Hi,
I installed the version on the title, and started doing some work with it. Here is a list of issues, big and small, as I go along:
- lots of extra calls in the console while doing normal actions (new file,
etc. etc.). won't list them all, I guess it might be a debugging thing.
alt-[key] doesn't work on menus? alt-f should make file menu open
path+startup windows with too small font size
text editor doesn't work, is he already gone? data properties' editor
works.
- don't know what popup mode is and saw no obvious difference - but I
didn't bother to look for it as well
nice new design for the windows
in the past I asked for pd to save window position also when windows are
saved on a 2nd screen (windows were always displayed on the 1st screen). today this backfired, as while working with 1 screen (on a "2-screen" patch) not all windows of the patch were visible. canvas values were "#N canvas 1422 54" on pd file. While I salute the possibilty of having windows automatically appear on multiple screens, how about also making sure that they appear only in 1 screen of there aren't any more? E.g. if this window's X value is 1422 and my x resolution is 1400, probably subtracting the screen resolution of the window value should be enough. can tcl/tk get these elements? Another representation of what I meant: if canvasres >= screenres, then canvasres=canvasres-screenres (apply to both x and y)
- some automatic redrawing while mousing over data structures (no edit
mode, no clicking, just moving the mouse). sliders blink, and uses cpu for redrawing.
- I get the impression that redrawing of data structures with toggle in
the inlet (i.e. visible/invisible) is slower now. maybe just an impression? but anyway, is it possible to enhance this, is the gui work related with that?
- I use the [hcs/sys_gui] command to make my own color scheme (white is
too agressive) and to place the pd window on a corner (wm geometry). These don't work now, are they going to be obsolete, or they have to adapt to the new tcl code?
- font window: only works on patch if it's opened on a (any) patch window.
if I do it from pd window, only affects pd window.
- x and y, x only + y only have all the same result. but I think that was
the normal behaviour before anyway (never used this feature)
- I can only use courier now, wasn't verdana the latest standard? (maybe
not)
- when I do a find in a pd window, it looks in the whole patch. wasn't it
before that it only looked deeper from the call window, instead of the whole patch? (i.e., if I do find on the subpatch "aaa", it only searches on patches deeper than it) In case I was dreaming before, is it possible to make find work on a deep-only modus? I think this might be more logical for big patches (one can always go to the grandparent patch to search the whole thing)
- nice touch the "search in xxx.pd for"
- what's exactly the "match whole work only" for? I searched for an
incomplete string ("edit-ind" instead of "edit-index"), and neither of them found "edit-ind". besides for obvious purposes, this would be useful to find objects that have a $0 before - as replacing $0 by its number doesn't work in the search function.
- audio meters (in sc style?) are nice. but:
- is it possible to start them in any menu/command besides the
clickbox? - out button in connected to the actual sound volume of the system, and not to pd dac~ output (independently of how many there are). the latter would make more sense.
- dragging of patch fragments (several objects) around the canvas: works
better in display purposes, but still uses lots of cpu (maybe even more?). anything can be done about that?
- Help->About: Pd doesn't work (so I can't tell exactly which version it
is)
- Help->Html manual: opens up in internet explorer instead of opening up
in my standard browser. that didn't happen before
- Help->Browser: "ERROR: menu_doc_browser non-directory" pops up in the
console. But the browser works. Btw, would it be possible to add scalable windows + key follow in the browser? (with key follow I mean that pressing "b" takes me to the first file with b in the window, "ba" to "ba...", etc etc
Anything else specific you want to be checked, Hans?
João
On Oct 14, 2009, at 8:31 AM, João Pais wrote:
Hi,
I installed the version on the title, and started doing some work
with it. Here is a list of issues, big and small, as I go along:
- lots of extra calls in the console while doing normal actions (new
file, etc. etc.). won't list them all, I guess it might be a debugging
thing.
yes, debug messages...
- alt-[key] doesn't work on menus? alt-f should make file menu open
Does that work on other versions of Pd?
- path+startup windows with too small font size
- text editor doesn't work, is he already gone? data properties'
editor works.
It just needs to be implemented. Any volunteers? I've never used it
so I don't know what its supposed to do. Or really, the better
approach IMHO is to make the in-place editing good enough so you don't
need the Text Editor.
