Squeak-dev September 1999

squeak-dev@lists.squeakfoundation.org

133 participants
357 discussions

[ENH] sockets and GC
by lex＠cc.gatech.edu 24 Sep '99

24 Sep '99

Squeak does finalization now on sockets which is nice. However, it would be nicer if socket allocation would force a GC when necessary. That way, you could have a style of programming where you never bother to destroy sockets when you are finished with them; you could just let them become garbage, much like any other memory. Of course, it would also turn some buggy programs into correct ones, if they currently leak a socket here and there :) Anyway, the below changeset does this. It adds a method retryWithGCIf: which can be used to try an operation, and, if it fails, to do a GC and then try it again. (A cool improvement would be to first try an incremental GC, and only if the routine fails again to try a full GC. I don't know how to do this, though). The changeset then inserts this routine in Socket's newTCP, newUDP, and acceptFrom: primitives. Lex 'From Squeak 2.5 of August 6, 1999 on 24 September 1999 at 9:45:51 am'! "Change Set: SocketGC Date: 24 September 1999 Author: Lex Spoon If socket creation fails, do a garbage collect and then try again "! !BlockContext methodsFor: 'controlling' stamp: 'ls 9/24/1999 09:45'! repeatWithGCIf: testBlock | ans | "run the receiver, and if testBlock returns true, garbage collect and run the receiver again" ans _ self value. (testBlock value: ans) ifTrue: [ Smalltalk garbageCollect. ans _ self value ]. ^ans! ! !Socket class methodsFor: 'instance creation' stamp: 'ls 9/24/1999 09:45'! acceptFrom: aSocket ^[ super new acceptFrom: aSocket ] repeatWithGCIf: [ :sock | sock isValid not ]! ! !Socket class methodsFor: 'instance creation' stamp: 'ls 9/24/1999 09:44'! newTCP "Create a socket and initialise it for UDP" ^[ super new initialize: UDPSocketType ] repeatWithGCIf: [ :socket | socket isValid not ]! ! !Socket class methodsFor: 'instance creation' stamp: 'ls 9/24/1999 09:44'! newUDP "Create a socket and initialise it for UDP" ^[ super new initialize: UDPSocketType ] repeatWithGCIf: [ :socket | socket isValid not ]! !

1 0

Re: File-ins/change set formats and a "diff view"
by Robert M. Fuhrer 24 Sep '99

24 Sep '99

Wow, I had no idea. Obviously I have some more exploring to do to find more of these goodies! Most of my earlier experience with Smalltalk is with Apple's Smalltalk-80 implementation circa 1986. Obviously a fair amount has been done since then... :-) At the same time, the present file-in format can't do part of what I was talking about, namely, presenting a diff view relative to the image on which the file-in was *created*. [I.e., if I create the file-in, people to whom I send it can see the diffs relative to my system, not theirs.] As implemented, the present mechanism can only completely replace your current version, possibly squashing existing changes that are in in your current version but not in the file-in. At 07:09 PM 9/23/99, you wrote: >Also, such "diffs" view is available in versions (alt-v/cmd-v in the >method list - in system browser, change sorter, senders/implementors). > >On Thu, 23 Sep 1999, Tim Rowledge wrote: > >> On Thu 23 Sep, Robert M. Fuhrer wrote: >> > It has occurred to me that it would be really great if somehow the file-in >> > format allowed a direct presentation of what the change actually was (for >> > changes that aren't simply additions/deletions of entire methods/classes). >> > >> > E.g., wouldn't it be nice to highlight the changed portions of the code, or to >> > present a side-by-side view of old/new (relative to the creator's OR the >> > receiver's image)? -------------------------------- Robert M. Fuhrer 34-231 (914) 945-3830 IBM T. J. Watson Research Center rfuhrer(a)watson.ibm.com Yorktown Heights, NY 10598 PGP Public Key available on request

2 1

Re: Spectrum Analyzer Morph
by Bruce ONeel 24 Sep '99

24 Sep '99

Hi, Yes, I've tried it out. The sonogram one seemed not to work for me, but I chalked it up to too slow a system. The other two options worked fine. It was really neat! My office mates wondered why I was making silly sounds... cheers bruce Dan Ingalls <Dan.Ingalls(a)disney.com> wrote: > Just Curious... > > Has anyone tried this out? It's a lot of fun to play with and a significant piece of real-time software for a friendly, garbage-collecting, object-oriented language. > > Just get the 'new morph...' menu, and choose SpectrumAnalyzerMorph from the Widgets category. Make sure you have a microphone hooked up, push 'start', adjust the gain, and start making silly noises. > > Enjoy > - Dan

2 1

Re: [ENH] Halo menu picks for manipulating TextMorph attributes
by Bob Arning 24 Sep '99

24 Sep '99

On Thu, 23 Sep 1999 18:39:01 -0400 "Robert M. Fuhrer" <rfuhrer(a)watson.ibm.com> wrote: >The only problem now is that the new handle has no balloon. I haven't yet >seen where balloons get hooked in; I would have expected to see code to >associate a balloon where the handle gets created, but that isn't the case. Check out HaloMorph>>addHandleAt:color:on:send:to:. It sends handle setBalloonText: (target balloonHelpTextForHandle: handle). and Morph>>balloonHelpTextForHandle: is where the balloons get their text. Cheers, Bob

2 1

New Complete Morphic Application (Sugar Miniature)
by Juan Manuel Vuletich 24 Sep '99

24 Sep '99

Hello Folks. I have just updated the latest (English version) Miniatures pages at SugarWeb. This includes a new Morphic Application. It is a Scrabble game, for several players, with a Client/Server architecture, that runs over TCP/IP (we have played from Windows clients on a Sun server, over Internet!). The Morphic client is very preety and worth taking a look at. You can see them at http://www.sugarweb.com/Miniatures/MiniaturesE.htm In case you don't know us, we are SUGAR (Smalltalk User's Group Argentina) and we have a mail list in Spanish and a web site that is mostly in Spanish too, but we try to have the most interesting parts in English also. The English home page is at http://www.sugarweb.com/indexE.html Hope you like it. Juan Manuel Vuletich jvuletic(a)dc.uba.ar

1 0

Re: Unix 2.5 VMs
by Bruce ONeel 24 Sep '99

24 Sep '99

Ah yes, I didn't watch where that reply was going. Sorry, wrong person. a very red faced bruce...

1 0

Unix 2.5 VMs
by Bruce ONeel 24 Sep '99

24 Sep '99

Hi, I'm the person who tries to keep ftp://st.cs.uiuc.edu/Smaltalk/Squeak up to date by building bundles for different systems with the VM, plugins, image, changes, etc. Would you like some help with updateing the unix VMs to 2.5? I'd be happy to try though I'm sure I'd need some help. Thanks very much for the unix port, I use the Solaris one daily (like to send you this note!). cheers bruce

1 0

Performance of algorithms in Squeak
by Henrik Gedenryd 24 Sep '99

24 Sep '99

How well do the Squeak algorthms do in general (algorithmically, that is)? How much effort has been put into them? I looked at the new Heap implementation, and came to think about how well optimized the quicksort implementation in Squeak is (Sedgewick's tricks, swapping triplets, selection-sorting small pieces, etc.). Then I thought more generally about the first question above. I also thought it might be a great Algorithms course assignment (college professors out there looking for projects: hint! hint!) to evaluate and improve algorithmic performance in Squeak, eg.: a) (find and) evaluate used algorithms, b) find out which classes & methods are heavily used and how, and thus what changes would actually have a wider impact on Squeak performance (my guess is the collection classes, hashing and so on), and c) implement better ones and evaluate them. How come we always come up with such great ideas for what other people might do? Best regards, Henrik

1 0

Re: How does double clicking work in Squeak?
by Henrik Gedenryd 24 Sep '99

24 Sep '99

Chris, Morphic has some built-in general support for generating drag:, click:, and doubleClick: messages, in addition to the usual mouseDown: etc. It is quite simple to use. (I now realize that Morphic may be of little help on a PDA.) All of these are implemented by one and the same mechanism, which is somewhat misleadingly named 'double click support', see the methods in the HandMorph category with this name. (Bob's code-not-so-snippet is more or less a simple re-implementation of some of this functionality.) However, Morphic doesn't normally use these itself for its clicks and drags, but regards mouseDowns as clicks, and uses 'hardwired' drags. This makes Morphic somewhat less simple to extend and modify. But here's how to use the stuff: First of all, the Morphs that wish to receive these events to mouseDown need to add waitForClicksOrDrag:event: to their mouseDown: method, something like this: mouseDown: event fromMorph: sourceMorph "ask to get click/doubleClick/drag messages" event hand waitForClicksOrDrag: sourceMorph event: event. .... Or as Morph>>click: states: click: evt "Handle a single-click event. This message is only sent to clients that request it by sending #waitForClicksOrDrag:event: to the initiating hand in their mouseDown: method. This default implementation does nothing." Don't forget to signal that the morph class wants the mouseDowns: handlesMouseDown: evt "always receive mouseDowns to initiate drag/doubleClick detection" ^true Then you just implement your click:, doubleClick:, and drag: methods. Best regards, Henrik < = > .

1 0

Re: Unicode support
by Hannes Hirzel 24 Sep '99

24 Sep '99

Hi Yoshiki, Thank you for the elaboration on the design of your implementation. OHSHIMA Yoshiki wrote: > Please note that > the design of my implementation is heavily influenced by > "Mule" (Multilingual enhancement to GNU Emacs), whose > development has began more than 12 years ago and still > agressively improving. > Roughly speaking, the character representation in my > implementation is somewhat similar to > SmallInteger/LargePositiveInteger integration. [...] > you can test it with vanilla VM. [...] > I found many people mentioned about the definition of > string and String on this thread, but I did't find any > article which mentioned what the definition of character > should be. There was a contribution by Peter W. Lount on this http://minnow.cc.gatech.edu/squeak.871 > On a system like Squeak, where the glyph of a character > should be controlled by the system itself, the character > should know how to represent itself. This is the reason why > I think the representation should carry enough information > (more than 16-bit). > One more thing I'd like to say is, the Unicode could be a > "local" encoding in my framework. There is so much software > which assumes Unicode, Squeak should be able to support it. > However, the local encoding would not have glyph, because > it's Unicode:-) Could you elaborate on this? Which other encodings are you considering? Thank you Hannes

1 0

← Newer
1
...
6
7
8
9
10
11
12
...
36
Older →

Jump to page: