Hi!
I just joined the mailinglist a few days ago. There was a mail from someone (sorry I accidentally deletetd it) complaining about the lack of Linux squeak to produce sound. I would like to take the opportunity to point out some other shortcomings of Linux squeak.
The one thing, that really annoys me is that squeak won't interrupt on CTRL-C. If I get stuck in an endless loop I have to kill the squeak process, restart squeak and replay the changes.
Mentioning starting squeak reminds me that I have tried to start squeak with the -xshm option to enable X shared memory extensions. If I try to resize the squeak work window it crashes. This won't happen when I start squeak without -xshm.
What confuses me is that some of the sourcecode in the image uses CR as the lineend character. Everytime I want to have a look at a method which has this line terminator I just get garbage in the method view. I have to do a format and accept to get it straight. Shouldn't the display method treat LF and CR equally?
Other thing I noticed, is that output to Transcript is very slow. Something like <? 80 timesRepeat: [Transcript show: '#'] ?> really gives you a coffee-break. I don't know if this is just a linux specific problem.
After all that bashing I want to make clear, that squeak has some aspects I really like. Just to pick one: PWS is a very ... "cool" thing. :)
I appreciate the rapid development of squeak and that it offers a playground for new ideas. This made me return to it. (After first contact with squeak about one year ago, my first reaction was to set up "The official Squeak haters page" at our university. -- Things can change :)
Alex.
For infinite loops problems, you can use alt-. (which also works on NT and perhaps every platform?)
For line-end problems, you can load CRLFFileStream.cs from:
http://www.cc.gatech.edu/fac/mark.guzdial/squeak/CrLfFileStream.cs
(although maybe there is a more official location). After loading it, change the method FileStream class>>concreteStream to say CrLfFileStream instead of StandardFileStream.
Regarding sound, at least three people have put together sound support for Linux. I'm one of those, and have put my stuff at:
http://www.cc.gatech.edu/~lex/squeak/sound/
It seems to work fine for me, though I don't exactly test it real hard. I don't know if the other two (or more!) have put their stuff on a Web page or not; my stuff has the potential disadvantage of requiring pthreads for recording to work.
And finally, I don't have any suggesting about crashing after window resizing. IMHO, though, little quirks like this are well worth the benefits of having the latest code available. If you want more safety, you can always turn off the really new features like xshm and jitter.
Happy Squeaking (and Linuxing)!
Lex
Alexander Lazarevic writes:
Hi!
I just joined the mailinglist a few days ago. There was a mail from someone (sorry I accidentally deletetd it) complaining about the lack of Linux squeak to produce sound. I would like to take the opportunity to point out some other shortcomings of Linux squeak.
The one thing, that really annoys me is that squeak won't interrupt on CTRL-C. If I get stuck in an endless loop I have to kill the squeak process, restart squeak and replay the changes.
Mentioning starting squeak reminds me that I have tried to start squeak with the -xshm option to enable X shared memory extensions. If I try to resize the squeak work window it crashes. This won't happen when I start squeak without -xshm.
What confuses me is that some of the sourcecode in the image uses CR as the lineend character. Everytime I want to have a look at a method which has this line terminator I just get garbage in the method view. I have to do a format and accept to get it straight. Shouldn't the display method treat LF and CR equally?
Other thing I noticed, is that output to Transcript is very slow. Something like <? 80 timesRepeat: [Transcript show: '#'] ?> really gives you a coffee-break. I don't know if this is just a linux specific problem.
After all that bashing I want to make clear, that squeak has some aspects I really like. Just to pick one: PWS is a very ... "cool" thing. :)
I appreciate the rapid development of squeak and that it offers a playground for new ideas. This made me return to it. (After first contact with squeak about one year ago, my first reaction was to set up "The official Squeak haters page" at our university. -- Things can change :)
Alex.
-- ------ Alexander Lazarevic ------------------------------------------------ -- Otto-von-Guericke-Universitaet mailto: lazarevi@cs.uni-magdeburg.de -- -- Computational Visualistics http://www.cs.uni-magdeburg.de/~lazarevi --
Sorry for the clutter. Read Andreas Raab message an Alt-. (period) works!! Thanks Andreas.
Phil Weichert wrote:
I am experiencing a similiar problem running under Windows 95. Neither ctrl-c alt- cause an interrupt. I am running OSR 2.1. Phil
Lex Spoon wrote:
For infinite loops problems, you can use alt-. (which also works on NT and perhaps every platform?)
Alex,
The one thing, that really annoys me is that squeak won't interrupt on CTRL-C. If I get stuck in an endless loop I have to kill the squeak process, restart squeak and replay the changes.
Use Alt-. (period) or redefine it by using InputSensor>>setInterruptKey: e.g.,
Sensor setInterruptKey: 16r263
should give you the desired Ctrl-C interrupt key.
Mentioning starting squeak reminds me that I have tried to start squeak with the -xshm option to enable X shared memory extensions. If I try to resize the squeak work window it crashes. This won't happen when I start squeak without -xshm.
Well ... this sounds like if there is some bounds check wrong.
[Note to Ian: you might check if you handle cases correctly, when Squeak's display form is smaller than the window - at one point this crashed the Windows VM]
What confuses me is that some of the sourcecode in the image uses CR as the lineend character.
Actually *all* sources in the image should have CR line endings ;-)
Everytime I want to have a look at a method which has this line terminator I just get garbage in the method view. I have to do a format and accept to get it straight. Shouldn't the display method treat LF and CR equally?
There is a fix in STP's set of fileIns (at ftp.create.ucsb.edu).
Bye, Andreas
Hi Andreas!
The one thing, that really annoys me is that squeak won't interrupt on CTRL-C. If I get stuck in an endless loop I have to kill the squeak process, restart squeak and replay the changes.
Use Alt-. (period) or redefine it by using InputSensor>>setInterruptKey: e.g.,
Sensor setInterruptKey: 16r263
should give you the desired Ctrl-C interrupt key.
Which it doesn't, but thanks for the pointer to InputSensor. I will try to figure out which value needs to be set here to get CTRL-C working. My first naive attempt was to do a
Sensor setInterruptKey: (TextConstants at: #CtrlC)
which didn't work either. I guess this is because of the difference of charctercodes and keycodes, isn't it?
Alex.
squeak-dev@lists.squeakfoundation.org