October 2010 - h323plus - lists.packetizer.com

crash when SessionID=0
by Marek Domaracky 13 Jun '11

13 Jun '11

Hi, I'm working on H.239 implementation and find out that devices might ask SessionID=0 for a new video channel. In particular I did a test with Mirial. I was checking the thread http://lists.packetizer.com/pipermail/h323plus/2009-December/001269.html and it's mentioned that the SessionID=0 is invalid. I know it's not following the standard as for RTPSession the range is from 1..255, but for compatibility, don't you think we should allow it? I did a test call from Mirial to Tandberg and Tandberg didn't have a problem with it. The lines involved are rtp.cxx from PAssert(id > 0 && id < 256, PInvalidParameter); to PAssert(id >= 0 && id < 256, PInvalidParameter); and h323.cxx // We must have a valid sessionID if (sessionID < 1 || sessionID > 255) to if (sessionID < 0 || sessionID > 255) Thanks, Marek

8 18

enabling DirectShow video input
by Alessandro Angeli 02 Nov '10

02 Nov '10

I successfully built the debug configurations of ptlib 2.4.5 (static), h323plus 1.21.0 and simpleplus 1.21.0 in VS2008SP1 + WinSDK 7.1 on Win7/x64. Audio and text work fine. However, video does not work: "No Video Grabber available Disabling Video Support!". DirectShow support in vidinput_directx.cxx is compiled successfully and the DirectShow branches in simpleplus/main.cxx are active, since P_DIRECTSHOW is #define'd as 1 in ptbuildopts.h. I tried #define'ing P_DSHOW to 1 as well in simpleplus, without any difference. As far as I can tell, the problem is that the DirectShow service is never registered on startup. I expanded the PCREATE_VIDINPUT_PLUGIN(DirectShow) macro in vidinput_directx.cxx to put a breakpoint in PPlugin_PVideoInputDevice_DirectShow_Registration::<ctor>() and that code is compiled but never executed. Actually, it appears to never be mapped into memory at all, which should imply that the linker is removing it from the final EXE, but adding /OPT:NOREF in the build config of simpleplus does not make a difference. However FakeVideo is registered and it also uses the PCREATE_VIDINPUT_PLUGIN() macro, and I can not spot any difference between the way FakeVideo and DirectShow are compiled and linked. The only reference to a similar issue I could find is from over one year ago: http://lists.packetizer.com/pipermail/h323plus/2009-August/001165.html But the current code already uses PCREATE_VIDINPUT_PLUGIN() and, as I said, FakeVideo uses the same macro but is correctly registered. -- Alessandro

4 9

calling the OpenVideoChannel function
by Isaac Kazadi 31 Oct '10

31 Oct '10

Hi all, I know there have been a few questions regarding this matter and i've looked throught the mailing list and didn't quite find the answer for it unless i missed it. I've downloaded and compiled ptlib 2.8.1 and the latest h323plus cvs and simpleh323, all seems to work fine except i've noticed that the every overriden functions from h323Endpoint are being called but the OpenVideoChannel/ OpenExtendedVideoChannel functions aren't getting called at all. I've checked and can confirm that H323_VIDEO is already defined, and i've tried tracing where the functions are being referenced, but didn't notice anything wrong. You help/suggestions will be greatly appreciated. Thanks Zhack

1 0

Bridging two H.323 calls together
by Mark Fawcett 27 Oct '10

27 Oct '10

All, I have successfully installed and used H.323plus and am very impressed by the strength-in-depth. I am building a B2BUA application and can make two outgoing calls and now need to switch the audio between them. I have done this initially using a new channel class derived from PIndirectChannel and taking audio data on the Write from one channel, buffering it internally in the application and then sending it to the other channel on its Read. I am wondering if there is a better or alternative method to achieve the same results - i.e. is it possible to use an H323plus object/method to connect the two H323Connections back-to-back without the need to go through the intermediate buffering stage in the application? Many thanks Mark

2 1

GNU Gatekeeper release 2.3.3
by Jan Willamowius 27 Oct '10

27 Oct '10

Hi all, the new GNU Gatekeeper release 2.3.3 is now available: http://www.gnugk.org/gnugk-2-3-3.html I would like to thank all sponsors who made it possible to add so many new features again. I'll append a list of changes below. Regards, Jan -- Jan Willamowius, Founder of the GNU Gatekeeper Project EMail : jan(a)willamowius.de Website: http://www.gnugk.org Support: http://www.willamowius.de/gnugk-support.html New features (gatekeeper configuration): - all executables now contain all database drivers: MySQL, PostgreSQL, ODBC, SQLite and Firebird! - much improved bandwidth management for registered endpoints per endpoint and per call - configure alternate gatekeepers per endpoint IP range with [RasSrv::AlternateGatekeeper] - call routing by destination IP with [Routing::Explicit] - gatekeeper-based TCS=0 call transfer: activated through H.450.2 emulator or status port (experimental) - H.450.2 emulator extended for unregistered calls - new switch: [CTI::MakeCall]TransferMethod= - BMV, a web front-end for SQLBill (in contrib/bmv2/) - configure where to send RAS replies with [ReplyToRasAddress] - ignore a 2nd column in SQL routing if it equals "IGNORE" - display H323Plus and PTLib version in GnuGk's version string - documented [RoutedMode] EnableH460P= and H460PActThread= switches New features (API): - extended status port commands RouteToAlias, RouteToGateway, BindAndRouteToGateway with caller-ID parameter - updated status port command: TransferCall by call-id and with transfer method - new status port command RerouteCall (experimental) - new status port command: DisconnectCallId - added vendor string in RouteRequest event - document status port command: DisconnectEndpoint Config changes: change to the format of [Gatekeeper::Main] AlternateGKs= Bug fixes: - apply ENUM policy also to Setup and Facility messages - fix multicast discovery - fix crash on failover with DisableRetryChecks=1 - avoid bug in H323Plus so H.460.19 feature indicator can be removed - disregard IPv6 addresses if the machine has any - DisconnectIP now disconnects all calls on the IP as specified - fix crash when setting a non-writable trace file in the config file - signal change of destination alias in parent policy - don't accept new calls or retry calls when shutdown is in progress, avoids possible crash - check if routing policies have been configured, before using them - only rewrite sourceCallSignalAddress if proxying - fix NAT detection for unregistered callers - fire unregister event also when endpoint expired - fix order of home IPs

1 0

Problem compiling h323plus 1.21.0
by alvaro muniz 22 Oct '10

22 Oct '10

Hi, I just download ptlib 2.4.5, h323plus 1.21.0 and openmcu 2.2.5. When i try to compile h323plus using make install, it fail, because /bin/install try to find libh323_linux_x86_.so.1.21.0, but in the lib folders it find libh323_linux_x86_.so.1.21.beta0. How i solve this issue? I download the right version? http://www.h323plus.org/source/ Im working on ubuntu 10.10 Thank you so much Alvaro

1 0