Hi all,
We know how to change the background color of canvas, is it possible to change the background color of objects on canvas in a similar way ?
I'd like to have a grey canvas with white objects.
Many thanks
Cheers
Alessandro Fogar
It is not really possible to do this because the objects don't really have a background - they are just lines and text. It should be possible to make PD use TK canvas polygons so that they have a background color (like Guenter's objects) but there are some issues with the stacking of the objects. Notice that since the current objects are 'transparent' when they overlap nothing is done about the order that they are drawn. PD will need a notion of depth in addition to horizontal/vertical position before this will be possible in a nice way. It would be nice, though.
Karl
On Fri, 29 Sep 2000, sfogar@libero.it wrote:
Hi all,
We know how to change the background color of canvas, is it possible to change the background color of objects on canvas in a similar way ?
I'd like to have a grey canvas with white objects.
Many thanks
Cheers
Alessandro Fogar
I have hacked together a GTK+ based gui for pd - here is a screenshot:
http://gigue.peabody.jhu.edu/~karlmac/pdgui-shot.png
Please don't get too excited by the fact that the objects have a white background while the canvas background is grey - I just did that as an experiment (and of course took a screenshot of one of the few patches that it worked well with it :)). The gui is in a workable state (a few of you saw it at the PD users group meeting at ICMC) but has a little way to go before it is really ready to be used instead of the Tcl/TK gui. I am announcing this now because I don't have much time to work on it at the moment and I thought someone might be interested in hacking on it a little. Here is the download (http only):
http://gigue.peabody.jhu.edu/~karlmac/download/pd-gui.tar.gz
If anyone actually wants to do some work on this let me know - I have a cvs repository that I can give you some access to. Also, if anyone wants to just try and use this send a note to the list and I will give you some compilation instructions. One good thing is that it doesn't require changes to pd so it can be used interchangeably with the current pd-gui.
Enjoy!
Karl
Hi Karl,
I'll try your GTK version...
many thanks
Alessandro Fogar
Karl MacMillan wrote:
It is not really possible to do this because the objects don't really have a background - they are just lines and text. It should be possible to make PD use TK canvas polygons so that they have a background color (like Guenter's objects) but there are some issues with the stacking of the objects. Notice that since the current objects are 'transparent' when they overlap nothing is done about the order that they are drawn. PD will need a notion of depth in addition to horizontal/vertical position before this will be possible in a nice way. It would be nice, though.
Karl
On Fri, 29 Sep 2000, sfogar@libero.it wrote:
Hi all,
We know how to change the background color of canvas, is it possible to change the background color of objects on canvas in a similar way ?
I'd like to have a grey canvas with white objects.
Many thanks
Cheers
Alessandro Fogar
-- _____________________________________________________ | Karl W. MacMillan | | Computer Music Department | | Peabody Institute of the Johns Hopkins University | | karlmac@peabody.jhu.edu | | www.peabody.jhu.edu/~karlmac |
hi there,
since there has been some chatting about gui stuff recently, i thought i'd mention something i've been thinking about a bit recently.
often i run into that situation where i'm dragging a number box, and i hit the edge of the display, and i have to stop dragging, move the mouse back to the number box and start dragging again. then i was thinking that it would be grat if, while dragging on a number box (or maybe a new kind of object) that the mouse got 'trapped' over the object, so that you could keep scrolling an infinite distance in any direction without beng limited by screen realestate.
i actually went out and got myself a trackball, hoping that i might be able to impliment this feature, and i would even be able to eliminate the inherent jerkiness off recentering the mouse when it reaches the edge of the mousemat.
thenn .. (are you still with me?... this is getting kind of long) i was thinking that it might also be useful to have this 'mouse trap' be able to maintain two values simultaneously, for the vertical and horizontal movement of the mouse. like the xy mouse reader objects in csound.
soanyhow, i was wondering if anyone had tried this, thought about it, wouldf like it etc? i havent played withh externals much, and i'm not sure if it's possible for an external to make the Xcalls that would be needed.
pix.
-- This .sig last modified Thu Jul 01 00:00:00 GMT 1970
David Z has done something good along these lines: he's fixed it on the Mac so that if your mouse cursor hits the top of the screen it just jumps down so you can keep dragging it up!
Others have invented rules like dx * (1 + dy * dy) (so that the further off center you drag the faster it goes) but it's counterintuitive for users.
I don't know how to pull David's trick off in TK; but I think it's easy enough for externs to track the mouse by imitating the way the number box does it.
On another topic, I've noticed that at least 5 people are interested in using Pd with alternative front ends... I'm working now on command-line options to make this easier; I hope to put a "test" version (Linux only) out within a week. Oh yes, new "data" stuff too...
cheers Miller
On Mon, Oct 02, 2000 at 01:33:47AM +0200, pix wrote:
hi there,
since there has been some chatting about gui stuff recently, i thought i'd mention something i've been thinking about a bit recently.
often i run into that situation where i'm dragging a number box, and i hit the edge of the display, and i have to stop dragging, move the mouse back to the number box and start dragging again. then i was thinking that it would be grat if, while dragging on a number box (or maybe a new kind of object) that the mouse got 'trapped' over the object, so that you could keep scrolling an infinite distance in any direction without beng limited by screen realestate.
i actually went out and got myself a trackball, hoping that i might be able to impliment this feature, and i would even be able to eliminate the inherent jerkiness off recentering the mouse when it reaches the edge of the mousemat.
thenn .. (are you still with me?... this is getting kind of long) i was thinking that it might also be useful to have this 'mouse trap' be able to maintain two values simultaneously, for the vertical and horizontal movement of the mouse. like the xy mouse reader objects in csound.
soanyhow, i was wondering if anyone had tried this, thought about it, wouldf like it etc? i havent played withh externals much, and i'm not sure if it's possible for an external to make the Xcalls that would be needed.
pix.
-- This .sig last modified Thu Jul 01 00:00:00 GMT 1970
|On another topic, I've noticed that at least 5 people are interested in using |Pd with alternative front ends... I'm working now on command-line options |to make this easier; I hope to put a "test" version (Linux only) out within a |week. Oh yes, new "data" stuff too...
while on frontends, how about running pd + pd-gui on two different machines=? so that you can start pd -nogui bla.pd on for example remote machine, later connect to it comme ./mypdfrontend -host remote.machine.com .. when you quit the frontend, the barebones pd just remians runnign, much like vnc (http://www.uk.research.att.com/vnc/) works .. the tk interface already does (some of its) communication via local network with the pd-backend ?
also, miller can you make receive's (and send's) symbol settable?
netreceive access control?
bzt.L
-- I x d V . O R G
|> + |> netreceive access control? | |what for ?
yes sure, there's a few ways of obfuscating everything (a little), but in a longer run i want to be able to restrict access to my netreceives like what hosts are allowd to connect at all, and what hosts are allowed to send what messages (and maybe other things) because i dont want every psycho outhere to be able to quit my pd's and do other stupid things ...
On Mon, 2 Oct 2000, [ISO-8859-1] _-�-_ wrote:
|On another topic, I've noticed that at least 5 people are interested in using |Pd with alternative front ends... I'm working now on command-line options |to make this easier; I hope to put a "test" version (Linux only) out within a |week. Oh yes, new "data" stuff too...
while on frontends, how about running pd + pd-gui on two different machines=? so that you can start pd -nogui bla.pd on for example remote machine, later connect to it comme ./mypdfrontend -host remote.machine.com .. when you quit the frontend, the barebones pd just remians runnign, much like vnc (http://www.uk.research.att.com/vnc/) works .. the tk interface already does (some of its) communication via local network with the pd-backend ?
Using the gui and pd on separate machines should be fairly easy - all of the communication between the two is via standard tcp sockets.
Karl
... yes; I've made it easier still (I hope) with the "-guicmd" option.
I'll have to get on a Macintosh to see how variable sends and receives are done there (I agree it's time to put them in!)...
cheers Miller
Using the gui and pd on separate machines should be fairly easy - all of the communication between the two is via standard tcp sockets.
Karl
--
Great!
This means that getting my gtk+ gui going is as simple as:
- download the source from http://www.peabody.jhu.edu/~karlmac/download/
- tar zxvf pd-gui.tar.gz somewhere convenient.
- cd to pd-gui and run ./autogen.sh (this will make and run the configure script).
- make (don't make install - just make).
After this if all went well you will have a binary in pd-gui/src called pd_gui. You can either make a symlink in the pd/bin directory to this binary called pd-gui or use the new -guicmd option. The symlink version is done with:
cd /path/to/pd/bin mv pd-gui pd-gui-tk ln -s /path/to/pd-gui/src/pd_gui pd-gui
Have fun,
Karl
On Mon, 2 Oct 2000, Miller Puckette wrote:
... yes; I've made it easier still (I hope) with the "-guicmd" option.
I'll have to get on a Macintosh to see how variable sends and receives are done there (I agree it's time to put them in!)...
cheers Miller
Using the gui and pd on separate machines should be fairly easy - all of the communication between the two is via standard tcp sockets.
Karl
--
Hi Karl,
I tried to use the gtk_gui but, on my system when I try to start pd, the following message appears after the sounddevice opening:
Function not in object 0x80870e0 hash table >pdtk_watchdog< recv: Connection reset by peer (104)
If I don't start pd with -rt I obtain:
pd: exiting
After the sounddevice opening.
Can you help me please ?
Cheers
Alessandro Fogar
Karl MacMillan wrote:
Great!
