Squeak-dev February 2001

squeak-dev@lists.squeakfoundation.org

144 participants
725 discussions

Re: emacs and squeak, again...
by Yoshiki Ohshima 25 Feb '01

25 Feb '01

Hello, Ah yes, of course the purpose should not be to port an existing editor to Squeak, but to create something better. >> I don't have a feel for how large these tasks are, but it doesn't seem >> like it should be *that* bad. The really hard part seems to be the >> mapping in #1. Eg, perhaps there could be one Squeak window per buffer, >> and the "mini buffer" could accept commands via pop-up windows, and the >> messages window is Transcript. One has to come up with parallels like >> this for all of Emacs's fundamental data structures. > > The Emacs distribution has 275,860 lines of C (including .h files) and > 507,035 lines of Emacs Lisp. I suspect that a lot of that C doesn't > map trivially to Smalltalk. :) So yes, it's probably *that* bad. We don't have to port the Emacs lisp interpreter code (including the memory management and other things) of Emacs. And a certain part of the other C code is only Lisp code written in C syntax so actually we could map most of them to Squeak:-) In this respect, it is not that bad. On the other hand, why I think it would be "lots of labor" is that the presense of the "imcompatibility" of Elisp and Squeak. For example, Elisp is dynamic scoped language while Squeak is not. Look at this: (setq c 3) => 3 "c is globally bound to 3" (defun func (a) c) => func "function func returns the value bound to c" (func 1) => 3 "returns 3, as expected" (defun caller (a) (let ((c 2)) (func a))) => caller "caller returns the value of func, but?" (caller 1) => 2 "not 3. the returned value is a 'temp' of function foo" This is just an example, but there would be issues that make simple mapping from Elisp function to Squeak method impossible. Rewriting those "elisp specific code" would not be difficult, but I think it can be a lot of labor. -- Yoshiki

2 1

[GOODIE] ZipViewer and Archive-Zip (new and improved)
by Ned Konz 25 Feb '01

25 Feb '01

Here's an improved version that also displays file contents, allows adding entire directory trees, and has better button resizing behavior. It also has balloon help on the buttons. And it's integrated with the FileList. from preamble: "Change Set: ZipViewer-nk Date: 25 February 2001 Author: Ned Konz This is a moderately full-featured graphical Zip file explorer. It only works in Morphic. It requires my Archive-Zip change set. "!

2 4

Re: Plugin Installers
by Karl Ramberg 25 Feb '01

25 Feb '01

I should note that the part of interest is the three bottom lines in the picture. I had to move the Transcript to bring it into view, that's all the fours in the beginning of the third line. After that I just moved the mouse around. The 89 occured mousing over the title bar of the transcript. The last two numbers are from taking the screenshot. Karl Karl Ramberg wrote: > > Ok, I attached a screenshot in pict format. It mostly picks up > zeros but then a single four pops up now and then. > Also another thing is that the squeak part of the screen is slightly dimmed > all the time while in Netscape. Not in iCab. > Karl > > John M McIntosh wrote: > > > > > > > >The sad part is that the plugin does not work. I don't know why yet. > > >One issue is that all mouse events also sends mousedown, ie the world > > >menu comes up when > > >I move the mouse around. > > >The other issue is loading stuff like .sts and .pr files. This gives > > >walkbacks. > > > > > >Karl > > > > > > > Karl could you check to see if you have a 2.8 version of the plugin > > lurking in your plugin folder? > > > > I'm quite puzzled by the behavior since looking at the VM code > > doesn't tell me why this appears to happen. Of course I have normal > > behavior on my machine. > > > > So if you could added to the end of processMouseEvent > > > > EventSensor>>processMouseEvent: t1 > > ... > > > > Transcript show: mouseButtons; space > > > > On mouse movement it should print 0. It will stop printing 0 if you > > stop moving the mouse > > > > On a mouse click it should print 4 4 0 0 > > > > Mouse down, mouse up, and 2 moves until it decides that no more > > events should be passed up from the VM if you stop moving the mouse. > > > > -- > > -- > > =========================================================================== > > John M. McIntosh <johnmci(a)smalltalkconsulting.com> 1-800-477-2659 > > Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com > > =========================================================================== > > Custom Macintosh programming & various Smalltalk dialects > > PGP Key: DSS/Diff/46FC3BE6 > > Fingerprint=B22F 7D67 92B7 5D52 72D7 E94A EE69 2D21 46FC 3BE6 > > =========================================================================== > > ------------------------------------------------------------------------ > Name: screen.pict > screen.pict Type: PICT Image (image/x-pict) > Encoding: x-uuencode > Description: Unknown Document

