Re: [h323plus] [Opalvoip-devel] [Openh323-devel] opalvoip vs h323plus
..deleted
H323plus has a completely different focus to that of OpalVoIP and is designed to undertake advanced H.323 development (hence the plus) and unleach the full potential of what H.323 is capable of where are OpalVoIP is more of a generalist which to date has focused a lot of their attention (of late) on SIP.
Here is a specific question for those people using OpenH323 (and now H323Plus) instead of OPAL.
What would have to changed in, or added to, OPAL in order for you to move from OpenH323 (other than all of H.45x and H.46x stuff we are already committed to migrating) ?
In other words, if you configure Opal with only H.323 support (which you can do now by running "./configure --disable-sip --disable-iax", what makes it so different from OpenH323 that it is worth splitting our joint efforts and maintaining two codebases?
If you can answer this question, I promise that the Opal project will do it's very best to fulfill the requirements so we can all work on one codebase rather than two.
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
Welcome Craig
h323plus currently contains a large array of features such as directory services (H.350),app sharing (H.239) conference controls (H.230), SNMP (H.341) etc which you did not mention that are not currently available in Opal. As I offered previously to you or anyone else, you are quite welcome to port any of these ideas to opal (like my H.460.x work craig mentioned) with my support so as to provide a more compelling argument for migration to opal.
But honestly I think you should direct your questions to the various projects that were "orphaned" on the executive decision to migrate to OpalVoIP and plead the case to them.
You are most welcome to participate on the h323plus mailing list and provide input etc but please try and refrain from touting other projects, it was frowned apon on the openH323 list, so it should be no different now.
Simon
-----Original Message----- From: h323plus-bounces@lists.packetizer.com [mailto:h323plus-bounces@lists.packetizer.com]On Behalf Of Craig Southeren Sent: Saturday, 3 November 2007 1:18 PM To: Robert Jongbloed Cc: h323plus@lists.packetizer.com; Opalvoip-devel@lists.sourceforge.net; 'Discussion on enhancements and development issues with the OpenH323library'; GNU Gatekeeper Developer Subject: Re: [h323plus] [Opalvoip-devel] [Openh323-devel] opalvoip vsh323plus
..deleted
H323plus has a completely different focus to that of OpalVoIP and is designed to undertake advanced H.323 development (hence the plus) and unleach the full potential of what H.323 is capable of where are OpalVoIP is more of a generalist which to date has focused a lot of their attention (of late) on SIP.
Here is a specific question for those people using OpenH323 (and now H323Plus) instead of OPAL.
What would have to changed in, or added to, OPAL in order for you to move from OpenH323 (other than all of H.45x and H.46x stuff we are already committed to migrating) ?
In other words, if you configure Opal with only H.323 support (which you can do now by running "./configure --disable-sip --disable-iax", what makes it so different from OpenH323 that it is worth splitting our joint efforts and maintaining two codebases?
If you can answer this question, I promise that the Opal project will do it's very best to fulfill the requirements so we can all work on one codebase rather than two.
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
Simon Horne wrote:
Welcome Craig
h323plus currently contains a large array of features such as directory services (H.350),app sharing (H.239) conference controls (H.230), SNMP (H.341) etc which you did not mention that are not currently available in Opal. As I offered previously to you or anyone else, you are quite welcome to port any of these ideas to opal (like my H.460.x work craig mentioned) with my support so as to provide a more compelling argument for migration to opal.
I'd like to see all of these ported over the OPAL, and I'm totally prepared to put in the work to do it, but I can only do it with *your* support
Given that you know the code best, what specific feature (i.e. what source files) would you recommend I start with?
But honestly I think you should direct your questions to the various projects that were "orphaned" on the executive decision to migrate to OpalVoIP and plead the case to them.
I don't see that any projects have been orphaned by the recent creation of OpalVoip, or at least, they are no more orphaned than they were before.
As you have correctly pointed out elsewhere, Robert and I effectively stopped working on OpenH323 quite some time ago. Our decision to move Opal onto a new site does not change that, so any projects that that are using OpenH323 will continue to use it, or can use h323plus.
I'm also a little confused by the reference to "executive decision". Robert and I have received many requests by Opal users over the past few years to move the code to it's own site. We have put it off for all this time but we finally decided to do it. If anything, we probably should have done it much earlier :)
Our decision to create OpalVoip was no more "executive" than your decision to create h323plus, and was probably done for much the same reasons :)
You are most welcome to participate on the h323plus mailing list and provide input etc but please try and refrain from touting other projects, it was frowned apon on the openH323 list, so it should be no different now.
