Squeak-dev June 1999

squeak-dev@lists.squeakfoundation.org

139 participants
408 discussions

Re: RE: Hacking Squeak With Apple MPW
by Bob Arning 08 Jun '99

08 Jun '99

On Mon, 7 Jun 1999 17:17:44 -0400 agree(a)carltonfields.com wrote: >I was likewise unable to get a working executable under the pre-release version 4 compilers. On the other hand, the version 3 compilers work just fine and seem to generate superior performing code. That's interesting. I was able to generate VM's last summer/fall and they ran better that the distributed ones. This spring I tried and the image crashed as soon as I tired to do anything. Guess I have to dig out the old compiler. Cheers, Bob

2 1

Re: Lazily-evaluating Squeak objects
by Brian Rice 08 Jun '99

08 Jun '99

>You don't need to use dictionary. > >Suppose you need to build two objects, A and B, and each >needs a reference to the other. You create one (say A) >and then the other. When you create B, you can pass in >a reference to A, so you can ensure that you never have >a B without an A. However, for a short while you will >have an A without a B. This is not a problem, because >it doesn't last long. > >Often I make a Builder object that creates both A and B >and ensures that they point to each other. Outside the >Builder, I never see an uninitialized object, so the >builder hides the initialization. > >-Ralph Johnson So, you suggest to basically construct a "wrapper" object, B, for the reference variable(s) of the object A, and then to use the Builder encapsulation of initialization as necessary? The logic makes sense. It also would seem to work for the general case where reference variables may be changed to refer to self instead of another object. Well, I'll try this out and test the results. Thank you, Brian Rice

1 0

Source Code Problem
by Neal Nelson 08 Jun '99

08 Jun '99

I'm new to Squeak, although I have been using SmallTalkMT for the last six months. I have downloaded V2.4c and I for about half of the methods I get no source code. Instead I get either garbage or in some cases parts of code variously in what looks like C and Pascal or Modula. Some cases cause an error when I select the method in a browser. I downloaded the SourcesV2 file twice so I believe I have the necessary configuration. Does anyone have a suggestion as to what is wrong? Thanks.

3 2

Re: Problems with sound
by prb＠thinkage.on.ca 08 Jun '99

08 Jun '99

John.Maloney(a)disney.com writes: > One possible problem is that you tried to uncompress all the > sound data at once. .... > working room. The 60m you allocated to the VM may not have > been quite enough... After getting the GSM file uncompressed into memory the VM stats said there was about 4.5million bytes of memory free. I was assuming that this would be enough. I didn't get a low memory warning. The problem is reproducible with smaller files. About 300K bytes of GSM audio is enough to cause the VM to fault. > Another posssible problem is that the GSM file created > outside of Squeak may include header or block information. If > so, you'll need to parse the file format and extract the > actual GSM data before passing to Squeak's GSM codec. > > What tool did you use to do the compression? I used the 'toast' command from the GSM 6.10 library written in C by Jutta Degener and Carsten Bormann circa 1993. I've read the source, I'm reasonably confident there is no header on the GSM data. (If it is there it is well hidden.) The format is just a linear list of 33 byte frames. I don't yet know if the handling of the wasted 1/2 byte is the same as is implemented in squeak. They're both using 33 byte frames. These files I'm using will play in squeak and sound indistinuishable from the natively played version. It's just that a minute or so into the playback the squeak VM faults. I will see if I can generate a GSM file outside of squeak that works perfectly inside of squeak. Thanks - paul

4 5

Morphic Data Entry
by Neal Nelson 08 Jun '99

08 Jun '99

I'm new to Morphic and I'm having a problem creating fairly standard windows which basically consist of data entry forms. I've tried using TextMorph for my entry fields but I can't seem to get them to work as a normal entry field would. If I create then with text in them, they keep that length but if I create them empty (as the user needs to populate the field), they have zero length and any text entered isn't ever displayed but the pointer moves down a line for each character. If anyone could explain to me how I can generate entry fields in Morphic, I'd be more grateful. I'm stumped on my current project until I can entry forms working. Thanks.

1 0

Re: Article in Wired
by Dino 08 Jun '99

08 Jun '99

If you are calling to volunteer... I raise my hand. Although I am a beginner at Squeak, I feel if I manage documentation for Squeak, I'll probably learn a whole lot. I am hoping others may volunteer to manage such a large task... -- Dino agree(a)carltonfields.com wrote: > > It is time to start building a corpus of useful squeak documentation, and to make this a substantial collaborative effort by the community, aiding rather than hindering the excellent work of the good folks at Disney. Any thoughts how we might best proceed?

