So one of the things that Miller said that is necessary before putting the pd sources into CVS is a suite of regression test. Since there is already development going on, I think that we should start developing regression tests. I think this could be done by creating patches that do a lot of things, then output a text result. Then a Makefile could run them using pd -nogui to test them.
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them. Also bangs stop displaying after displaying once. And the second problem is demonstrated by maxlib's [velocity]. When I use it with the CVS pd, the only thing it outputs is "nan". When I reverted to the "pd" package using apt-get, all of the above problems disappeared. When I get more time, I can do a thorough bug report if anyone is interested.
.hc
zen \ \ \[D[D[D[D
Hallo, Hans-Christoph Steiner hat gesagt: // Hans-Christoph Steiner wrote:
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them.
Actually I also visited this problem with an earlier version, I guess one of the 0.36-test ones, then it went away, and with current cvs it is there again...
Strange.
ciao
On Wed, 30 Apr 2003, Frank Barknecht wrote:
Hallo, Hans-Christoph Steiner hat gesagt: // Hans-Christoph Steiner wrote:
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them.
Actually I also visited this problem with an earlier version, I guess one of the 0.36-test ones, then it went away, and with current cvs it is there again...
Strange.
Its working well here .. wondering. Which sound driver are you using ? I am trying to reproduce.
Guenter
On Wed, 30 Apr 2003, guenter geiger wrote:
On Wed, 30 Apr 2003, Frank Barknecht wrote:
Hallo, Hans-Christoph Steiner hat gesagt: // Hans-Christoph Steiner wrote:
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them.
Actually I also visited this problem with an earlier version, I guess one of the 0.36-test ones, then it went away, and with current cvs it is there again...
Strange.
Its working well here .. wondering. Which sound driver are you using ? I am trying to reproduce.
I am using alsa-source-0.9+0beta12-3 compiled with a 2.4.21-pre2 kernel with force-feedback joystick patches. I am currently using an ens1371 soundcard. Here are some other things that may be of interest:
ii tcl8.3 8.3.3-7 The Tool Command Language ii tk8.3 8.3.3-8 ii xserver-xfree86 4.2.1-3
root@winky:~>gcc --version 2.95.4
.hc
zen \ \ \[D[D[D[D
On Tue, 29 Apr 2003, Hans-Christoph Steiner wrote:
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them. Also bangs stop displaying after displaying once. And the second problem is demonstrated by maxlib's [velocity]. When I use it with the CVS pd, the only thing it outputs is "nan". When I reverted to the "pd" package using apt-get, all of the above problems disappeared. When I get more time, I can do a thorough bug report if anyone is interested.
Strange, which debian version are you using ? (actually sid comes with the CVS pd version because of the jack support, same with the Planet CCRMA package).
Either the problem is specific to your system or the bug has been introduced recently.
regarding 2)
Yes regression tests would be a good thing, but it is unlikely that you can write a regression test that catches the bugs you mentioned.
Anyhow, the first big task that has to be done in CVS now is to merge with 0.37.
Greetings,
Guenter
On Wed, 30 Apr 2003, guenter geiger wrote:
On Tue, 29 Apr 2003, Hans-Christoph Steiner wrote:
I am motivated by the fact that pd devel_0_36 from CVS has some key bugs in it on my Linux box. There are two clear demonstrations of the problems: the first is that the bold border around a message doesn't go away after I click on the message. All of the message I click get stuck with this bold border around them. Also bangs stop displaying after displaying once. And the second problem is demonstrated by maxlib's [velocity]. When I use it with the CVS pd, the only thing it outputs is "nan". When I reverted to the "pd" package using apt-get, all of the above problems disappeared. When I get more time, I can do a thorough bug report if anyone is interested.
Strange, which debian version are you using ? (actually sid comes with the CVS pd version because of the jack support, same with the Planet CCRMA package).
Either the problem is specific to your system or the bug has been introduced recently.
regarding 2)
Yes regression tests would be a good thing, but it is unlikely that you can write a regression test that catches the bugs you mentioned.
Anyhow, the first big task that has to be done in CVS now is to merge with 0.37.
I'll wait until 0.37 is merged in, then I start this process again. 0.37 might just fix my issues. I'll write a little -nogui test patch to highlight the [velocity]/"nan" problem then too.
.hc
zen \ \ \[D[D[D[D
-
Frank Barknecht -
guenter geiger -
Hans-Christoph Steiner