Everything will be in context, and cross-posts may happen :)
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
I'd say from GnuGk point of view it would be more important to have long-lived PWLib bugs fixed. H.45x and H.46x stuff is useful mostly for development of custom software only. It's nice to have it, but it's not mission critical.
----- Original Message ----- From: "Craig Southeren" craigs@postincrement.com To: "Robert Jongbloed" robert.jongbloed@bigpond.com Cc: h323plus@lists.packetizer.com; Opalvoip-devel@lists.sourceforge.net; "'Discussion on enhancements and development issues with the OpenH323 library'" openh323-devel@lists.sourceforge.net; "GNU Gatekeeper Developer" openh323gk-developer@lists.sourceforge.net Sent: Saturday, November 03, 2007 6:17 AM Subject: Re: [Openh323gk-developer] [Opalvoip-devel] [Openh323-devel]opalvoip vs h323plus
..deleted
H323plus has a completely different focus to that of OpalVoIP and is designed to undertake advanced H.323 development (hence the plus) and unleach the full potential of what H.323 is capable of where are OpalVoIP is more of a generalist which to date has focused a lot of their attention (of late) on SIP.
Here is a specific question for those people using OpenH323 (and now H323Plus) instead of OPAL.
What would have to changed in, or added to, OPAL in order for you to move from OpenH323 (other than all of H.45x and H.46x stuff we are already committed to migrating) ?
In other words, if you configure Opal with only H.323 support (which you can do now by running "./configure --disable-sip --disable-iax", what makes it so different from OpenH323 that it is worth splitting our joint efforts and maintaining two codebases?
If you can answer this question, I promise that the Opal project will do it's very best to fulfill the requirements so we can all work on one codebase rather than two.
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
Zygmuntowicz Michal wrote:
I'd say from GnuGk point of view it would be more important to have long-lived PWLib bugs fixed. H.45x and H.46x stuff is useful mostly for development of custom software only. It's nice to have it, but it's not mission critical.
Are there any bugs in particular you need fixed? Or are you just talking about general ongoing maintenance?
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
There are still at least 2 open bugs on the PWLib tracker that I submitted - PConfig problem with file contents caching and (recently committed) PIPSocket::Address::AsString threading issues. There are also some BSD issues with sockaddr structures - some systems require an additional field ss_len and/or sa_len to be initialized for sockaddr_storage and/or sockaddr_in.
----- Original Message ----- From: "Craig Southeren" craigs@postincrement.com To: "Zygmuntowicz Michal" m.zygmuntowicz@onet.pl Cc: "GNU Gatekeeper Developer" openh323gk-developer@lists.sourceforge.net; "Robert Jongbloed" robert.jongbloed@bigpond.com; h323plus@lists.packetizer.com; Opalvoip-devel@lists.sourceforge.net; "'Discussion on enhancements and development issues with the OpenH323library'" openh323-devel@lists.sourceforge.net Sent: Saturday, November 03, 2007 12:36 PM Subject: Re: [h323plus] [Openh323gk-developer] [Opalvoip-devel] [Openh323-devel]opalvoip vs h323plus
Zygmuntowicz Michal wrote:
I'd say from GnuGk point of view it would be more important to have long-lived PWLib bugs fixed. H.45x and H.46x stuff is useful mostly for development of custom software only. It's nice to have it, but it's not mission critical.
Are there any bugs in particular you need fixed? Or are you just talking about general ongoing maintenance?
Craig
Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
Zygmuntowicz Michal wrote:
There are still at least 2 open bugs on the PWLib tracker that I submitted
- PConfig problem with file contents caching and (recently committed)
PIPSocket::Address::AsString threading issues. There are also some BSD issues with sockaddr structures - some systems require an additional field ss_len and/or sa_len to be initialized for sockaddr_storage and/or sockaddr_in.
I've just checked, and I am not getting email alerts from the OpenH323 bug tracker even though I get them just fine from the patch tracker.
So, please accept my apologies for the delay in getting these issues addressed. I will look at them immediately
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
Hi Craig,
I read your comment for the PIPSocket::Address::AsString fix and you missed one thing. Even if IPv6 is enabled, getnameinfo is used only for IPv6 addresses (and not also for IPv4, as it should be) due to if (version == 6) check.
Another issue is that getnameinfo fails on BSD (and maybe some other) systems with EAI_FAIL if sa_len field of sockaddr is not initialized with a proper structure length. I saw that other projects (like PostgreSQL) check for the sockaddr.sa_len field presence in the configure script and initialize it with a proper value. We could do the same inside Psockaddr class.