1 0

Re: Article in Wired
by Bijan Parsia 08 Jun '99

08 Jun '99

--On Monday, June 07, 1999, 12:46 AM -0400 Dino <gte949e(a)prism.gatech.edu> wrote: > I strongly back this up.... One of the most difficult aspects when I > began programming in Squeak was due to lack of documentation. Since > everything in Smalltalk is dynamically binded, it can be extremely > difficult to trace what kind of variable goes with what class especially > when they have awkward names. I don't understand what you're talking about. An example of this problem and how it's due to dynamic binding and how it's not addressed by current tools would help. > I think all classes should have a running > example of itself along with a few comments along the way. Worthy ideal. > Something like the Java API would be incredible. One of the reasons Java > became popular fast... it was easy to learn, documentation and books > everywhere about it. If you don't even have an API, beginners will have a > difficult time. Some cannot tolerate spending large amounts of time > sifting through source code to find what they want. That's one reason so > few people even give Squeak a chance. This I really, really, really must protest. What makes you think that it's "so few" rather than "so many" who even give Squeak a chance? I think the Squeak community is vibrant, growing, and, in general, doing rather well. As for documentation, there *is* documentation available for large chunks of the Squeak system. The Blue & Purple books do a fairly decent job. Inside Smalltalk is widely applicible. (Yes, yes, yes, I grant that in-system, online, cheaper, freer, more easily accesible, etc. documentation is better, but that's a lot different than saying there's *no* documentation, or that the documenation is wholely inadquate, especially if you haven't *used* it.) Finally, and I've said this before, and I'm not the only one, Squeak has not yet hit is "for the public/newbies" release yet. Morphic, for example, is *still* experimental, though it's getting more polished with every release. I think this is all made quite clear. Yet another call for documentation (as opposed to *providing* documentation) isn't telling anyone anything new. > A language with dynamic binding and > no documentation is like trying to find a colored needle in a haystack > without knowing what color you're looking for. Again, I don't know what this means. I found the tools for navigating Squeak code inadequate *until* I read some books on Smalltalk. Now, they typically suffice and I'm often frustrated when trying to unpack, oh, java code. That's just my experience, of course, but it's a real one. Now I can pretty much figure out a class structure or a method without *too* much difficulty and without a sense of futility. And when I compare situations where I *do* have documentation and one's where I don't, documentation don't necessarily make much of a difference. (And yes, I have concrete examples. And yes, I think the documentation is decent.) The big picture docs that Lex describes also exist (at least some of them). One move I'm trying to make now is from understanding Squeak well enough to do stuff and learn new things in a comfortable way, to feeling totally at home designing and implementing ideas in Squeak in a natural and perspicuous way. For that shift, I've found Kent Beck's books invaluable. Cheers, Bijan.

2 1

Lazily-evaluating Squeak objects
by Brian Rice 08 Jun '99

08 Jun '99

I have a question of Smalltalk semantics. I am working on a self-contained information system which I am trying to implement in the Squeak environment. The nature of the system is such that some of information objects must contain references to themselves, or references to other objects that reference themselves. Of course, my implementation takes my information objects and maps them directly to Squeak objects, with variables containing the object's references. Naturally, instantiating an environment necessarily results in system deadlocks, due to self-referencing objects resulting in infinite recursion of message calls. I've trashed my special-purpose image several times trying to work around this. Now, I know that a solution would be to implement the system as a dictionary that contained the references, with rules keeping the dictionary true to the environment model. I've considered making a subclass of dictionary, and implementing it that way, but I'm inexperienced with systems Squeak programming. I'm not familiar with the system's basic classes, so that this would be difficult. The code would at least be complicated. So, my question is: Is there a simple solution to this problem, or at least one that has already been crafted, perhaps in other Smalltalk systems? Otherwise, pointers on modifying Dictionary's methods would help immensely. Thanks for your time, Brian Rice

1 0

Re: Publish-Subscribe
by Peter Smet 08 Jun '99

08 Jun '99

I have sent you the original 2 versions of Publish Subscribe. The WeakKeyDictionaries (publishers) have always been weak dependents of WeakArray, so that is not likely to be the source of the problems you have been having. Finalization was handled this way right from version 0. I think the problems are more likely to involve some sort of deadlock or other synchronization problem. If you don't need process-safe code, the quickest fix is simply to remove the #protected: block from the 3 or 4 methods in PostOffice that use it. Then test using the class side tests individually, but don't run the 'start' process - that test is virtually guaranteed to cause problems. I am going to look at the synchronization stuff in detail soon, but have some exams looming that I must take care of first. Let me know if this fixes things. >Peter: > >I'm working on my event system, again, and need at least a partial fix for >you stuff. Unfortunately, I tossed the older, working versions. The >problem seems to have started when you made the publishers dictionary a weak >dependent of WeakArray (if I recall correctly). I don't remember how you >handled finalization before that. I don't need Smalltalk-process-safe code, >now, just something to get me by. > >(I don't know whether you received my previous e-mail: the latest >Publish-Subscribe change when tested will hang a clean 2.4c /w updates.) > >Thanks. > >Shaping > >

2 1

Block temporary variables + Exceptions
by Pierre Roy 07 Jun '99

07 Jun '99

Hi folks, Where can I find a patch for block temporary variables ? And also an implementation of a decent Exception handling mechanism ? - P

1 0

← Newer
1
...
34
35
36
37
38
39
40
41
Older →

Jump to page: