Maybe with IOhannes' latest request, initbang?

~Kyle

On 8/22/06, Max Neupert abonnements@revolwear.com wrote:

Am 22.08.2006 um 21:40 schrieb Frank Barknecht:

...

Hallo, Max Neupert hat gesagt: // Max Neupert wrote:

...

i want to force pd to reload a patch. why? because i want to manipulate a gem patch running on a networked machine from another machine. and yes, i've had a look on netpd and serendipd.

How much reloading do you want? If another execution of the loadbang is enough, then you can send the message "loadbang" to the patch's receiver made from prepending "pd-" to the name of the subpatch or the abstraction. See att.

hallo frank & list,

it's not enough reloading. i want to manipulate the objects and continue patching on a remote machine, then save the altered patch and finally send the machine running the patch a signal make it reload the patch to hear/see the changes made. is this possible somehow?

thanks, max

---

PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list

Am 23.08.2006 um 03:13 schrieb Mathieu Bouchard:

On Tue, 22 Aug 2006, Max Neupert wrote:

...

it's not enough reloading. i want to manipulate the objects and
continue patching on a remote machine, then save the altered patch
and finally send the machine running the patch a signal make it
reload the patch to hear/see the changes made. is this possible
somehow?

If there's something stored in the right inlet of an object, should
that be reset or not?

salut matju, hi list,

well actually i thought that's anyway impossible, so those values
would be lost. and this would be okay for now. but on the other hand your counter-question implies that i may just
choose from both available options freely.. so i would be curious how
to do a patch reload at all an then one were we still have the stored
values of the objects...

max

Am 23.08.2006 um 03:57 schrieb Mathieu Bouchard:

On Wed, 23 Aug 2006, Max Neupert wrote:

...

well actually i thought that's anyway impossible, so those values
would be lost. and this would be okay for now.

So, what if you just close the patch and reopen it? You can do that
with pd messages, such as "; open filename directory". Or else,
even just restart pd with the same arguments.

hi list, matju,

i think this approach will do what i want. the message is actually:

[; pd open filename directory(

how do i close the patch remotely later?

[; pd close filename(

doesn't do the trick. where do i find a reference for those messages to pd?

inside src/m_glob.c are some hints, but i did not find something to
close a patch.

thanks,

m.

...

but on the other hand your counter-question implies that i may
just choose from both available options freely..

Not necessarily, I just wanted to know whether you had thought
about the losing the state of your objects and whether it mattered
to you.

...

so i would be curious how to do a patch reload at all an then one
were we still have the stored values of the objects...

it's not possible without tracking which object is which. This
might be best done with one of pd's state-saving systems.

thank you, i'll go though those help patches.

m.

Am 29.08.2006 um 18:42 schrieb Enrique Erne:

[; pd-patchname.pd menuclose 1(

found it in pd-extended: manuals/pd-msg/2.msg_and_pd/2.open_close.pd

...

how do i close the patch remotely later?

[; pd close filename(

doesn't do the trick. where do i find a reference for those messages to pd?

inside src/m_glob.c are some hints, but i did not find something
to close a patch.

thanks,

m.

[; pd open pix_film_transloc.pd /Users/max/Documents/(

[; pd-pix_film_transloc.pd menuclose(

or

[; pd-pix_film_transloc.pd menuclose 1(

works here pd-0.39-2 and pd-0.40-0test4

regards eni

On Aug 29, 2006, at 7:43 PM, Max Neupert wrote:

hm. still does not work as expected, i have no explanation, what's the magic?

<open_close.pd>

Am 29.08.2006 um 18:42 schrieb Enrique Erne:

...

[; pd-patchname.pd menuclose 1(

found it in pd-extended: manuals/pd-msg/2.msg_and_pd/2.open_close.pd

...

how do i close the patch remotely later?

[; pd close filename(

doesn't do the trick. where do i find a reference for those messages to pd?

inside src/m_glob.c are some hints, but i did not find something to close a patch.

thanks,

m.

On Wed, 30 Aug 2006, Max Neupert wrote:

now it does. all-though it crashes my pd completely if i try to close while there are more than one with the same name opened. this is probably not a feature.

Pd's symbol-table allows only one receiver per symbol. On top of this, pd_bind and pd_unbind implement a kind of splitter using a hidden [bindlist] object, which is only ever used internally, and only when there's more than one receiver per receive-symbol.

When you close several patches through that symbol, the message is handled a method in [bindlist], which terminates only when all receivers are finished handling the message.

When a canvas closes, it removes itself from the receivers list using pd_unbind. When the number of receivers goes from 2 to 1, pd_unbind deletes the [bindlist] object. At this point, [bindlist]'s method has not finished running, which is what causes the crash.

_ _ __ ___ _____ ________ _____________ _____________________ ... | Mathieu Bouchard - tél:+1.514.383.3801 - http://artengine.ca/matju | Freelance Digital Arts Engineer, Montréal QC Canada

That is an awesome solution!

~Kyle

On 8/29/06, Chris McCormick chris@mccormick.cx wrote:

On Tue, Aug 29, 2006 at 10:54:55PM -0400, Mathieu Bouchard wrote:

...

On Wed, 30 Aug 2006, Max Neupert wrote:

...

now it does. all-though it crashes my pd completely if i try to close while there are more than one with the same name opened. this is probably not a feature.

When a canvas closes, it removes itself from the receivers list using pd_unbind. When the number of receivers goes from 2 to 1, pd_unbind deletes the [bindlist] object. At this point, [bindlist]'s method has not finished running, which is what causes the crash.

Yikes!

I am not sure if this would suffer the same problem, but what I do for this kind of dynamic opening and closing of patches is have a bunch of subpatches in my main patch called:

[pd inst1] [pd inst2] [pd inst3] etc.

Then I send internal messages to them to add objects called the patch I want to dynamically open. When it's time to delete those patches, I issue a [clear( to the subpatch which closes and deletes the instance of the dynamically open patch. I am pretty sure I have had multiple versions of the same patch open in this way before.

Maybe you can use this method Max? If you'd like an example I can send one through.

Best rgds,

Chris.

---

chris@mccormick.cx http://mccormick.cx

---

PD-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list