Squeak-dev June 1999

squeak-dev@lists.squeakfoundation.org

139 participants
408 discussions

new display prims 91 & 92
by Tim Rowledge 03 Jun '99

03 Jun '99

After updating from the server last night, I noticed tweo new prims relating to checking and setting the display parameters have been added. Prim 92 appears not to be actually used. Prim 91 requires a new plat specific routine to be added. Is prim 92 actually going to be used in the future? What are the expected failure conditions? When new requirements are added to the support needed from platforms it would be most useful to get some kind of warning so that those of us who have to do the work outside disney can at least understand what is going on. -- Fractured Idiom:- ALOHA OY - Love; greetings; farewell; from such a pain you should never know Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim>

3 2

Re: Newbies Wanted
by Edward P Luwish 03 Jun '99

03 Jun '99

The new content looks very good, but my obsessive-compulsive side is going nuts over the misspelling of the page title (which, as an HTML title, was duplicated in the name of my bookmark) - It should be TutoRial, not TutoTial, although maybe it's an improvement? :-) Ed "Ted K." wrote: > Folks, > I've responded to your requests on the BankAccount tutorial. Curly quotes have been removed, and two pictures of the windows have been added. Thank you Steve Wessels, Russ Van Rooy, Tim Olsen, Paul Bernard, Andreas Raab, and Bert Freudenberg for all of the suggestions. (I'm the tutotial editor and John is the author. The tutorial was transcribed from a class John gave.) > > http://minnow.cc.gatech.edu/SqueakDoc.42 > > --Ted. > > Ted Kaehler, Walt Disney Imagineering, R&D > (home) 3415 Cork Oak Way, Palo Alto, CA 94303. voice (650) 424-1070 > http://www.webPage.com/~kaehler2/ > Remember, we are the lowest form of intelligence that can do this sort of thing.

2 1

RE: new display prims 91 & 92
by Tim Rowledge 03 Jun '99

03 Jun '99

On Wed 02 Jun, Raab, Andreas wrote: > here's a brief description: [SNIP] > Considering your question about some kind of warning what warning would you > expect?! Pretty much the info you've just given. Platform affecting VM changes are very disconcerting when they just appear - and as we've successfully demonstrated in the past, early discussions can help make them both more useful and better implemented. tim -- Useful random insult:- The wheel's spinning but the hamster's dead. Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim>

1 0

[BUG][FIX] FileStream naming problems
by Tim Rowledge 03 Jun '99

03 Jun '99

When translating a plugin or the vm, the code has recently been changed to use a CrLfFileStream which resulted in my files going to all sorts of odd places. It turns out that by refering directly to CrLfFileStream we miss out the filename canonicalisation that was done when FileStream was used as previously. This is actually an endemic problem when one has a hierarchy where subclasses that are originally intended to be accessed indirectly get used directly. The 'root' class has code to keep everything tidy and then some subclass is used in what seems a reasonable manner and the tidy-code is missed out because of the method naming 'clash'. To fix the problem in this case, I was able to rename the filestream creation methods and redirect them a little more robustly. CrLfFileStream (or indeed HTMLFileStream) can now be used without the name-fixing code being missed. I'm sure there are better ways to do this in the long term, but this should do for now. tim -- I'd kill for a Nobel Peace Prize. Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim> Content-Type: APPLICATION/Octet-stream; NAME="FilestreamFixes/2June321pm.cs" Attachment converted: Anon:FilestreamFixes/2June321pm.cs (????/----) (0000A935)

1 0

Squeak sighting in Dr Dobbs
by Michael S. Klein 02 Jun '99

02 Jun '99

Paul Tremblett http://www.softwaresummit.com/1999/speakers/tremblett.htm wrote an article in July 1999 Dr. Dobb's Journal, using Squeak (and Java). Article (with some squeak screen shots) is at: http://www.ddj.com/articles/1999/9907/9907c/9907c.htm Unfortunately, I can't find X509.st, I looked at the ftp site (ftp.ddj.com) but could only find X509.txt). Anybody (Paul?) know where it is? -- Mike mklein(a)alumni.caltech.edu

2 1

Debugger bugged ?
by Pierre Roy 02 Jun '99

02 Jun '99

I have lots of "simulated message XXX not understood", especially with XXX = '<' when I use the debugger (either with the step or send commands). Does anyone else have the same problem ?

1 0

[Squeak] <-- Could we add one of these?
by Nathan Wallace 02 Jun '99

02 Jun '99

I'd like to see all messages to the squeak list prefixed with [Squeak]. My inbox fills up very quickly with stuff from mailing lists, these prefixes are common and very helpful. Can we add it? Thanks, Nathan -- Nathan Wallace http://e-gineer.com

2 1

Re: Process priority and portable OS threads
by Bob Arning 02 Jun '99

02 Jun '99

On Wed, 2 Jun 1999 07:06:32 -0400 "Stephen Pair" <spair(a)advantive.com> wrote: >--- process priority --- > >In Squeak (under Windows), when I fork a process at user background >priority, it seems to never run (or at least it takes a very long time). Is >there a way to have user background priority processes to run in a more >timely fashion. I was creating about 100 processes to test the threaded >call out, and everything worked great, except when I use the user level >priority, the UI (MVC) seems to slow down. I'd like to be able to make >threaded call outs, leaving plenty of processing to handle the UI, but also >having the call-out execute in a more timely manner. > >Can anyone shed some light on how to do this? Stephen, Here is some code that may help. The basic idea is to alter the priority of processes from time to time so that each gets a chance to run. The code below did work, but I can't guarantee it to be problem-free. Hope it helps. Cheers, Bob 'From Squeak 2.2beta of Sept 16, 1998 on 20 January 1999 at 8:39:02 pm'! !BlockContext methodsFor: 'scheduling' stamp: 'RAA 1/20/1999 20:38'! toggleWith: parameterList "Run the receiver block with alternating priorities as given by <parameterList> this allows a process to run at high priority for best speed while dropping down to a lower priority so that the UI does not completely die. Example: [self doSomethingLengthy] toggleWith: {6. 600. 3. 400} " | newPriority removed wait parms newProcess | parms _ ReadStream on: parameterList. newProcess _ [ self value. newProcess _ nil. ] newProcess. [ newProcess priority: 4; resume. [newProcess isNil or: [newProcess suspendedContext isNil or: [newProcess isInDebugger]]] whileFalse: [ parms atEnd ifTrue: [ parms reset. ]. newPriority _ parms next. wait _ parms next. removed _ Processor remove: newProcess ifAbsent: [nil]. removed ifNotNil: [ newProcess offList; priority: newPriority. newProcess resume. ]. (Delay forMilliseconds: wait) wait. ]. ] forkAt: 8. ^newProcess! !

1 0

Re: Newbies Wanted
by Tim Olson 02 Jun '99

02 Jun '99

>Hey Ted and John, thanks for the tutorial. I'm not a total newbie ,but ( as >the zen practitioners say ) I try to keep a 'beginners mind' . I have just >gone through the tutorial and enjoyed it a lot. I especially like the >Morphic part at the end. I think now that I will be able to make some of my >squeak projects Morphic-enabled. >thanks, >Russ Van Rooy >smalltalkin the day away in Seattle > >----- Original Message ----- >From: Ted K. <Ted.Kaehler(a)disney.com> >To: <squeak(a)cs.uiuc.edu> >Sent: Tuesday, June 01, 1999 12:25 PM >Subject: Newbies Wanted > I like the tutorial, as well -- it touches on a lot of interesting things, like object mutation, etc. One thing: does a "virgin" Squeak image start off in Morphic, yet? This tutorial assumes you are running in a Morphic world to begin with, but as I recall, the stock Squeak image still starts up in MVC. We might need to add some stuff at the beginning to get into an initial Morphic world. -- tim

2 1

Process priority and portable OS threads
by Stephen Pair 02 Jun '99

02 Jun '99

--- portable OS threads --- Last night, I got an OS threaded call-out mechanism working. I've created three pluggable primitives primitiveCallOut primitiveThreadedCallOut primitiveThreadedCallOutResult I've abstracted the machine code necessary to actually do the call out to ensure portability. This abstraction and the machine code reside in the image (in instances of ExternalByteArray). These primitives simply call the machine code for the desired calling mechanism. Thus, these primitives are highly portable, except where the native threads come in to the picture. My first cut is using tthreads http://www.upl.cs.wisc.edu/~stanis/tthread.html ; thanks to Greg A. for the pointer), a C++ abstraction that provides portability between pthreads (POSIX) and win32 threads. Does the Mac support pthreads? I'm trying to determine what level of portability I actually have here. --- process priority --- In Squeak (under Windows), when I fork a process at user background priority, it seems to never run (or at least it takes a very long time). Is there a way to have user background priority processes to run in a more timely fashion. I was creating about 100 processes to test the threaded call out, and everything worked great, except when I use the user level priority, the UI (MVC) seems to slow down. I'd like to be able to make threaded call outs, leaving plenty of processing to handle the UI, but also having the call-out execute in a more timely manner. Can anyone shed some light on how to do this? Also, I was planning to default to user interrupt priority to handle call-ins to Squeak, does that sound reasonable? Thanks, Stephen Content-Type: application/octet-stream; name="Stephen Pair.vcf" Content-Disposition: attachment; filename="Stephen Pair.vcf" Attachment converted: Anon:Stephen Pair.vcf 6 (????/----) (0000A921)

3 2

Jump to page: