@whale-av Aha ok, interesting. Do you mean dynamically resizing (as in changing the variable memory allocation) ? If it is the case I don't know how to.

@weightless , thanks for your input. True I am still a bit confused about data structure and access in PD. Like the link between tables and arrays.

Unfortunately the experience I was talking about is for the moment such a mess that I don't want to risk anyone's sanity by posting it. I will try to take some time to extract a minimal example.

In one case I have a [tabread] flow fed into a chain of logical gates (a kind of classifier, data is coming from a live spectral analysis of the mic signal ). Depending on the value, the chain can be short or long before completion. I had problem's with data messing in the pipe, mostly because the order is not conserved (e.g. first value needing a bigger chain ends up getting out later than second value with a shorter chain) . I have resolved the problem by adding a [metro 5] as a feeding rate, however it is quite inefficient because 5ms is a very conservative time estimate for any single chain. Ideally I would like to build a data bucket after the [tabread], which will release values when a completion bang is received, the tricky part is that the bucket has to remain open for receiving new values.

I hope that this all makes some sense.
Cheers.

(improved clarity edits)

Hey ! I am just starting with PD. Very nice piece of code I must say.
My question: Is there an easy way to schedule a queue of processes in pd-vanilla ?
For instance in one of my experiments I trigger a series of bangs launching several subpatche process of yet unknown duration (based on tabread loops) . A given process should wait for it's predecessor completion before starting. Simply speaking I am looking for a kind of "return 1" equivalent.

Thanks in advance for your advices.