Re: [Jack-Devel] A couple of questions about Jackdmp

Le 18 avr. 2011 à 11:32, John Emmas a écrit :

> 
> On 18 Apr 2011, at 09:38, Stéphane Letz wrote:
> 
>> 
>>> In recent mailings I've noticed some references to the VC++10 runtime and redistributables.
>> 
>> Only to compile JackRouter ASIO driver, not for jack2 itself.
>> 
> Ah - that's the bit I hadn't realised.  Looking through my code though, although I modified the jackdmp source quite extensively, the vast majority of my changes seem to be for accommodating Cygwin rather than VC++.  For example wherever the code says this:-
> 
> #ifdef WIN32
> 
> I've needed to change it to this:-
> 
> #if (defined(WIN32) && !defined(__CYGWIN__))
> 
> 
> With regard to Visual C++ the only major change I needed to implement was to use the threading library 'pthread_win32' as opposed to regular pthread.  

Hum... jack2 on WIndows does not use "pthread": it directly uses Windows threads (look at JackWinThread.cpp file)


> The other changes seem to be very simple things (redefining Int64 to be 'long long' for example).
> 
> 
>>> 
>>> Q2.  If I understand Jackdmp's command line parameters, the parameter "-T" makes Jackdmp "temporary".  I think this means that if a client launches Jackdmp using the "-T" parameter, Jackdmp should unload itself when there are no more clients connected.  Have I understood that correctly?
>> 
>> Yes.
>> 
>>> It's the only thing that doesn't seem to work in my Windows build so I'd like to figure out where I should start looking for the problem.
>>> 
> I'm glad I looked into this.  While browsing through the source I came across this function in JackEngine.cpp:-
> 
> void JackEngine::ReleaseRefnum(int ref)
> {
>    fClientTable[ref] = NULL;
> 
>    if (fEngineControl->fTemporary) {
>        int I;
> 
>        for (I = REAL_REFNUM; I < CLIENT_NUM; I++) {
>            if (fClientTable[I])
>                break;
>        }
> 
>        if (I == CLIENT_NUM) {
>            // last client and temporay case: quit the server
>            jack_log("JackEngine::ReleaseRefnum server quit");
>            fEngineControl->fTemporary = false;
> 
> #ifndef WIN32
>            exit(0);
> #endif
>        }
>    }
> }
> 
> I assume that if I removed  #ifndef WIN32  in the above code, my server would magically unload itself when there are no more clients left - but I'm curious to know why the condition was ever there.  Is it still present in the latest code and/or can anyone remember why it's needed?
> 
> John

The whole "temporary" machinery has been redesigned since and works also on Windows.

Stéphane 

1303119606.373_0.ltw:2,a <F2FD6625-FD67-461E-92D3-2D9EADBB285B at grame dot fr>