hi krzysztof, hi devs ...
during a performance this evening i found an incompatibility of toxy and 0.38:
when opening a patch that has visible subpatches, only the subpatch will be displayed. this happened after i removed the menu from my other patches...
also i realized that in one of the windows, some gui elements of a patch froze (at a few number boxes in my case) ... at that time, i already removed the menu with toxy ... all the numberboxes where working for some time ...
system: devel_0_38, toxy (well compiled for devel_0_37, maybe that's the problem)
anyway, if you have an idea, what might cause or even better solve this behaviour, please let me know ...
cheers ... tim
hi Tim,
sorry it has happened during a show. Using Miller's 0.38test10, I could not reproduce that behaviour...
Perhaps, your Pd version sprintfs gui names the %p way (former %x, now it is %lx in test10)? If not, would be nice to see an example patch...
Krzysztof
Tim Blechmann wrote: ...
during a performance this evening i found an incompatibility of toxy and 0.38:
Perhaps, your Pd version sprintfs gui names the %p way (former %x, now it is %lx in test10)? If not, would be nice to see an
that's not the problem ... i'm using %lx in devel_0_38 ...
example patch...
i attached one for the subpatch opening ... (patch.pd -> opens patch2.pd with message to pd) error message: invalid command name ".x8368430.m.windows"
it's more difficult for frozen gui elements (since they worked for about one hour) ... i couldn't reproduce their behaviour, yet ...
cheers ... tim
I think, the menu of a *main* patch should never be destroyed, without updating the $menu_windowlist variable. Same in .37, but .38 seems to be less forgiving.
hm ... i'm not really familiar with the tcl/tk code ... is it possible to do that from toxy or will i have to do that in pd itself?
cheers ... tim
hi Tim,
the easiest short-term solution would be commenting out the line
menu_fixwindowmenu [lindex $i 1]
in proc pdtk_fixwindowmenu of pd.tk. This line's task is duplicated with canvas menu -postcommand, anyway, so it is not necessary.
If you do not want to mess with pd.tk, the pdtk_fixwindowmenu may be redefined using tot (still a short-term solution).
A permanent solution: another message to canvas ('hidemenu'?).
Krzysztof
Tim Blechmann wrote:
I think, the menu of a *main* patch should never be destroyed, without updating the $menu_windowlist variable. Same in .37, but .38 seems to be less forgiving.
hm ... i'm not really familiar with the tcl/tk code ... is it possible to do that from toxy or will i have to do that in pd itself?
the easiest short-term solution would be commenting out the line
menu_fixwindowmenu [lindex $i 1]
in proc pdtk_fixwindowmenu of pd.tk. This line's task is duplicated with canvas menu -postcommand, anyway, so it is not necessary.
If you do not want to mess with pd.tk, the pdtk_fixwindowmenu may be redefined using tot (still a short-term solution).
thanks ... i'll try that ...
A permanent solution: another message to canvas ('hidemenu'?).
that sounds pretty good, but i won't have the time to look into it for the next few weeks ...
cheers ... tim