The current help patch for [list] family objects doesn't mention what kind of messages [list] outputs.

I remember reading somewhere (maybe on puredata.info) that for messages consisting of two or more atoms/elements, [list] will output a list, and for empty lists and 1-element messages [list] will output a bang or float/symbol messages, respectively.

But as far as I can tell, [list] doesn't do any conversion at its outlet-- everything it outputs will have the selector "list", and it's up to other objects like [route] to do any conversion they deem necessary.

For example, take the bug with sigbinops:

[put anything you want in this message box(
|
[list]
|

[+~ ]

error: inlet: expected 'signal' but got 'list'

So I'm guessing that [list append/prepend/split] always output messages with the "list" selector. This would seem consistent with the behavior at the inlets.

Do I have this right?

Thanks, Jonathan

--- On Thu, 2/4/10, Claude Heiland-Allen claudiusmaximus@goto10.org wrote:

From: Claude Heiland-Allen claudiusmaximus@goto10.org Subject: Re: [PD] [list] output To: "Jonathan Wilkes" jancsika@yahoo.com Cc: pd-list@iem.at Date: Thursday, February 4, 2010, 6:22 PM My top 3 related annoyances (in pd-0.41-4 on Linux).

1. "1 2   "--[route

1   ]--[*~] // inlet~ no method for list 2. "1     "--[list trim ]--[*~] // inlet~ no method for list 3. "list 1"--[route list]--.... // outputs on "unmatched" outlet

So would you also want [list( -> [route list] to output on the "list" outlet?

-Jonathan

ypatios wrote:

hello

[1, 1 2, a, a b, a 2 b ( | [list ] | [print ]

never believe what print is telling you! print does an intpretion of the incoming message and will generate the output accordingly.

gfmasdr IOhannes

--- On Thu, 2/4/10, ypatios ypatios@gmail.com wrote:

From: ypatios ypatios@gmail.com Subject: Re: [PD] [list] output To: "Jonathan Wilkes" jancsika@yahoo.com Cc: pd-list@iem.at Date: Thursday, February 4, 2010, 6:31 PM hello

But as far as I can tell, [list] doesn't do any conversion at its outlet--

everything it outputs will have the selector "list", and it's up to other

objects like [route] to do any conversion they deem necessary.

For example, take the bug with sigbinops:

[put anything you want in this message box(

|

[list]

|

[+~ ]

error: inlet: expected 'signal' but got 'list'

[1, 1 2, a, a b, a 2 b ( | [list ] | [print ]

gives the following (0.42-5):

###### print: 1 print: 1 2 print: symbol a print: list a b print: list a 2 b

######

so [list ] does some kind of conversion. what i don't understand is why there is no selector printed in the first two messages, where only numbers are present.

As far as your output from [print] goes:

For "1" and "1 2", they are an implicit float and implicit list, respectively. This is a convenience feature of Pd: you don't have to type "float 1" every time you want to send a number from a msg box to a [+] object, for example.

For "list a b" and "list a 2 b", you would expect this because all non-list messages are converted to lists at the _inlets_ of the [list] object.

So that leaves "symbol a" unexplained: why isn't it "list a" if my theory is correct? It's because the [print] object decided to print it that way: it receives a message "list a" and calls the print_list method, which decides to print "symbol a" instead of "list a".

Here's a trick that shows clearly what is happening in Pd 0.42-5, using the -n flag of [print], which suppresses the word "print":

[1( | [print]

[1( | [print -n]

For the first one, the print_method (x_interface.c) prints the word "print", followed by a colon, and the float itself. For the second one, the word "print" is absent, so print_method then decides not to print a colon either. Only the float is printed.

However, consider the following:

[1( | [list] | [print -n]

Now the [print] calls the method print_list because the incoming message is a list. How do I know this? Because of a bug: instead of just printing the float-atom (as above), the output is ": 1", which matches the way print_list is telling Pd to format a list with at least one element that is not a symbol. [1]

Likewise, the output "symbol a" you referred to above enters [print] as "list a", and gets formatted as "symbol a" because it's a one-element list, with that element being a symbol-atom.

This is complicated by the fact that objects like [route] do a similar trick.

Ok, now I'm pretty convinced that [list] always outputs a list.

[1]- On older versions of Pd, you can still see a discrepancy if, for example, you compare the way [print] formats a gpointer, and a list containing one gpointer.

-Jonathan

anyways, you should provide [+~ ] operator with an argument if you want it to expect messsages instead of a signal (check help patch) :-)

ciao

-- ypatios