Squeak-dev May 2001

squeak-dev@lists.squeakfoundation.org

143 participants
597 discussions

Re: InputSensor kbdTest
by Bob Arning 28 May '01

28 May '01

Bert, The problem is that processKeyboardEvent: evt "Never update keyboardBuffer if we have an eventQueue active" and kbdTest utlimately depends on the keyboardBuffer. Here is a version that seems to work in Morphic. Cheers, Bob ===== 'From Squeak3.1alpha of 5 February 2001 [latest update: #4089] on 28 May 2001 at 7:18:36 am'! "Change Set: kbdTest Date: 28 May 2001 Author: Bob Arning A version of #kbdTest that works in Morphic"! !EventSensor methodsFor: 'keyboard' stamp: 'RAA 5/28/2001 07:17'! kbdTest "Sensor kbdTest" "This test routine will print the unmodified character, its keycode, and the OR of all its modifier bits, until the character x is typed" | char evt type pressCode charCode | [ char _ nil. inputSemaphore signal. eventQueue ifNil: [ keyboardBuffer peek ifNotNil: [ char _ self characterForKeycode: self keyboard. mouseButtons _ self primMouseButtons. ]. ] ifNotNil: [ (evt _ eventQueue nextOrNil) ifNotNil: [ type _ evt at: 1. EventTypeMouse = type ifTrue: [self processMouseEvent: evt]. EventTypeKeyboard = type ifTrue: [ pressCode _ evt at: 4. pressCode = EventKeyChar ifTrue: [ charCode _ evt at: 3. charCode ifNil: [^self]. mouseButtons _ (mouseButtons bitAnd: 7) bitOr: ((evt at: 5) bitShift: 3). char _ Character value: charCode. ]. ]. ]. ]. char ifNotNil: [ {char. char asciiValue. char asciiValue radix: 16. mouseButtons} printString, ' ' displayAt: 10@10. ]. char = $x ] whileFalse! ! ===== On Mon, 28 May 2001 11:48:45 +0200 (CEST) Bert Freudenberg <bert(a)isg.cs.uni-magdeburg.de> wrote: >On Wed, 23 May 2001, Bob Arning wrote: > >> On Wed, 23 May 2001 18:12:02 -0700 (PDT) doug edmunds <being_doug(a)yahoo.com> wrote: >> >That's what I had tried, but apparently from >> >within Morphic. This leads to the question, >> >why would this work in MVC but not Morphic? >> >> It stems from a rework of the event handling mechanisms. > >But the EventSensor tries to be compatible whit the InputSensor. Looking >at the code I see no reason why it should not work. > > InputSensor>>kbdTest > InputSensor>>keyboardPressed > EventSensor>>primKbdPeek > inputSemaphore signal > EventSensor>>ioProcess > primGetNextEvent > >Is it possible the ioProcess is not run immediately when the signal is >raised?

1 0

Re: [BUG] can't move a window twice without intervening mouse motion
by Bert Freudenberg 28 May '01

28 May '01

On Fri, 25 May 2001 JArchibald(a)aol.com wrote: > => 5/25/01 4:37:40 PM EDT, Andreas.Raab(a)disney.com => > << What platform are you on?! I seem to be unable to duplicate the problem. >> The bug only shows when fastDragWindowForMorphic is on. -- Bert

1 0

Re: convincing squeak to use LF instead of CR as the line separator...
by John Hinsley 28 May '01

28 May '01

"Greg A. Woods" wrote: > > [ On Thursday, May 24, 2001 at 23:32:07 (+0100), John Hinsley wrote: ] > > Subject: Re: convincing squeak to use LF instead of CR as the line separator... > > > a bigger problem for me is trying to > > read Macified code and text outside of Squeak. God, it's *horrible*! > > It's not too bad in Emacs if you get emacs to do the conversion on the > fly. The only problem is in getting the file "type" detection to work > properly. Oh if only people would/could always put: > > -*- mode: smalltalk; coding: mac -*- > > on the first line of their Smalltalk files (for example)..... :-) Well, I get round the problem (more often it's a case of people running Win 2000 boxes coming to me with a floppy containing some java file knocked out under Solaris or Mac, so I just run recode or unix2dos or mac2dos or whatever on it. It's pointless explaining that a decent OS should come with such tools...for most of them doing command line java is the closest they'll ever come to a text file :-) But for just reading I tend to rely on Gvim doing it automagically. I won't enter into the Emacs vs Gvim wars! But am I right in guessing that, from the point of view of filing in code, Squeak doesn't care about the LF/CR stuff? So I can "hard convert" files to my hearts content? > > well, the Unix choice of a <linefeed> as the text line separator was > "unconventional" once upon a time.... But surely there's a history to this..... I just wonder why, given that Unix preceeded DOS and Mac, they both decided to do things differently? (But I'd imagine that the Smalltalk 80 people were all developing on Macs, so that the choice sort of followed on from that....) I wonder which tradition OS X follows? Cheers John -- ****************************************************************************** Marx: "Why do Anarchists only drink herbal tea?" Proudhon: "Because all proper tea is theft." ******************************************************************************

2 1

squeakland plugin ScheduledControllers nil
by Baveco, Hans 28 May '01

28 May '01

I have a minor problem with the Squeakland plugin: the basic plugin image has apparently ScheduledControllers set to nil. When the plugin (no morphs on the screen) starts with downloading and filing-in some code and then tries to open the filed-in application's morph, it complains that ScheduledControllers is nil. After opening a random morph (eg the browser), I can restart opening the application and everything works as before (in the last non-squeakland plugin I didn't have this problem). Is this something that should be fixed in the plugin? or should I change the sts file to open some morph before filing-in...? Hans