- don't know what popup mode is and saw no obvious difference - but I
didn't bother to look for it as well
Popup mode means the Pd window will popup to the front if any message
is printed there. I think I'll move that to a plugin.
nice new design for the windows
in the past I asked for pd to save window position also when
windows are saved on a 2nd screen (windows were always displayed on the 1st
screen). today this backfired, as while working with 1 screen (on a "2-screen" patch) not all windows of the patch were visible. canvas values were
"#N canvas 1422 54" on pd file. While I salute the possibilty of having windows automatically appear on multiple screens, how about also
making sure that they appear only in 1 screen of there aren't any more? E.g. if this window's X value is 1422 and my x resolution is 1400, probably subtracting the screen resolution of the window value
should be enough. can tcl/tk get these elements? Another representation of
what I meant: if canvasres >= screenres, then canvasres=canvasres-screenres (apply
to both x and y)
I don't have a multi-monitor setup, so I can't test this. Ideally
you'd edit the code to get it working properly. It should be pretty
straightforward, the code in question is in pdtk_canvas.tcl in the
proc called 'pdtk_canvas_new'. You can see it gets the 'geometry'
from Pd as an argument to the proc, and is then set using "wm
geometry $mytoplevel $geometry" If you just do the math before
running the 'wm geometry' and put the right values into $geometry,
then 'wm geometry' should do the right thing.
- some automatic redrawing while mousing over data structures (no edit
mode, no clicking, just moving the mouse). sliders blink, and uses
cpu for redrawing.
Yeah, there seems to be a bug in Tcl/Tk that makes the hover/cursor
change events trigger a window configure event, as if the window was
resized. I haven't figured that one out, can someone try their hand
at this one?
- I get the impression that redrawing of data structures with toggle
in the inlet (i.e. visible/invisible) is slower now. maybe just an impression? but anyway, is it possible to enhance this, is the gui
work related with that?
Do you have an example patch?
- I use the [hcs/sys_gui] command to make my own color scheme (white
is too agressive) and to place the pd window on a corner (wm geometry).
These don't work now, are they going to be obsolete, or they have to adapt
to the new tcl code?
They should work fine, but the color scheme variables are part of Pd-
extended, this is still Pd-vanilla. I think you'd be better off
porting your color scheme to a Tcl plugin. Its not hard, plus you can
do a lot more with it.
http://puredata.info/dev/PdGuiRewriteTheming
- font window: only works on patch if it's opened on a (any) patch
window. if I do it from pd window, only affects pd window.
Ok, it was working for me on Windows and Mac OS X, but was behaving
oddly. I fixed the odd behavior.
- x and y, x only + y only have all the same result. but I think
that was the normal behaviour before anyway (never used this feature)
I've never used it either, I think it might be really old cruft...
- I can only use courier now, wasn't verdana the latest standard?
(maybe not)
It was Bitstream Vera Sans Mono with Pd-extended.
- when I do a find in a pd window, it looks in the whole patch.
wasn't it before that it only looked deeper from the call window, instead of the whole patch? (i.e., if I do find on the subpatch "aaa", it only
searches on patches deeper than it) In case I was dreaming before, is it possible to make find work on a deep-only modus? I think this might be more logical for big patches
(one can always go to the grandparent patch to search the whole thing)
This is also just a new interface to old Pd code, so its a separate
issue. File a bug/feature report if you want to follow up on it.
- nice touch the "search in xxx.pd for"
- what's exactly the "match whole work only" for? I searched for an
incomplete string ("edit-ind" instead of "edit-index"), and neither of them found "edit-ind". besides for obvious purposes, this would be
useful to find objects that have a $0 before - as replacing $0 by its number doesn't work in the search function.
That's the idea, but is perhaps buggy. That was introduced in Pd
0.42, so it separate from the GUI rewrite.
- audio meters (in sc style?) are nice. but:
- is it possible to start them in any menu/command besides the
clickbox?
Same as old versions: [pd meters $1(
- out button in connected to the actual sound volume of the
system, and not to pd dac~ output (independently of how many there are). the
latter would make more sense.
This is also just a new interface to old Pd code, so its a separate
issue. File a bug/feature report if you want to follow up on it.
- dragging of patch fragments (several objects) around the canvas:
works better in display purposes, but still uses lots of cpu (maybe even
more?). anything can be done about that?
There are many things to do on that, but it'll take a fair amount of
work, so we'll need more contributions to get that done.
- Help->About: Pd doesn't work (so I can't tell exactly which
version it is)
Yeah... not implemented yet... it tells you the version at startup in
the Pd window.
- Help->Html manual: opens up in internet explorer instead of
opening up in my standard browser. that didn't happen before
It uses the internal Windows method for opening links, same as
before. That code is old.
- Help->Browser: "ERROR: menu_doc_browser non-directory" pops up in
the console. But the browser works. Btw, would it be possible to add scalable
windows + key follow in the browser? (with key follow I mean that pressing "b"
takes me to the first file with b in the window, "ba" to "ba...", etc etc
It would be very nice for sure, its just a matter of someone doing the
work. Any volunteers to redo the help browser?