----- Original Message ----- From: "Craig Southeren" craigs@postincrement.com To: "GNU Gatekeeper Developer" openh323gk-developer@lists.sourceforge.net Cc: h323plus@lists.packetizer.com; Opalvoip-devel@lists.sourceforge.net; "'Discussion on enhancements and development issues with the OpenH323library'" openh323-devel@lists.sourceforge.net; "Robert Jongbloed" robert.jongbloed@bigpond.com Sent: Sunday, November 04, 2007 11:28 AM Subject: [Openh323gk-developer] Outstanding Ptlib problems (was opalvoip vsh323plus)
Zygmuntowicz Michal wrote:
There are still at least 2 open bugs on the PWLib tracker that I submitted
- PConfig problem with file contents caching and (recently committed)
PIPSocket::Address::AsString threading issues. There are also some BSD issues with sockaddr structures - some systems require an additional field ss_len and/or sa_len to be initialized for sockaddr_storage and/or sockaddr_in.
I've just checked, and I am not getting email alerts from the OpenH323 bug tracker even though I get them just fine from the patch tracker.
So, please accept my apologies for the delay in getting these issues addressed. I will look at them immediately
Craig
----------------------------------------------------------------------- Craig Southeren Post Increment – VoIP Consulting and Software craigs@postincrement.com.au www.postincrement.com.au
Phone: +61 243654666 ICQ: #86852844 Fax: +61 243656905 MSN: craig_southeren@hotmail.com Mobile: +61 417231046 Jabber: craigs@jabber.org
"Science is the poetry of reality." Richard Dawkins
Craig Southeren asked:
Here is a specific question for those people using OpenH323 (and now H323Plus) instead of OPAL.
What would have to changed in, or added to, OPAL in order for you to move from OpenH323 (other than all of H.45x and H.46x stuff we are already committed to migrating) ?
The main point for GnuGk is that there is very little to be gained from a move to OPAL. Its not that I dislike anything thats in there, we just don't need 90% of the framework OPAL provides.
But because we would have to change some many places in the code where the interface is a little different, we would introduce a lot of bugs when doing the switch - not even considering the amount of work we'd have to put in. This simply doesn't make sense for a zero gain.
It totally makes sense to use OPAL for any endpoint application. They have a lot to benefit, but GnuGk is using mainly the packet decoding stuff and that has been extremely stable since a number of years. Every once in a blue moon we hit a PWLib bug or Simon implements a new H.great-and-new standard, but basically the OpenH323 code base is just too stable and proven to drop for GnuGk's needs. ;-)
Cheers, Jan
-----Original Message----- From: openh323-devel-bounces@lists.sourceforge.net [mailto:openh323- devel-bounces@lists.sourceforge.net] On Behalf Of Jan Willamowius
....
The main point for GnuGk is that there is very little to be gained from a move to OPAL. Its not that I dislike anything thats in there, we just don't need 90% of the framework OPAL provides.
....
And that is fine, you can stay with OpenH323.
There are two classes of legacy user: ones such as GnuGk for which OpenH323 v1.18 does exactly what they want and they need no more features. Except maybe the odd bug fix, the code is as it is.
The second class are people with ongoing application development are in continuous need of the "next new thing" in the feature list. These are the people we were hoping to make OPAL sufficiently attractive, due to its new whiz bang wonder gizmos, to convert to OPAL from OpenH323.
Oh, and suppose there is a third class of those who have not yet started their project. They should have no reason not to use OPAL. We don't want them to have a reason. And a major reason for moving OPAL to a its own project was not to "fork" the code, but to reduce the confusion between OpenH323 and OPAL.
I will reiterate what I and Craig have been saying all along: we are NOT abandoning OpenH323 completely. We just have not been doing any NEW work there as it is a duplication of effort. It is more akin to Ethereal and Wireshark, Ethereal is what it is at that point of time and all new stuff is in Wireshark. It is not, and never was, intended that this be a true "fork" where A goes it's way and B goes theirs. This is the whole reason why I am so disappointed with the ensuing result.
Remember, this whole thread started with someone asking the question "Should I use OpenH323 or OPAL, which is better?", a fair question but one that I did not want anyone to have to ask.
Robert Jongbloed OPAL/OpenH323 Architect and Co-founder.
participants (5)
-
Craig Southeren -
Jan Willamowius -
Robert Jongbloed -
Simon Horne -
Zygmuntowicz Michal