hello!
i have a to me weird appaering problem...
send~/receive~ connection stops working. after i rename it, it works again for a while then likes to stop again. in the patch there are quite some connections of this type.
i use pd 39.2 ext.test7 on intel mac mini.
yann

Hi!
I saw in send~-help.pd this mechanism works only for default block size
(64 samples). Is this your case?

Alberto

The behaviour you describe sounds typical of when you have more than one [send~]
or [receive~] of the same name. They must be unique pairs. Use [throw~] and [catch~] for n:n connectivity.

If you get odd problems with sends and receives delete them *both* and replace the send before the receive. The order in which you create them has an effect on how they behave. Usually you never notice this, but sometimes it causes subtle bugs. The best way imho is to delete them all and rebuild the patch if you get this happening.

It's not a bug, it's relatively undocumented hole into which the unlucky may fall. I have fallen in the same hole many times.:(

regards,
Andy

you can have only one [send~] object per name, but multiple [receive~] objects with the same name are ok.

conversely, you can have multiple [throw~] objects per name, but only one [catch~] object per name.

so:

[s~ synth]

[r~ synth] [r~ synth] [r~ synth]

is ok, as is:

[throw~ vox] [throw~ vox] [throw~ vox][throw~ vox]

[catch~ vox]

hi!
i did delete them both made new ones with a different name but did not care about in what order yet. i will try that next.
so it makes a difference if i delete objects or rename them?
i always check my patch in the pd window and look for last error.
there it then reports just the send~ xzy object which do not work.
it does not tell me why...
another thing is that sometimes after saving changes, sometimes, on the next program start it is partly still the version before the changes.
i chowned the whole folder but it did not change that behaviour.
i use the same folder from os x and linux. could that create the confusion?
yann

@geraeuschtest said:

hi!
i always check my patch in the pd window and look for last error.
there it then reports just the send~ xzy object which do not work.
it does not tell me why...

It won't spot them because they are not errors. Sometimes breaking when you
change evaluation orders is a feature of Puredata. If that seems cruel say a prayer for
those poor souls who use Max. Position of objects not just connectivity affects the evaluation order in a Max/MSP patch, so just moving an object can break your patch.

Obviously it's not an ideal situation either way. The Max way actually gives you an easier opportunity to experiment with correcting an order problem. Improvements in Pd that might sort this out are ways of tracking the order with numbered connections or dumping a trace of the netlist graph traversal. But don't expect those soon.
It's one of those things you just have to live with at the moment.

@geraeuschtest said:

another thing is that sometimes after saving changes, sometimes, on the next program start it is partly still the version before the changes.
i chowned the whole folder but it did not change that behaviour.
i use the same folder from os x and linux. could that create the confusion?
yann

That's most odd. I can't explain it. I doubt your getting partial file writes. It's probably something else.

if i copy and paste the whole patch and subpatches, would then the code be rewritten without the chaos?
y

if it's not too hard to paste the patch here, then somone can probably check it out for you and see what is going funny.

maybe i do this in the future...
for now i like to get some answers to my questions to find the funny stuff myself.
in pd i consider myself still as some sort of newbee.
greets
y