmission"); OpalGloballyUniqueID id = info.arq.m_callIdentifier.m_guid; if (id == NULL) { PTRACE(2, "RAS\tNo call identifier provided in ARQ!"); info.SetRejectReason(H225_AdmissionRejectReason::e_undefinedReason); return H323GatekeeperRequest::Reject; } H323GatekeeperRequest::Response response; PSafePtr<H323GatekeeperCall> oldCall = FindCall(id, info.arq.m_answerCall); if (oldCall != NULL) response = oldCall->OnAdmission(info); else { // If on restarted in thread, then don't create new call, should already // have had one created on the last pass through. if (!info.IsFastResponseRequired() && info.CanSendRIP()) { PTRACE(2, "RAS\tCall object disappeared after starting slow PDU handler thread!"); info.SetRejectReason(H225_AdmissionRejectReason::e_undefinedReason); return H323GatekeeperRequest::Reject; } ************************************************************ ***** Here newcall is allocated in a plain ptr. ************************************************************ H323GatekeeperCall * newCall = CreateCall(id, info.arq.m_answerCall ? H323GatekeeperCall::AnsweringCall : H323GatekeeperCall::OriginatingCall); PTRACE(3, "RAS\tCall created: " << *newCall); response = newCall->OnAdmission(info); ************************************************************ ******* This tracks the new call inside the endpoint WHEN NOT REJECTED ************************************************************ if (response != H323GatekeeperRequest::Reject) { mutex.Wait(); info.endpoint->AddCall(newCall); oldCall = activeCalls.Append(newCall); if (activeCalls.GetSize() > peakCalls) peakCalls = activeCalls.GetSize(); totalCalls++; PTRACE(2, "RAS\tAdded new call (total=" << activeCalls.GetSize() << ") " << *newCall); mutex.Signal(); AddCall(oldCall); } } ************************************************************ ******* But call is leaked when rejected, we fixed it using: ************************************************************ else { PTRACE(2,"RAS\tDeleting rejected call: "<<*newCall); delete newCall; } ************************************************************ ***** ***** I've traced code path to my best and think newcall cannot ***** be deleted by onAdmission, and our code has not got any ***** double-delete, but we discovered this in final testing ***** and using a H323GatekeeperCall derived call class ( ***** which counts instances, that is how we discovered it ). ***** ****** End of comments/modifications. ************************************************************ #ifdef H323_H248 switch (response) { case H323GatekeeperRequest::Confirm : if (oldCall->AddCallCreditServiceControl(info.acf.m_serviceControl)) info.acf.IncludeOptionalField(H225_AdmissionConfirm::e_serviceControl); break; case H323GatekeeperRequest::Reject : if (oldCall != NULL && oldCall->AddCallCreditServiceControl(info.arj.m_serviceControl)) info.arj.IncludeOptionalField(H225_AdmissionReject::e_serviceControl); break; default : break; } #endif return response; } <<<<<<<<<<<<<<<<<<<<< Regards. Francisco Olarte.

3 6