1 1

Re: Zooming interface
by Richard A. O'Keefe 25 Feb '01

25 Feb '01

So my claim here is that via what libraries are for, it should not be difficult to find out what has existed over the last 40 years, nor should it be difficult to get it. Compare this with other fields, the relatively small amount of stuff done since 1960, the tremendous ease of copying using xerox machines, video tapes, digital media and communication. If you can't find it readily then there is something wrong with the larger university processes re computer "science" (and I suspect there may just be). Can you say - funding cuts - skyrocketing journal prices - a dollar that catches cold every time America sneezes? I knew you could. The will is there, the skill is there, but the money ain't.

2 1

Re: [BUG][FIX][Slowdown] Add preference too slow (Symbol>>hashtoo slow)
by Yoshiki Ohshima 25 Feb '01

25 Feb '01

Hi, > We discussed this, and propose to fix the problem a > different way. MethodDictionary is a subclass of Dictionary > instead of IdentityDictionary (which didn't exist when > MethodDict was defined ;-). MethodDictionary overrides > "scanFor:" to provide IdentityDictionary speed, but it > inherits #keys from Dictionary and so produces a Set here > instead of an IdentitySet. Our proposal is to move > methodDictionary to being a subclass of IdentityDictionary. > Hopefully this will fix the problem in a more general way. Your fix solves the particlar MethodDictionary case, but still there are other many Dictionaries that uses Symbols as keys. I think their speed would be also important. I checked several expressions in my 3.0 final image. -- Dictionary allInstances size. => 522 Dictionary allInstances inject: 0 into: [:sub :next | sub + next size]. => 4900 "The total number of elements in all Dictionaries. Not too many." (Dictionary allInstances select: [:d | d isEmpty not]) size. => 420 "So, roughly 80% of Dictionaries have at least one element." keyClasses _ (Dictionary allInstances select: [:d | d isEmpty not]) collect: [:d | d keys someElement class]. "keyClasses is a collection of the keys' class." keyClasses size. => 420 "same as the third expression." (keyClasses select: [:c | c = Symbol]) size => 293 "This means 69% of the non-empty Dictionary uses Symbols as its key" (keyClasses select: [:c | c = String]) size => 114 "This means 27% of the non-empty Dictionary uses Strings as its key" [Dictionary allInstances do: [:d | d rehash]] timeToRun => ... "seems take forever!" -- ... Umm, I was about to write "so let's use faster implementation of Symbol>>hash!" I still think so, but because those "Symbol-key Dictionaries" typically used for classPools, converting the classPools to IdentityDictionaries seems a good thing anyway. -- Yoshiki

2 1

Re: emacs and squeak, again...
by Alan Kay 25 Feb '01

25 Feb '01

Hey, come on folks ..... While we are at it, let's reimplement the horse and buggy -- they were really popular for quite a while and had a certain degree of elegance. Yes, there are some good features in emacs (as in Unix), but why not try designing a 21st century editor (and a 21st century operating environment), instead of "driving into the future but steering only by looking in the rearview mirror" (as McLuhan so aptly phrased it)? Our main task here is to get to something better than Smalltalk and LISP, good as they ONCE WERE...... Cheers, Alan At 11:03 AM -0800 2/25/01, Randal L. Schwartz wrote: > >>>>> "Lex" == Lex Spoon <lex(a)cc.gatech.edu> writes: > >Lex> Can you say why you think this would be such a huge task? E-lisp has >Lex> the advantage of being a very simple language. It seems the "only" big >Lex> tasks required are: > >Lex> 1. Map Emacs's data structures such as "buffers" and "frames" into >Lex> something relevant for Squeak. > >Lex> 2. Implement all the primitives (which requires a good solution to #1). > > >Lex> I don't have a feel for how large these tasks are, but it doesn't seem >Lex> like it should be *that* bad. The really hard part seems to be the >Lex> mapping in #1. Eg, perhaps there could be one Squeak window per buffer, >Lex> and the "mini buffer" could accept commands via pop-up windows, and the >Lex> messages window is Transcript. One has to come up with parallels like >Lex> this for all of Emacs's fundamental data structures. > >The Emacs distribution has 275,860 lines of C (including .h files) and >507,035 lines of Emacs Lisp. I suspect that a lot of that C doesn't >map trivially to Smalltalk. :) So yes, it's probably *that* bad. > >-- >Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095 ><merlyn(a)stonehenge.com> <URL:http://www.stonehenge.com/merlyn/> >Perl/Unix/security consulting, Technical writing, Comedy, etc. etc. >See PerlTraining.Stonehenge.com for onsite and open-enrollment Perl training!

