Patches item #1670440, was opened at 2007-02-27 16:49
Message generated for change (Comment added) made by millerpuckette
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1670440&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Ivica Ico Bukvic (ico_bukvic)
Assigned to: Miller Puckette (millerpuckette)
Summary: 8.5 tcl help browser bug
Initial Comment:
This help browser bug using CVS and tcl/tk 8.5 has been present since at least last fall. After digging through the code, I found the following:
proc doc_make_listbox {base dir count} {
# check for [file readable]?
if { [info tclversion] >= 8.5 } {
# requires Tcl 8.5 but probably deals with special chars better
# destroy {expand}[lrange [winfo children $base] [expr {2 * $count}] end]
} else {
The comment before "destroy" in effect disables destroying of children widgets and causes errors whenever I have more than one column open and try to click on a column preceding one that is already created. The error reported is that the "listbox1-list already exists" and the only way around is to close the help browser and reopen it. This obviously prevents going up the file tree once a particular subfolder has been opened which obviously means it is a non-lethal bug, but nonetheless annoying.
Thus, the "else" clause only affects non 8.5 versions and as such causes pd with tcl/tk 8.5 to be broken.
I just checked this out and the 8.4 syntax does work on 8.5 but in the current code it is never accessed due to the existing "if" statement. The new fix simply comments the "if" and "else" statements making the 8.4 style of closing children widgets mandatory for all versions, and since it appears that 8.4 code works fine with tcl/tk 8.5, this appears to be a universal fix for this issue.
Please see attached (diff -uw against the fresh CVS checkout).
----------------------------------------------------------------------
>Comment By: Miller Puckette (millerpuckette)
Date: 2007-12-18 14:34
Message:
Logged In: YES
user_id=313747
Originator: NO
I think this one got silently adopted earlier.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1670440&group_…
Bugs item #1359216, was opened at 2005-11-17 19:14
Message generated for change (Comment added) made by claudiusmaximus
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1359216&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: None
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: printing open bracket { (ASCII123) freezes gui
Initial Comment:
On WinXP and Miller's pd039_test7 typing '{' (ascii
123) after print in an object box hangs the gui,
apparently because the tcl is waiting for a closing '}'
before proceeding.
In the source code, dopost() in s_print.c, it looks
like four characters are escaped: '\', '{', '}', and ';'.
But if they are escaped they should not freeze the gui...
static void dopost(const char *s)
{
...
if (c == '\\' || c == '{' || c == '}' || c
== ';') <- if it's a dangerous character for tcl
upbuf[ptout++] = '\\'; <- add a
backslash before the dangerous character
upbuf[ptout] = s[ptin]; <- then put the
character
...
sys_vgui("pdtk_post {%s}\n", upbuf); <- then
send it to tcl interpreter
If I try entering them in a print box I get different
results. '\' and '}' are rejected:
\}: dropped
\\: dropped
';' is accepted
'{' freezes the gui.
In g_editor.c:
void canvas_key(t_canvas *x, t_symbol *s, int ac,
t_atom *av)
{
...
if (keynum == '\\' || keynum == '{' || keynum ==
'}') <-a dangerous character
{
post("%c: dropped", (int)keynum); <-should be
dropped, but for '{' this doesn't get printed
return;
}
...dopost() is called from post() so '{' is causing
trouble even when it is escaped with a backslash.
Submitted by Martin Peach
martinrp(a)vax2.condordia.ca
----------------------------------------------------------------------
Comment By: ClaudiusMaximus (claudiusmaximus)
Date: 2007-12-18 16:10
Message:
Logged In: YES
user_id=769033
Originator: NO
This bug is still present in pd-0.40-3.
I noticed it when pdlua tried to print a "syntax error" message containing
Lua source code on a line containing '{'.
----------------------------------------------------------------------
Comment By: ClaudiusMaximus (claudiusmaximus)
Date: 2006-05-02 11:17
Message:
Logged In: YES
user_id=769033
In pd-0.39-2, typing { in a new object box hangs pd-gui
until you type a corresponding }, at which point the
following is printed in the console:
\{: dropped
\{: dropped
\}: dropped
\}: dropped
The text typed (exclusively) between { and } is entered into
the object box when the } is typed, too.
