<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Afonso Zimmermann escreveu:
<blockquote cite="mid:4A9D09B3.8020700@intelbras.com.br" type="cite">
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
Moises Silva escreveu:
<blockquote
cite="mid:c4d05cbe0908311607w5dcd8d93y9383109646023743@mail.gmail.com"
type="cite">
<div>Hello Afonso,</div>
<div> </div>
<div>Thank you for bringing this to my attention. I hope you don't
mind I am moving this thread to the asterisk-r2 mailing list in order
to keep a record of our discussion. Please reply also to the
asterisk-r2 mailing list to keep it there.</div>
<div> </div>
<div>First thing first, thanks for the patches. Not much people
take
the time to contribute real code to openr2. Your help is very much
appreciated.</div>
<div> </div>
<div>Having said that, for the record, I have uploaded the patches
to
my site:</div>
<div> </div>
<div><a moz-do-not-send="true"
href="http://www.moythreads.com/patches/openr2/asterisk-1.4-calling-category.patch">http://www.moythreads.com/patches/openr2/<u><font
color="#0066cc">asterisk-1.4-calling-category.patch</font></u></a></div>
<div> </div>
<div><a moz-do-not-send="true"
href="http://www.moythreads.com/patches/openr2/openr2-calling-category.patch">http://www.moythreads.com/patches/openr2/<u><font
color="#0066cc">openr2-calling-category.patch</font></u></a></div>
<div> </div>
<div>This should help those interested in seeing the patches,
but, I
think, the Asterisk patch as it is now will not work quite well, unless
I don't understand how global variables work in Asterisk. The problem
is that setting a global variable will affect other channels as well,
doesn't it? this means that if 2 incoming channels from PBX A at the
same time report a different category, the call that gets processed in
second place may override the category reported by the first call and
therefore the 2 outgoing calls to PBX B will report the same category,
which is incorrect, this is a race condition, and as more incoming
calls arrive the probability of hitting the race condition is
increased. Let me know if this makes sense.</div>
<div> </div>
<div>In the other hand, I see you are using sort of an old patch,
try
using the latest 1.4 patch, or better yet, if you want latest 1.4 code
try using the mfcr2-1.4 branch (it's under team/moy in Digium's SVN),
you can create any patch you want for 1.4 by using svn diff on that
branch. You will see that the MFCR2_CATEGORY variable is set now in
dahdi_new.</div>
<div> </div>
<div>Another question for you, is, do you need for something in
particular the "mfcr2_call_accepted = 1" lines you added in
dahdi_answer? seems a bit redundant since openr2_chan_accept_call() API
should trigger on_call_accepted callback in Asterisk when the call is
finally accepted (either for fwd or backward call) and then in that
callback I am already setting mfcr2_call_accepted member to 1, which is
better since only when on_call_accepted is triggered we can be sure
both parties are ready to proceed.</div>
<div> </div>
<div>As of the patch for openr2 itself. I will include it with
minor
changes. The only thing I want to change is that you are changing the
behavior of openr2_proto_get_category() in a non-backwards compatible
way. This will break r2test application and any other application
making use of category strings with underscores like
"national_subscriber". I think that in order to keep backwards
compatibility we must accept both "national_subscriber" and "national
subscriber".</div>
<div> </div>
<div>Once I commit the patch, the solution to your problem would be
to do this on the receiving side of the call:</div>
<div> </div>
<div>[openr2-incoming-from-pbx-a] </div>
<div>exten => _X.,1,Answer() ; of course, you dont need to
answer
if you dont want to<br>
exten => _X.,n,Set(_MFCR2_CATEGORY=${MFCR2_CATEGORY})</div>
<div>exten => _X.,n,Dial(DAHDI/g2/${EXTEN}) ; assuming g2 goes
to
pbx b</div>
<div> </div>
<div>The key point to remember here is that you are Set()'ing
_MFCR2_CATEGORY variable to whatever MFCR2_CATEGORY variable is. The
MFCR2_CATEGORY variable has the value of the incoming category, then