1 0

Re: Plugin Installers
by Dan Shafer 25 Feb '01

25 Feb '01

Definitely not working. I gave NS 100MB of room and I got the same problem. It may be bogus; I was monitoring memory usage and it didn't appear to me that we actually ran out of RAM in Netscape when the error came up. Here's the .out transcript if it'll help: out of memory 47021448 Behavior>isVariable 47021356 Behavior>basicNew 47021264 Behavior>basicNew 47021172 Behavior>basicNew 47021080 Behavior>basicNew 47020988 Behavior>basicNew 47020896 Behavior>basicNew 47020804 Behavior>basicNew 47020712 Behavior>basicNew 47020620 Behavior>basicNew 47020528 Behavior>basicNew 47020436 Behavior>basicNew 47020344 Behavior>basicNew 47020252 Behavior>basicNew 47020160 Behavior>basicNew 47020068 Behavior>basicNew 47019976 Behavior>basicNew 47019884 Behavior>basicNew 47019792 Behavior>basicNew 47019700 Behavior>basicNew 47019608 Behavior>basicNew 47019516 Behavior>basicNew 47019424 Behavior>basicNew 47019332 Behavior>basicNew 47019240 Behavior>basicNew 47019148 Behavior>basicNew 47019056 Behavior>new 47018964 Delay class>timeoutSemaphore:afterMSecs: 47018872 Semaphore>waitTimeoutMSecs: 32290100 EventSensor>ioProcess 32290316 [] in EventSensor>startUp 32290408 [] in BlockContext>newProcess --- Michael Rueger <m.rueger(a)acm.org> wrote: > Hi all, > > I built a couple of web pages that should make it easier to install the > Squeak Plugin for Mac and Win. > http://www.squeaklet.com/NPSqueak/download.htm > > You need Javascript enabled for most of the functionality (auto detect > etc.). You do need to restart the browser after the installation (part > of the test though). > Look here > http://www.squeaklet.com/NPSqueak/plugin/supported.html > for current state of testing and support. > > Please make sure you don't have an old plugin installed, I'm working on > this issue to give the option to first remove an old installation. > > The installers for mac and win include the Squeak 3.0 version. > > > Michael > > -- > "To improve is to change, to be perfect is to change often." > Winston Churchill > +------------------------------------------------------------+ > | Michael Rueger m.rueger(a)acm.org ++1 (310) 937 7196 | > +------------------------------------------------------------+ > __________________________________________________ Do You Yahoo!? Get email at your own domain with Yahoo! Mail. http://personal.mail.yahoo.com/

2 1

[BUG] PhraseTileMorph>>tileRows tries to copy a World (formerOwner)
by Ned Konz 25 Feb '01

25 Feb '01

I was just playing with the scripting and found that if I set a tile down and then dropped it into a script, I'd get a walkback that said that Worlds shouldn't be copied. It looks like PhraseTileMorph>>tileRows has been changed to use veryDeepCopy, but its former owner is not being set aside (any more), though the comment still stands. -- Ned Konz currently: Stanwood, WA email: ned(a)bike-nomad.com homepage: http://bike-nomad.com

2 2

PDF viewer/writer
by cole＠griffoninc.com 25 Feb '01

25 Feb '01

Is anyone working on PDF support in Squeak? Michael Cole

2 1

Re: [QUESTION?] Scope (visibility) of variables declared in blocks?
by Bob Arning 25 Feb '01

25 Feb '01

On Wed, 21 Feb 2001 23:54:50 +0100 (CET) Bert Freudenberg <bert(a)isg.cs.uni-magdeburg.de> wrote: >On Wed, 21 Feb 2001, Elzbieta Chludzinska wrote: > >> When I declare a value within a block: >> >> [ | tmp | ... ], >> >> shouldn't tmp's visibility be limited to the block in which it was declared? >> >> Shouldn't iterates have a limited visibility as well? For example, in: >> >> someCollection do: [ :i | ... ] >> >> shouldn't i's visibility be limited to block in which it's the iterate? > >Yes it should. This was postponed until someone finally implements >closures. Currently the declaration inside a block actually creates >another method temp, so it's visible outside, too. Well, that has actually been fixed a bit now. "i" in this case is no longer visible outside the block. Look for change sets with "scope" in the name. There are two in the low 3400's and one in the low 3500's. Cheers, Bob

2 1

← Newer
1
...
6
7
8
9
10
11
12
...
73
Older →

Jump to page: