please forgive my stubborness......... can you send me an example? i'm just wondering, because i once did a patch with dynamically created abs, in which other abs were created dynamically .. and so on, and it worked all fine (with several messages, too). also doing something like:
[112(
|
[t b b f]
| |
| [repeat 0]
| |
| [; pd-patch.pd obj 10 10 myabs(
|
[; pd-myabs.pd loadbang(
works fine on my machine, altough it takes a few miliseconds to create all [myabs], but all [myabs] get a loadbang. i never experienced yet that pd does NOT keep execution order because of the time it takes to build the gui or dsp-chain or whatever. but obviously i did not try everything.
cheers roman
----- Original Message ----- From: "B. Bogart" ben@ekran.org
*cough*
I unfortunatly found many issues with things not happening in the correct time. Well that is to say when you dynamically create an array of objects the time between the message being send and the object actually being ready to receive messages is not 0ms. I can't remember the delay I needed, but indeed some abstractions would not get the
bang
message if I sent it immediatly after the message to create the last abstraction. This is on OSX, which has had some GUI issues in the
past,
could be related.
Oh and this is also because I'm not creating my whole dynamic patch
with
one message, but multiple messages. (This is because I'm creating one abstraction for each file in a directory)Also the number of files is indeterminant so I can't create one message to do it all...
I guess the best way to test if the last object is done would be to
have
it send a message "I'm done!" to the parent patch and then we know all the abstractions are ready... Maybe I'll do this...
Back to replacing my TOTs with PYs.
b>
Roman Haefeli wrote:
hi ben
----- Original Message ----- From: "B. Bogart" ben@ekran.org
Hey Tom,
If you look at pt.layerfx in pixelTANGO I'm sending a bang to all dynamic abstractions after all the creation messages have been sent
and
an additional (arbitrary) delay. It seems to work...
are you sure, that a delay is necessary? in my experience, the determinisme of pd is absolutely reliable. i mean, when i use a [t b
b]
and i use the first 'bang' to create the whole patch and the second
one
the send the 'loadbang' it works perfectly. afaik, pd rather drops the audio-processing for while than dropping a message. at least i experienced that if i create big patches dynamically, i get audio drop-outs.
It would be damn handy to have tcl/tk report when a dynamic patch has finished building.
in case i am right, this wouldn't be necessary anymore... or would
it?
cheers roman