Anything else specific you want to be checked, Hans?
It would be great if you can use it for daily use, and report issues.
I'm starting a project where I will be doing the same thing. That's
where we will really discover bugs and issues.
.hc
João
-- Friedenstr. 58 10249 Berlin (Deutschland) Tel +49 30 42020091 | Mob +49 162 6843570 Studio +49 30 69509190 jmmmpais@googlemail.com | skype: jmmmpjmmmp
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
http://at.or.at/hans/
- alt-[key] doesn't work on menus? alt-f should make file menu open
Does that work on other versions of Pd?
the only other version I have is vanilla 0.42-5. guess what, it doesn't
work as well. probably it never did.
would it be part of the job now to make it work?
- text editor doesn't work, is he already gone? data properties' editor
works.
It just needs to be implemented. Any volunteers? I've never used it so
I don't know what its supposed to do. Or really, the better approach
IMHO is to make the in-place editing good enough so you don't need the
Text Editor.
actually I never used the text editor, and don't know if anyone did (the
data structures editor yes, but that's independent). maybe it's better
just to take it out?
- remembered something that I sugested already: what about return closes
an object while typing? eliminates the need for an extra click.
- how about also making the line breaks on comments to work, would it go
into this work batch?
- in the past I asked for pd to save window position also when windows
are saved on a 2nd screen (windows were always displayed on the 1st screen). today this backfired, as while working with 1 screen (on a "2-screen" patch) not all windows of the patch were visible. canvas values were "#N canvas 1422 54" on pd file. While I salute the possibilty of having windows automatically appear on multiple screens, how about also making sure that they appear only in 1 screen of there aren't any more? E.g. if this window's X value is 1422 and my x resolution is 1400, probably subtracting the screen resolution of the window value should be enough. can tcl/tk get these elements? Another representation of what I meant: if canvasres >= screenres, then canvasres=canvasres-screenres (apply to both x and y)
I don't have a multi-monitor setup, so I can't test this. Ideally you'd
edit the code to get it working properly. It should be pretty
straightforward, the code in question is in pdtk_canvas.tcl in the proc
called 'pdtk_canvas_new'. You can see it gets the 'geometry' from Pd as
an argument to the proc, and is then set using "wm geometry $mytoplevel
$geometry" If you just do the math before running the 'wm geometry' and
put the right values into $geometry, then 'wm geometry' should do the
right thing.
the problem is that the only language I can program in is Pd. C and others
I don't really know how to do anything.
It should be very easy to test for you, in case you want to: just open
your pd file in a text editor, locate any line that defines a windows -
format "#N canvas 914 187 335 381 gui 0;", being "gui" the window name -,
and add 1000 or something to the first number (x position). If you can't
see it when you open the patch in pd, it means it's still not ok.
- I get the impression that redrawing of data structures with toggle in
the inlet (i.e. visible/invisible) is slower now. maybe just an impression? but anyway, is it possible to enhance this, is the gui work related with that?
Do you have an example patch?
I have a complicated patch with some structures in it, I can send to it
you later. but I guess any patch will do - including your solitude.
- I use the [hcs/sys_gui] command to make my own color scheme (white is
too agressive) and to place the pd window on a corner (wm geometry).
These don't work now, are they going to be obsolete, or they have to adapt to the new tcl code?They should work fine, but the color scheme variables are part of Pd- extended, this is still Pd-vanilla. I think you'd be better off porting
your color scheme to a Tcl plugin. Its not hard, plus you can do a lot
more with it.
sorry, but I can't understand what "you can drop them into pd/startup"
means. I see no folder or file with that name here.
- x and y, x only + y only have all the same result. but I think that
was the normal behaviour before anyway (never used this feature)
I've never used it either, I think it might be really old cruft...
I would ask the higher powers (miller/iem people) and get rid of it.
- Help->About: Pd doesn't work (so I can't tell exactly which version it
is)
Yeah... not implemented yet... it tells you the version at startup in
the Pd window.
Pd version 0.43-0devel-20091008, in case it's important to know.
Anything else specific you want to be checked, Hans?
It would be great if you can use it for daily use, and report issues.
I'm starting a project where I will be doing the same thing. That's
where we will really discover bugs and issues.
should be doing that from tomorrow, maybe. doesn't the ubuntu version
comes with the pd-ext material, like the windows one?
On Oct 14, 2009, at 8:19 PM, João Pais wrote:
- alt-[key] doesn't work on menus? alt-f should make file menu open
Does that work on other versions of Pd?
the only other version I have is vanilla 0.42-5. guess what, it
doesn't work as well. probably it never did. would it be part of the job now to make it work?
Sure, sounds like something good to have. Do you want to take on that
project?
- text editor doesn't work, is he already gone? data properties'
editor works.
It just needs to be implemented. Any volunteers? I've never used
it so I don't know what its supposed to do. Or really, the better
approach IMHO is to make the in-place editing good enough so you
don't need the Text Editor.actually I never used the text editor, and don't know if anyone did
(the data structures editor yes, but that's independent). maybe it's
better just to take it out?
- remembered something that I sugested already: what about return
closes an object while typing? eliminates the need for an extra click.
- how about also making the line breaks on comments to work, would
it go into this work batch?
- in the past I asked for pd to save window position also when
windows are saved on a 2nd screen (windows were always displayed on the 1st
screen). today this backfired, as while working with 1 screen (on a "2- screen" patch) not all windows of the patch were visible. canvas values
were "#N canvas 1422 54" on pd file. While I salute the possibilty of having windows automatically appear on multiple screens, how about also
making sure that they appear only in 1 screen of there aren't any more? E.g. if this window's X value is 1422 and my x resolution is 1400, probably subtracting the screen resolution of the window value
should be enough. can tcl/tk get these elements? Another representation of
what I meant: if canvasres >= screenres, then canvasres=canvasres-screenres
(apply to both x and y)I don't have a multi-monitor setup, so I can't test this. Ideally
you'd edit the code to get it working properly. It should be
pretty straightforward, the code in question is in pdtk_canvas.tcl
in the proc called 'pdtk_canvas_new'. You can see it gets the
'geometry' from Pd as an argument to the proc, and is then set
using "wm geometry $mytoplevel $geometry" If you just do the math
before running the 'wm geometry' and put the right values into
$geometry, then 'wm geometry' should do the right thing.the problem is that the only language I can program in is Pd. C and
others I don't really know how to do anything. It should be very easy to test for you, in case you want to: just
open your pd file in a text editor, locate any line that defines a
windows - format "#N canvas 914 187 335 381 gui 0;", being "gui" the
window name -, and add 1000 or something to the first number (x
position). If you can't see it when you open the patch in pd, it
means it's still not ok.
Tcl is not hard, and there is nothing to be lost by trying it. :-D
- I get the impression that redrawing of data structures with
toggle in the inlet (i.e. visible/invisible) is slower now. maybe just an impression? but anyway, is it possible to enhance this, is the gui
work related with that?Do you have an example patch?
I have a complicated patch with some structures in it, I can send to
it you later. but I guess any patch will do - including your solitude.
Ideally there would be a patch that clearly illustrates the problem
and nothing else.
- I use the [hcs/sys_gui] command to make my own color scheme
(white is too agressive) and to place the pd window on a corner (wm
geometry). These don't work now, are they going to be obsolete, or they have to
adapt to the new tcl code?They should work fine, but the color scheme variables are part of
Pd-extended, this is still Pd-vanilla. I think you'd be better off
porting your color scheme to a Tcl plugin. Its not hard, plus you
can do a lot more with it.sorry, but I can't understand what "you can drop them into pd/ startup" means. I see no folder or file with that name here.
Its not getting installed yet on Windows. You can create 'startup' in
the folder where you installed Pd 0.43/gui-rewrite (\Program Files\pd
by default). Then you can drop and *-plugin.tcl file there to have it
loaded when Pd runs. Here are some of the plugins available:
http://pure-data.svn.sourceforge.net/viewvc/pure-data/branches/pd-gui-rewrit...
.hc
- x and y, x only + y only have all the same result. but I think
that was the normal behaviour before anyway (never used this feature)
I've never used it either, I think it might be really old cruft...
I would ask the higher powers (miller/iem people) and get rid of it.
- Help->About: Pd doesn't work (so I can't tell exactly which
version it is)
Yeah... not implemented yet... it tells you the version at startup
in the Pd window.Pd version 0.43-0devel-20091008, in case it's important to know.
Anything else specific you want to be checked, Hans?
It would be great if you can use it for daily use, and report
issues. I'm starting a project where I will be doing the same
thing. That's where we will really discover bugs and issues.should be doing that from tomorrow, maybe. doesn't the ubuntu
version comes with the pd-ext material, like the windows one?
"Free software means you control what your computer does. Non-free
software means someone else controls that, and to some extent controls
you." - Richard M. Stallman
-
Hans-Christoph Steiner -
João Pais