This means that getting my gtk+ gui going is as simple as:
- download the source from http://www.peabody.jhu.edu/~karlmac/download/
- tar zxvf pd-gui.tar.gz somewhere convenient.
- cd to pd-gui and run ./autogen.sh (this will make and run the configure script).
- make (don't make install - just make).
After this if all went well you will have a binary in pd-gui/src called pd_gui. You can either make a symlink in the pd/bin directory to this binary called pd-gui or use the new -guicmd option. The symlink version is done with:
cd /path/to/pd/bin mv pd-gui pd-gui-tk ln -s /path/to/pd-gui/src/pd_gui pd-gui
Have fun,
Karl
On Mon, 2 Oct 2000, Miller Puckette wrote:
... yes; I've made it easier still (I hope) with the "-guicmd" option.
I'll have to get on a Macintosh to see how variable sends and receives are done there (I agree it's time to put them in!)...
cheers Miller
Using the gui and pd on separate machines should be fairly easy - all of the communication between the two is via standard tcp sockets.
Karl
--
-- _____________________________________________________ | Karl W. MacMillan | | Computer Music Department | | Peabody Institute of the Johns Hopkins University | | karlmac@peabody.jhu.edu | | www.peabody.jhu.edu/~karlmac |
On Wed, 4 Oct 2000, Alessandro Fogar wrote:
Hi Karl,
I tried to use the gtk_gui but, on my system when I try to start pd, the following message appears after the sounddevice opening:
Function not in object 0x80870e0 hash table >pdtk_watchdog< recv: Connection reset by peer (104)
This makes some sense - the watchdog functions are not implemented so you should not run pd with -rt. This should not kill the socket, however.
If I don't start pd with -rt I obtain:
pd: exiting
After the sounddevice opening.
Can you help me please ?
I'm not quite certain what the problem is - the pd: exiting means that pd encountered some problem. Did you use all of the same options as with -rt when you did it without? Also, what pd version are you using?
Karl
Cheers
Alessandro Fogar
Karl MacMillan wrote:
Great!
This means that getting my gtk+ gui going is as simple as:
- download the source from http://www.peabody.jhu.edu/~karlmac/download/
- tar zxvf pd-gui.tar.gz somewhere convenient.
- cd to pd-gui and run ./autogen.sh (this will make and run the configure script).
- make (don't make install - just make).
After this if all went well you will have a binary in pd-gui/src called pd_gui. You can either make a symlink in the pd/bin directory to this binary called pd-gui or use the new -guicmd option. The symlink version is done with:
cd /path/to/pd/bin mv pd-gui pd-gui-tk ln -s /path/to/pd-gui/src/pd_gui pd-gui
Have fun,
Karl
On Mon, 2 Oct 2000, Miller Puckette wrote:
... yes; I've made it easier still (I hope) with the "-guicmd" option.
I'll have to get on a Macintosh to see how variable sends and receives are done there (I agree it's time to put them in!)...
cheers Miller
Using the gui and pd on separate machines should be fairly easy - all of the communication between the two is via standard tcp sockets.
Karl
--
-- _____________________________________________________ | Karl W. MacMillan | | Computer Music Department | | Peabody Institute of the Johns Hopkins University | | karlmac@peabody.jhu.edu | | www.peabody.jhu.edu/~karlmac |
Hi Karl,
I'm using pd 0.31.
If I start pd without my usual .pdrc I experience the same problem, starting pd with -verbose doesn't help.
When I tried without the -rt I used my usual options but as I said before without any options I obtain the same behaviour.
Ciao
Alessandro
Karl MacMillan wrote:
On Wed, 4 Oct 2000, Alessandro Fogar wrote:
Hi Karl,
I tried to use the gtk_gui but, on my system when I try to start pd, the following message appears after the sounddevice opening:
Function not in object 0x80870e0 hash table >pdtk_watchdog< recv: Connection reset by peer (104)
This makes some sense - the watchdog functions are not implemented so you should not run pd with -rt. This should not kill the socket, however.
If I don't start pd with -rt I obtain:
pd: exiting
After the sounddevice opening.
Can you help me please ?
I'm not quite certain what the problem is - the pd: exiting means that pd encountered some problem. Did you use all of the same options as with -rt when you did it without? Also, what pd version are you using?
Karl
|... yes; I've made it easier still (I hope) with the "-guicmd" option. | |I'll have to get on a Macintosh to see how variable sends and receives are |done there (I agree it's time to put them in!)...
oke, in the meantime a quick plaster for receive only against x_connective.c from pd-0.31, ... it works sofar .. ;)
cd pd/src patch x_connective.c x_connective.diff make
-
unknown＠example.com -
_-¯-_ -
Alessandro Fogar -
Karl MacMillan -
Miller Puckette -
pix -
sfogar＠libero.it