----------------------------------------------------------------------
Comment By: Thomas Grill (xovo)
Date: 2005-12-07 01:49
Message:
Logged In: YES
user_id=350252
second:
it's pretty clear that the bug must be in the socket
receiver code for pd.tk inside t_tkcmd.c. In pd_readsocket
starting with line 168 it says:
/* search for locations that terminate a complete TK
command. These are carriage returns which are not inside
any braces. Braces can be escaped with backslashes (but
backslashes themselves can't.) */
When a "{: dropped" post comes in, the code in pd_readsocket
must get confused because a brace is opened but none closed.
There's no command-delimiting newline character found in
that case.
----------------------------------------------------------------------
Comment By: Thomas Grill (xovo)
Date: 2005-12-07 01:08
Message:
Logged In: YES
user_id=350252
first, Matju 's fixes for the dopost function in s_print.c
static void dopost(const char *s)
{
if (sys_printhook)
(*sys_printhook)(s);
else if (sys_printtostderr)
fprintf(stderr, "%s", s);
else
{
char upbuf[MAXPDSTRING];
int i,j=0;
for(i=0; s[i] && j<MAXPDSTRING-16; i++)
{
if (strchr("\\\"[]$\n",s[i])) upbuf[j++]='\\';
if (s[i]=='\n') upbuf[j++]='n';
else upbuf[j++] = s[i];
}
upbuf[j] = 0;
sys_vgui("pdtk_post \"%s\"\n",upbuf);
}
}
----------------------------------------------------------------------
Comment By: Mathieu Bouchard (matju)
Date: 2005-12-03 06:57
Message:
Logged In: YES
user_id=801174
fixed in DesireData today.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1359216&group_…
Bugs item #1853216, was opened at 2007-12-18 16:06
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1853216&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: v0.40.2
Status: Open
Resolution: None
Priority: 5
Private: No
Submitted By: ClaudiusMaximus (claudiusmaximus)
Assigned to: Nobody/Anonymous (nobody)
Summary: class_set_extern_dir not exported (breaks loaders)
Initial Comment:
The problem:
Help patches for externals loaded by the loader functionality of Pd are not found.
They can be found if class_set_extern_dir() is used, but this doesn't work on all platforms (unresolved symbol errors when trying to load the loader).
The solution:
Declare class_set_extern_dir() in m_pd.h and export it from pd/src/m_class.c or pd/src/s_loader.c .
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478070&aid=1853216&group_…
Patches item #1551815, was opened at 2006-09-04 10:05
Message generated for change (Settings changed) made by zmoelnig
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1551815&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
>Status: Closed
>Resolution: Accepted
Priority: 5
Private: No
Submitted By: oskude (oskude)
Assigned to: Miller Puckette (millerpuckette)
Summary: align control checkboxes to left
Initial Comment:
"-anchor w" in checkbutton doesnt have any effect.
moved to pack, so they aligh to the left...
----------------------------------------------------------------------
Comment By: Miller Puckette (millerpuckette)
Date: 2007-12-18 05:18
Message:
Logged In: YES
user_id=313747
Originator: NO
took it.
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-09-16 19:04
Message:
Logged In: YES
user_id=27104
Nice little detail. I added to the Pd-extended build system
as
pure-data/packages/patches/pd_controls-anchor-fix-0.40-pre.patch
It will be in the nightly builds starting tomorrow.
----------------------------------------------------------------------
Comment By: oskude (oskude)
Date: 2006-09-15 10:06
Message:
Logged In: YES
user_id=1383707
[snip]
checkbutton .controls.switches.audiobutton -text {compute
audio} \
-variable ctrls_audio_on \
-anchor w \
-command {pd [concat pd dsp $ctrls_audio_on \;]}
checkbutton .controls.switches.meterbutton -text {peak meters} \
-variable ctrls_meter_on \
-anchor w \
-command {pd [concat pd meters $ctrls_meter_on \;]}
pack .controls.switches.audiobutton
.controls.switches.meterbutton -side top
[snip]
looks like this:
[ ] compute audio
[ ] peak meters
as you see, -anchor -w doesnt have any effect when used in
checkbutton, they are not aligned to left(w:west)
so changing to this:
[snip]
checkbutton .controls.switches.audiobutton -text {compute
audio} \
-variable ctrls_audio_on \
-command {pd [concat pd dsp $ctrls_audio_on \;]}
checkbutton .controls.switches.meterbutton -text {peak meters} \
-variable ctrls_meter_on \
-command {pd [concat pd meters $ctrls_meter_on \;]}
pack .controls.switches.audiobutton
.controls.switches.meterbutton -side top \
-anchor w
[snip]
looks like this:
[ ] compute audio
[ ] peak meters
using -anchor w in pack aligns the checkbuttons to left...
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-09-13 21:25
Message:
Logged In: YES
user_id=27104
can you elaborate on what this patch does?
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1551815&group_…
Patches item #1549377, was opened at 2006-08-30 17:39
Message generated for change (Settings changed) made by zmoelnig
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
>Status: Closed
>Resolution: Accepted
Priority: 9
Private: No
Submitted By: IOhannes m zmölnig (zmoelnig)
Assigned to: Miller Puckette (millerpuckette)
Summary: fix crash on sending empty "obj" mesage to closed canvas
Initial Comment:
as described in bug #1525832, pd crashes when you send
a "msg" or "obj" message (without coordinates and
arguments) to a canvas that is not visible (e.g. closed).
this is due to the fact, that inthis case, pd
automatically goes into edit mode, so you can type in
the content of the object/message (it's like creating
an empty object-box with Ctrl-1).
this is a not so clever thing, when the canvas is closed.
attached is a patch that fixes this, by checking
whether the canvas is visible and refusing to continue
otherwise
----------------------------------------------------------------------
Comment By: Miller Puckette (millerpuckette)
Date: 2007-12-18 05:07
Message:
Logged In: YES
user_id=313747
Originator: NO
took it.
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2007-12-17 21:48
Message:
Logged In: YES
user_id=564396
Originator: YES
this still crashes
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2006-10-10 15:28
Message:
Logged In: YES
user_id=564396
raised priority, since a crasher should be fixed.
added a patch to reliably reproduce the bug.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Patches item #1551815, was opened at 2006-09-04 01:05
Message generated for change (Comment added) made by millerpuckette
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1551815&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
Status: Open
Resolution: Works For Me
Priority: 5
Private: No
Submitted By: oskude (oskude)
Assigned to: Miller Puckette (millerpuckette)
Summary: align control checkboxes to left
Initial Comment:
"-anchor w" in checkbutton doesnt have any effect.
moved to pack, so they aligh to the left...
----------------------------------------------------------------------
>Comment By: Miller Puckette (millerpuckette)
Date: 2007-12-17 20:18
Message:
Logged In: YES
user_id=313747
Originator: NO
took it.
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-09-16 10:04
Message:
Logged In: YES
user_id=27104
Nice little detail. I added to the Pd-extended build system
as
pure-data/packages/patches/pd_controls-anchor-fix-0.40-pre.patch
It will be in the nightly builds starting tomorrow.
----------------------------------------------------------------------
Comment By: oskude (oskude)
Date: 2006-09-15 01:06
Message:
Logged In: YES
user_id=1383707
[snip]
checkbutton .controls.switches.audiobutton -text {compute
audio} \
-variable ctrls_audio_on \
-anchor w \
-command {pd [concat pd dsp $ctrls_audio_on \;]}
checkbutton .controls.switches.meterbutton -text {peak meters} \
-variable ctrls_meter_on \
-anchor w \
-command {pd [concat pd meters $ctrls_meter_on \;]}
pack .controls.switches.audiobutton
.controls.switches.meterbutton -side top
[snip]
looks like this:
[ ] compute audio
[ ] peak meters
as you see, -anchor -w doesnt have any effect when used in
checkbutton, they are not aligned to left(w:west)
so changing to this:
[snip]
checkbutton .controls.switches.audiobutton -text {compute
audio} \
-variable ctrls_audio_on \
-command {pd [concat pd dsp $ctrls_audio_on \;]}
checkbutton .controls.switches.meterbutton -text {peak meters} \
-variable ctrls_meter_on \
-command {pd [concat pd meters $ctrls_meter_on \;]}
pack .controls.switches.audiobutton
.controls.switches.meterbutton -side top \
-anchor w
[snip]
looks like this:
[ ] compute audio
[ ] peak meters
using -anchor w in pack aligns the checkbuttons to left...
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-09-13 12:25
Message:
Logged In: YES
user_id=27104
can you elaborate on what this patch does?
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1551815&group_…
Patches item #1549377, was opened at 2006-08-30 08:39
Message generated for change (Comment added) made by millerpuckette
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
Status: Open
Resolution: None
Priority: 9
Private: No
Submitted By: IOhannes m zmölnig (zmoelnig)
Assigned to: Miller Puckette (millerpuckette)
Summary: fix crash on sending empty "obj" mesage to closed canvas
Initial Comment:
as described in bug #1525832, pd crashes when you send
a "msg" or "obj" message (without coordinates and
arguments) to a canvas that is not visible (e.g. closed).
this is due to the fact, that inthis case, pd
automatically goes into edit mode, so you can type in
the content of the object/message (it's like creating
an empty object-box with Ctrl-1).
this is a not so clever thing, when the canvas is closed.
attached is a patch that fixes this, by checking
whether the canvas is visible and refusing to continue
otherwise
----------------------------------------------------------------------
>Comment By: Miller Puckette (millerpuckette)
Date: 2007-12-17 20:07
Message:
Logged In: YES
user_id=313747
Originator: NO
took it.
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2007-12-17 12:48
Message:
Logged In: YES
user_id=564396
Originator: YES
this still crashes
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2006-10-10 06:28
Message:
Logged In: YES
user_id=564396
raised priority, since a crasher should be fixed.
added a patch to reliably reproduce the bug.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Patches item #1094912, was opened at 2005-01-03 11:56
Message generated for change (Comment added) made by zmoelnig
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1094912&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: wishlist
>Status: Pending
>Resolution: Wont Fix
Priority: 6
Private: No
Submitted By: Tim Blechmann (timblech)
Assigned to: Miller Puckette (millerpuckette)
Summary: message-based access to the audio api
Initial Comment:
attached is a patch that simplifies message-based
access to the audio api:
pd understands the following messages:
audio-samplerate
audio-delay
audio-dacblocksize
audio-scheduler
audio-device
audio-device-in
audio-device-out
audio-dacblocksize and audio-scheduler require devel_0_38
----------------------------------------------------------------------
>Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2007-12-17 21:53
Message:
Logged In: YES
user_id=564396
Originator: NO
i guess this will never get fixed.
and it is most likely a better idea to create an (external) object that
interfaces with the the audioapi.
e.g.
this would set the samplerate to 48000
[samplerate 44000(
|
[audioapi]
this would query the current samplerate:
[samplerate(
|
[audioapi]
|
[44100\
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2006-10-18 09:58
Message:
Logged In: YES
user_id=564396
as carmen has pointed out on the list, miller had just very
recently articulated a very similar idea.
even though it is unlikely that he will use the patch here,
i reopen this issue (and hopefully will close it again soon
when an equivalent mechanism will be part of pd)
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-10-17 07:41
Message:
Logged In: YES
user_id=27104
this has been here for a long while with no activity, so I
am setting it to Pending. It'll be automatically set to
closed after a while if no one tends to it.
