switching between each other and creating a crossfade

Pd Vanilla's example patches (I don't know if PlugData has them):
Read B.07.sampler.pd until B.13.sampler.overlap.pd

dumb questions

I didn't say that.

I thought the guard points was there to make a fast crossfade between the end and the start of the loop... is this right?

The "4" in [tabread4~] [tabosc4~] ect. stands for four-point interpolation. It inter-sample interpolates. This is useful when you read an array at different sampling-rate than the patch/system's sampling-rate. This happens if you change the read-speed or of the array or if it has different samplerate. By interpolating between samples you get a value anytime, even between samples.
The interpolation algorithm requires 4 points, and to work as intended, it requires an array-size in the power of 2 plus 3 samples have to be added at the edges of the array.
You can read about this in more detail in the array-guard helpfile and further links mentioned there.
There is also [tabread~] that does not inter-sample interpolate.

I thought you could add some extra samples to the length of the table and it would happen automatically

Take an array in the size of power of two (256 for example) and use array-guard to add the guardpoints.

But I don't think that the click is about inter-sample-interpolation-artifacts.
Check out the patch I uploaded in my last post.

crossfade between the end and the start of the loop

Yes, this would be another common way to get rid of the click.

Also, you can read Perlin directly, without array, check Perlin helpfile.

here is a patch to play with, as always in DSP there are tradeoffs to choose.

array-clicks.pd

Also you can drive Perlin noise directly without glitches, without an additional array.

@mezko
Pd is a fairly low-level language, more so than it is common in the music software world. Another way to say this is, Pd is closer to C-Sound than to any VST. I'd say it is even closer to C-sound than to a modular-software, such as VCV-Rack or Reaktor.
You have to understand things more in depth than "I want this, I tryed that, it klicks, please make it work." If you want things easier, use sth different than Pd.
Also read this post by @Boran-Robert https://forum.pdpatchrepo.info/topic/13054/starting-a-pure-data-wiki-database-examples-collection/14

When you are asking questions in the forum, please make it as easy as possible for us to answer and make some effort yourself!
Did you already search? Did you read the helpfile of tabosc4~ ? Did you read the helpfile of array-guard ? (here is the original post: https://forum.pdpatchrepo.info/topic/14301/add-delete-guard-points-of-an-array-for-4-point-interpolation-of-tabread4-ect )

What is it that you do not understand?
This forum is very helpful, but "I don't understand" is not not enough to give any answer.
If you want low-level, I recommend learn the basics by going through Vanilla's example patches and manual.

Don't want to sound like a gatekeeper or so but this is basic netiquette.

On your question about clicking:
a) What is a click? We hear a click when there is an abrupt change in the waveform.
b ) When does it happen? At the wrap around of the reading [phasor~]>[tabread4~]
c) There is a jump between the last sample and the first sample, so it clicks
d) How to solve this?
e) One way is to use a triangle-wave reading the array instead of a saw-tooth ramp. This gives us a continous loop back and forth without glitch.

I think @willblackhurst solution is similar, but instead of reading back and forth it builds an array by mirroring the original array, to build a continous loop.

As last step you can make the triangle drive the tabread4~ between guardpoints, but I don't think missing guardpoints are the reason for clicks.

(I am also missing externals to test the patch)

True, and I must admit that I am bored with this.
The array sum changes after the first or second sequence or after each sequence. This also happens in my sleek patch: https://forum.pdpatchrepo.info/topic/15834/shuffling-audio-only-allowing-repeats-once-all-options-have-been-played/11

I think OPs requirement has been solved by your patch or others in the thread?

Also permutations have been discussed in this forum before:
https://forum.pdpatchrepo.info/topic/10707/all-possible-permutations-with-repetitions-of-harmonic-volume-from-top-to-bottom/8

https://forum.pdpatchrepo.info/topic/10746/permutations-second-part-can-anybody-get-this-patch-to-work

(I did not read them entirely)

And there are non-Vanilla externals, mentioned in Dizzys thread and fd_lib [combi], an object I don't understand yet.

EDIT and this https://forum.pdpatchrepo.info/topic/5142/generating-all-possible-permutations-of-a-fixed-number-set
and that
https://grrrr.org/2010/11/04/list-permutation-pure_data/

This probably comes closer, a new sequence has all the possibilities but the last number of the last sequence. A new permutation.

randompermutations2.pd

@jameslo wrote:

I also wonder if it's not the worst idea to just generate random permutations and simply throw out the ones you don't like, e.g. ones that start with the same # as the end of the previous permutation.

Yes I think it is not the worst idea!

EDIT: patch does not work https://forum.pdpatchrepo.info/topic/15858/shuffling-in-pd-vanilla

you have to add the excluded # back into the pool at the right point.

There was a little bug: The center output of [list split] needs a [route bang] (as I read in the helpfile now.
randomwmemory.pd

randomwmemory.pd

There are many ways to do this ... @Dizzy-Dizzy Please read the patches and try to understand them.

https://forum.pdpatchrepo.info/topic/14635/how-to-construct-a-background-noise-filter-in-pd

If non-Vanilla externals are ok: zexy/urn
Otherwise for example:
random-wo-repitition.pd
(not cheap!)
With that for example:
https://forum.pdpatchrepo.info/topic/14522/is-it-possible-to-group-different-soundfiles-in-only-one-message/2

[list-find] from list-abs library

What are you referring to by normalize action?
normalize message to an array?
Pd is deterministic! See 2.control.examples/08.depthfirst.pd

@oid said:

this won't work if you have more than one instance of the middle abstraction

I knew I was missing sth.

Can we do

[savestate]
/
[route menusave]
|
[menusave(
|
[sendcanvas]

using the right outlet bang of [savestate] in the middlepatch, then [savestate] in the child as intended.

I agree with @alexandros init state should work, but in an abstraction it doesn't.

if this is the case maybe you could do a work around by sending dirty message to the middle patch, similar to how I did there:
https://forum.pdpatchrepo.info/topic/15732/how-to-save-restore-multiple-0-array-abstractions-in-a-patch/4 maybe with metro instead of ping, this way you get asked to save before closing.

[namecanvas $0-parent]

[loadbang]
|
[1(
|
[metro 20000|
|
[dirty 1 (
|
[s $0- parent]

@witteruis ctrl+click on the red error line to find the error.
Read here below Fig. 2.1.7 https://msp.ucsd.edu/Pd_documentation/resources/chapter2.htm

[osc~] [osc~]
|      |
[max~   ]

or with [min~], this switches the signals themselfs

Decision of using control-rate or signal-rate basicly depends on where the trigger is coming from: signal-rate or control-rate?
(this is the short answer)

I'm not clear on what value "le+23" is meant to indicate.

1e+23 is just a very big number (+ - 1e+32 is Pd's 32-bit single precision space). By multiplying the output of [rzero~ 1] to extremes, then clipping with [clip~ -2 0], turns it into a boolean, either -2 or 0.

Here is @jameslo's example of sigrate flipflop https://forum.pdpatchrepo.info/topic/13918/signal-logic-clocks-division-and-sequences/4
rpole~version should be drop-in replacement.

I tried those

please share your patch

https://forum.pdpatchrepo.info/topic/15750/audio-rate-flipflop