you want to inherit that variable to the outgoing channel so Asterisk
can see it in the outgoing channel, that is done in Asterisk through
variable inheritance and you do that by prefixing the variable with an
underscore, so, when you do Set(_MFCR2_CATEGORY=${MFCR2_CATEGORY}), you
are inheriting the value of the incoming channel to the outgoing
channel and therefore the correct category will be used.</div>
<div> </div>
<div>I will commit the changes some day this week and let you know
so
you can please test it before I create a new release (openr2-1.2.1)
with this patch.</div>
<div><br>
-- <br>
Moises Silva<br>
Software Developer<br>
Sangoma Technologies Inc. | 50 McIntosh Drive, Suite 120, Markham ON
L3R 9T3 Canada<br>
t. 1 905 474 1990 x 128 | e. <a moz-do-not-send="true"
href="mailto:moy@sangoma.com">moy@sangoma.com</a><br>
<br>
</div>
<div class="gmail_quote">On Mon, Aug 31, 2009 at 10:30 AM, Afonso
Zimmermann <span dir="ltr"><<a moz-do-not-send="true"
href="mailto:afonso.zimmermann@gmail.com">afonso.zimmermann@gmail.com</a>></span>
wrote:<br>
<blockquote
style="border-left: 1px solid rgb(204, 204, 204); margin: 0px 0px 0px 0.8ex; padding-left: 1ex;"
class="gmail_quote">Hi Moises,<br>
<br>
I'm Afonso Zimmermann, and i work on Cleiber's work now. While a work
in a bridge scenario, I noticed openr2 doesn't passthrough the calling
category passed for other PBX. Scenario is attached.<br>
<br>
When user A calls user B in normal situations, PBX A forward a calling
category "National Subscriber" (II-1 in Brazil), and Asterisk forwad
"National Subscriber" too. But, if PBX A use ani other category (for
example "International Subscriber"), the Asterisk forward "National
Subscriber" too. It doesn't cause any pratical problem to the call, but
is a incorrect indication to PBX B.<br>
<br>
The problem in this case is in pbx_builtin_setvar function, used to set
a Asterisk variable. This function set the variable, but the buffer is
cleared before the openr2 can retrieve this value, retrieveing always
"Null" value. The asterisk patch changes this in asterisk-1.4.24.1,
creating a global variable in asterisk, and the openr2 patch changes
the openr2_proto_get_category function in version 1.1.0, to Asterisk
can understand the calling category value. This also added the "Test
Equipment" category (II-3 in ITU and Brazil too), but this is very,
very, very rarely used. This affect bridge calls, both in
accept_on_offer mode or not.<br>
</blockquote>
</div>
<br>
<pre wrap=""><hr size="4" width="90%">
_______________________________________________
--Bandwidth and Colocation Provided by <a moz-do-not-send="true"
class="moz-txt-link-freetext" href="http://www.api-digital.com">http://www.api-digital.com</a>--
asterisk-r2 mailing list
To UNSUBSCRIBE or update options visit:
<a moz-do-not-send="true" class="moz-txt-link-freetext"
href="http://lists.digium.com/mailman/listinfo/asterisk-r2">http://lists.digium.com/mailman/listinfo/asterisk-r2</a></pre>
</blockquote>
Hi Moises,<br>
<br>
About the Global variables, yes, it is. I don't notice this problem. I
tested with 1 call. But in a real scenario with many calls, it may
happen.<br>
<br>
I used "mfcr2_call_acepted = 1" to remove from asterisk CLI the
following warning lines:<br>
<br>
"Cannot send ANSWER signal, failed to answer call!"<br>
<br>
It appear 3 times before answer.<br>
<br>
I will test this.<br>
<pre class="moz-signature" cols="72">--
Afonso Zimmermann
</pre>
<pre wrap="">
<hr size="4" width="90%">
_______________________________________________
--Bandwidth and Colocation Provided by <a class="moz-txt-link-freetext" href="http://www.api-digital.com">http://www.api-digital.com</a>--
asterisk-r2 mailing list
To UNSUBSCRIBE or update options visit:
<a class="moz-txt-link-freetext" href="http://lists.digium.com/mailman/listinfo/asterisk-r2">http://lists.digium.com/mailman/listinfo/asterisk-r2</a></pre>
</blockquote>
Hi Moises,<br>
<br>
The real messege with appear is:<br>
<br>
"Cannot answer call if the call is not accepted first"<br>
<br>
<pre class="moz-signature" cols="72">--
Afonso Zimmermann
</pre>
</body>
</html>