----------------------------------------------------------------------
Comment By: Hans-Christoph Steiner (eighthave)
Date: 2006-01-29 03:34
Message:
Logged In: YES
user_id=27104
I am liking the idea of sending messages to pd to get all of
this kind of information more and more. I am thinking that
a generic "get" message would be useful here. Like this:
[;pd get audio-samplerate(
[;pd get audio-delay(
[;pd get audio-dacblocksize(
[;pd get audio-scheduler(
[;pd get audio-device(
[;pd get audio-device-in(
[;pd get audio-device-out(
Then also:
[;pd get path(
[;pd get libs(
[;pd get version(
[;pd get dsp(
Then you could get similar messages from each canvas/patch:
[;pd-my_patch.pd get editmode(
[;pd-my_patch.pd get vis(
[;pd-my_patch.pd get namespace(
[;pd-my_patch.pd get canvasname(
You would retrieve these messages with a [receive] of the
same name, i.e.:
[r pd]
|
[route version]
|
----------------------------------------------------------------------
Comment By: Miller Puckette (millerpuckette)
Date: 2005-05-18 06:36
Message:
Logged In: YES
user_id=313747
I'm scared of this one... how important is it to have all the
different audio parameters handled by separate messages? I
think the "audio_dialog" message should suffice for setting
audio parameters...
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1094912&group_…
Patches item #1549377, was opened at 2006-08-30 17:39
Message generated for change (Comment added) made by zmoelnig
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: bugfix
Status: Open
Resolution: None
>Priority: 9
Private: No
Submitted By: IOhannes m zmölnig (zmoelnig)
Assigned to: Miller Puckette (millerpuckette)
Summary: fix crash on sending empty "obj" mesage to closed canvas
Initial Comment:
as described in bug #1525832, pd crashes when you send
a "msg" or "obj" message (without coordinates and
arguments) to a canvas that is not visible (e.g. closed).
this is due to the fact, that inthis case, pd
automatically goes into edit mode, so you can type in
the content of the object/message (it's like creating
an empty object-box with Ctrl-1).
this is a not so clever thing, when the canvas is closed.
attached is a patch that fixes this, by checking
whether the canvas is visible and refusing to continue
otherwise
----------------------------------------------------------------------
>Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2007-12-17 21:48
Message:
Logged In: YES
user_id=564396
Originator: YES
this still crashes
----------------------------------------------------------------------
Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2006-10-10 15:28
Message:
Logged In: YES
user_id=564396
raised priority, since a crasher should be fixed.
added a patch to reliably reproduce the bug.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1549377&group_…
Patches item #1528580, was opened at 2006-07-25 21:19
Message generated for change (Comment added) made by zmoelnig
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1528580&group_…
Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: puredata
Group: None
>Status: Pending
>Resolution: Fixed
Priority: 5
Private: No
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: starting with "-jack" alone does not create ports
Initial Comment:
Starting pd 0.39-2 with the "-jack" flag alone does not create audio i/o ports at startup. This behavior is different from, for example, starting it with "-alsa". The following patch make jack behave as alsa:
------------------------------
--- pd-0.39-2/src/s_audio.c.orig 2005-09-28 13:29:44.000000000 -0700
+++ pd-0.39-2/src/s_audio.c 2006-07-23 16:05:08.709122608 -0700
@@ -321,10 +321,16 @@
else
#endif
#ifdef USEAPI_JACK
- if (sys_audioapi == API_JACK)
+ if (sys_audioapi == API_JACK) {
+ if (audio_naudioindev == -1 && audio_naudiooutdev == -1) {
+ /* if nothing specified then open the default */
+ audio_naudioindev = audio_naudiooutdev = 1;
+ audio_audioindev[0] = audio_audiooutdev[0] = DEFAULTAUDIODEV;
+ audio_audiochindev[0] = audio_audiochoutdev[0] = SYS_DEFAULTCH;
+ }
jack_open_audio((naudioindev > 0 ? realinchans[0] : 0),
(naudiooutdev > 0 ? realoutchans[0] : 0), rate);
-
+ }
else
#endif
#ifdef USEAPI_OSS
------------------------------
-- Fernando
----------------------------------------------------------------------
>Comment By: IOhannes m zmölnig (zmoelnig)
Date: 2007-12-17 21:37
Message:
Logged In: YES
user_id=564396
Originator: NO
at least with 0.41 this seems to be fixed.
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=478072&aid=1528580&group_…