1 0

Re: Embedding Squeak in REALbasic?
by Tom _ 28 May '01

28 May '01

The biggest problem I've had in addressing embedding issues is the inability to implement a robust callback facility. I would love to see a sound embeddable, perhaps reentrant, version of the VM. Another approach would be to have the two sides run in separate threads and communicate via a simple message queue facility. They can still exchange pointers efficiently, but their calling conventions and stacks don't have to mesh. You can still give the user the illusion of callbacks. If preemption is a problem, you can either use system threads and turn off preemption, or you can use SGI's state thread library. Tom. __________________________________________________ Do You Yahoo!? Yahoo! Auctions - buy the things you want at great prices http://auctions.yahoo.com/

1 0

RE: running directly from RAM on wince?
by Noel J. Bergman 28 May '01

28 May '01

>> Build for Win32, and test on your PC. These are not WinCE >> specific APIs. :-) > In theory yes, but in practice ... In practice, they work just fine in the PC Emulator, where I tend to do all of my first round testing for PocketPC. I didn't mean to give any impression that he could or should run them native on WinMe or Win2K. You won't see important device performance differences in the emulator. For example, the cost to page in a 1K page on the iPAQ is approximately 0.2 milliseconds when the memory mapped file is located in the internal object store, and approximately 1 millisecond when the file is located on a CF II solid state storage card. You won't see any of that in the PC Emulator. Also, iterating through memory in byte order is 6 to 7 times faster on the iPAQ than iterating over the same region of memory in page order. Both of which imply that co-locality is really important to performance on the iPAQ. >> Shall we keep this between you, Ohshima-san, Ned and I until we get it >> working? Just so as not to swamp the list. I don't know if >> everyone wants to follow the birthing pains. > Please keep me in the loop - I've got a bit of experience with > that kind of stuff. Absolutely! Just didn't want to bore everyone with the birthing pains. But anyone who WANTS to see/help ... --- Noel

1 0

Re: Design Principles Behind Smalltalk
by John Hinsley 28 May '01

28 May '01

Gary wrote: > Message answer = Nil ??? :O) > Gary Ahem. Well, I can't directly answer the question. But "Future Work" in the context of Smalltalk might be taken to imply Self (or Squeak ;-) ). "Have other aspects of human thought been identified as metaphors?" Might be taken to imply either hermeneutics or cognitive psychology or both. There has been some discussion around this on the Self Interest list (it's archived at yahoo). You might find that some of the threads by or references by Albertina Lourenci (whose field is architectural ecosystems) are of use. Be warned: these threads are extremely free form and wide ranging in their references. Certainly when I encounter a 3 way discussion between Albertina, Jecel Assunpcao Jr and David Ungar, I feel well humbled! Cheers John Incidentally, trying to reply directly to your posting crashed Netscape: are you doing something wierd in Outlook with formatting? -- ****************************************************************************** Marx: "Why do Anarchists only drink herbal tea?" Proudhon: "Because all proper tea is theft." ******************************************************************************

1 0

Re: comp.lang.smalltalk.squeak ?
by Bob Hartwig 28 May '01

28 May '01

> >I'm not that much in a flaming mood, only a little discouraged. It sounds a >little bit like Mr. Hartwig is suggesting something along the lines of "it's >THEIR platform, so give them the least possible utility and that's plenty >good enough, but for MY platform I want this and this and this..." Am I >misinterpreting anything here? > I was just suggesting that the problem has already been solved by Yahoo's digest option. If I didn't have a filtering mail client, I would subscribe to the digest and my problem would be solved. By the way, if a bidirectional list to newsgroup gateway was established, that would be a nice solution too, and would make everybody happy. -Bob

1 0

Re: pending file primitives (EH?)
by Richard A. O'Keefe 28 May '01

28 May '01

3) query file/filesystem characteristics, for example on the Windows platform, you have to ask the filesystem that contains a specific filename if it's case sensitive as different filesystems vary True in principle in UNIX also. That's why (f)pathconf() exist.

1 0

Re: convincing squeak to use LF instead of CR as the line separator...
by Richard A. O'Keefe 28 May '01

28 May '01

Concerning the text/binary distinction: I first met such a distinction on the Burroughs B6700. The Burroughs file system associated a "file type" byte with each file, there was a byte for APL Workspaces, there was a byte for BASIC source code, there was a byte for Algol object code, and so on. The snag, of course, was that there wasn't really any way to generate your _own_ file types. The Macintosh has the same feature, but generalised to 4 bytes. I would have thought that it would have made more sense for a file type to itself be a file system object, so that you could create them ad lib and attach whatever properties you wanted to them. Macintosh tools like browsers and FTP agents typically "fudge" this information from file name extensions, if they can't get it by other means. This can have unpleasant consquences. Windows can be told to associate types/icons with files by extension as well, and so can some UNIX desktops. Since Celeste has to deal with MIME types and HTTP uses MIME types, it might be simplest if a view of file encoding and type that was as close to MIME as reasonable could be applied uniformly. This might require system- dependent backup, such as mappings between extensions and MIME types. Indeed within Plan 9 are the fruits of lots of man-years of experience with handling multi-lingual text too! And that's certainly true.

1 0

← Newer
1
...
8
9
10
11
12
13
14
...
60
Older →

Jump to page: