Hi again.<br><br>I now seem to have another problem. Running the testOnDemandRTSPserver with .ts or .mpg files makes it use all cpu power. A quick profiling suggests that there is something in the BasicTaskScheduler that is doing this:  
<br><br>Flat profile:<br><br>Each sample counts as 0.01 seconds.<br>  %   cumulative   self              self     total<br> time   seconds   seconds    calls  us/call  us/call  name<br> 19.79      0.57     0.57                             BasicTaskScheduler::SingleStep(unsigned int)  
<br> 10.42      0.87     0.30                             HandlerSet::~HandlerSet()<br>  9.72      1.15     0.28                             BasicTaskScheduler::turnOffBackgroundReadHandling(int)<br>  6.94      1.35     0.20
                               DelayQueue::synchronize()<br>  6.94      1.55     0.20                             DelayQueue::removeEntry(long)<br>  6.25      1.73     0.18                             DelayQueue::handleAlarm()  
<br>  5.90      1.90     0.17                             DelayQueue::~DelayQueue()<br>  4.51      2.03     0.13                             Timeval::operator>=(Timeval const&) const<br>  4.17      2.15     0.12                             TimeNow()  
<br>  3.13      2.24     0.09                             operator*(short, DelayInterval const&)<br>  2.78      2.32     0.08                             DelayQueue::~DelayQueue()<br>  2.43      2.39     0.07                             TransportStreamAnalyze::maskTables(unsigned int)  
<br>  1.39      2.43     0.04                             BasicTaskScheduler::createNew()<br>  1.39      2.47     0.04                             BasicTaskScheduler0::scheduleDelayedTask(long, void (*)(void*), void*)<br>
    1.39      2.51     0.04                             TransportStreamAnalyzeTS::analyze(unsigned char const*)<br>  1.39      2.55     0.04                             TransportStreamAnalyzePID::parsePESHeader(unsigned char const*, unsigned int&)  
<br>  1.39      2.59     0.04                             __gnu_cxx::slist<SubscribedPIDIndex, std::allocator<SubscribedPIDIndex> >::end()<br>  1.04      2.62     0.03                             DelayQueue::findEntryByToken(long)  
<br>  0.69      2.64     0.02                             BasicTaskScheduler::BasicTaskScheduler()<br><br>This does of course effect all applications that use the Live555 TaskScheduler. Can anyone confirm this? I find it strange that noone has noticed this before.
<br><br>-Morgan-<br><br><br>On 30/10/06, Morgan Tørvolt < <a href="mailto:morgan.torvolt@gmail.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)"> morgan.torvolt@gmail.com</a>> wrote:<br>> Darn. I had the latest version, but I never checked the timestamps on 
<br>> the files. :-/<br>> <br>> Changing a symbolic link fixed the problem *flush*. That was an easy fix. <br>> <br>> Thanks...<br>> <br>> -Morgan Tørvolt-<br>> <br>> On 30/10/06, Ross Finlayson < 
<a href="mailto:finlayson@live555.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">finlayson@live555.com</a>> wrote:<br>> > >In OnDemandServerMediaSubsession.cpp  you have a while (1) loop, 
<br>> > >grabbing a port with an even number<br>> ><br>> > Not anymore!  That code changed starting with version 2006.08.07 -<br>> > i.e., more than 2 months ago.  You should upgrade. <br>> > 
<br>> > Everybody - please: Upgrade to the latest version of the code before<br>> > spending (wasting) a lot of time digging into alleged problems.<br>> ><br>> > >  (why you need an even number, cannot guess).  
<br>> ><br>> > Because the RTP standard specifies that even port numbers should be<br>> > used for RTP, with the next highest (and thus, odd) port number being<br>> > used for RTCP.  (Actually, this rule is not quite as manditory as it  
<br>> > used to be, but it remains the default and usual behavior.)<br>> ><br>> > >What needs to be done to fix this is addressing the port number issue.<br>> ><br>> > Right now I'm not convinced that there is anything that needs to be 'fixed'.  
<br>> > --<br>> ><br>> > Ross Finlayson<br>> > Live Networks, Inc.<br>> > <a href="http://www.live555.com/" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://www.live555.com/ 
</a><br>> > _______________________________________________ <br>> > live-devel mailing list<br>> > <a href="mailto:live-devel@lists.live555.com" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">
 live-devel@lists.live555.com</a><br>> > <a href="http://lists.live555.com/mailman/listinfo/live-devel" target="_blank" onclick="return top.js.OpenExtLink(window,event,this)">http://lists.live555.com/mailman/listinfo/live-devel  
</a><br>> ><br>> <br>