Squeak-dev January 2001

squeak-dev@lists.squeakfoundation.org

112 participants
318 discussions

[ENH] ConnectionQueue (added blocking waitForConnection/waitForConnectionIfDestroyed methods)
by Peter Schuller 25 Jan '01

25 Jan '01

Here it goes again with proper [ENH] prefix and copy to squeak(a)3plus4.de. The attached changes set adds: ConnectionQueue>>waitForConnection ConnectionQueue>>waitForConnectionIfDestroyed: and modifies other methods in said class to support this functionality. Both are blocking mechanisms for waiting for a new incoming connections. -- / Peter Schuller, InfiDyne Technologies HB PGP userID: 0x5584BD98 or 'Peter Schuller <peter.schuller(a)infidyne.com>' Key retrival: Send an E-Mail to getpgpkey(a)scode.infidyne.com E-Mail: peter.schuller(a)infidyne.com Web: http://scode.infidyne.com

1 0

Re: What's in a name?
by Dan Ingalls 25 Jan '01

25 Jan '01

>Therefore we am proposing ... Er, well, I is, but on behalf of we all at SqC ;-) Sorry 'bout that. - D

1 0

Re: Convention regarding exceptions
by Bob Arning 25 Jan '01

25 Jan '01

Duane, Tim, On Wed, 24 Jan 2001 16:54:16 -0800 Duane Maxwell <dmaxwell(a)exobox.com> wrote: >The reason there are so few is that exception handling was a relatively >recent addition to Squeak - about a year or so. IIRC, the version in the >base image still needs a little final tweaking. I think Tim Rowledge has >the necessary modifications to make it complete. I know several people >have been lobbying for inclusion. When/if that happens, some adventurous >Squeakers may embark on a mission to add rational exception handlers to >the rest of the image. Tim's changes arrived as part of one of SMA's packages back in November. Unfortunately, Tim's changes collided with those by Ian and Marcus that were needed by jitter3 and were set aside until they could be resolved. I don't know if Tim was informed of this nor whether any further action was taken. Perhaps if Tim or someone else could reconcile the differences, this matter could be wrapped up. Cheers, Bob

2 2

Re: C Parser
by Richard A. O'Keefe 25 Jan '01

25 Jan '01

Eric wrote: >Unless someone can point out a useful C editor for Windoze, I'm planning to >write something to help me out. You really can't go past Emacs. It does more than you can imagine, and it's completely free.

2 1

Convention regarding exceptions
by Peter Schuller 25 Jan '01

25 Jan '01

I was just about to re-submit the changes to the ConnetionQueue I made when I decided to improve it some more before doing so. (Thanks for all the info on the submission process!) I know have a question regarding Squeak/Smalltalk exception usage. I know what I want to do, but I don't know if it would be deemed appropriate. ConnectionQueue now has a waitForConnectionIfDestroyed: method that will block forever until a socket becomes available, or evaluate the given block if the receiver is destroyed by another process. But I still have waitForConnection. However I am not sure what to do about the default behavior should the receiver be destroyed while this method is blocking. First of all I would personally opt for throwing an exception. After all, if the calling code wanted to handle the case of a destroyed socket, it can use the waitForConnectionIfDestroyed. If not, and the receiver is destroyed anyway because of a bug, an exception that actually says that the ConnectionQueue was destroyed is more informative that the generic error which will ensue if it just answered nil. Now, here is my dilemma: Coming from a Java background my first reaction is to create a new Exception subclass (ConnectionQueueDestroyedException for example). This will enable the calling code to handle this particular exception very easily simply by relying on the class matching functinality of Squeak's exception handling (i.e.: [ .... ] on: ConnectionQueueDestroyedException do: [ ... ].). However, looking at the existing API, Exception classes are *very* sparse; it seems most errors are signaled by a normal Exception with the appropriate text. Is this also becaue of historical reasons (i.e. exception handling not always being supported by Squeak), or would creating a new Exception subclass just for "my" little case go against what Squeak:ers want? I think there is every reason to create a new Exception subclass; but searching for "Exception" in the class browser yields only 5 classes that appear to be actual exception types. This makes me apprehensive. -- / Peter Schuller, InfiDyne Technologies HB PGP userID: 0x5584BD98 or 'Peter Schuller <peter.schuller(a)infidyne.com>' Key retrival: Send an E-Mail to getpgpkey(a)scode.infidyne.com E-Mail: peter.schuller(a)infidyne.com Web: http://scode.infidyne.com

2 1

NEWBIE Question: == vs =?
by Phuong Bui 24 Jan '01

24 Jan '01

What exactly is the difference between the two boolean mesages: #== and #= ? Can someone give an example where #== results in true but not #= ? (and vice versa?) Much thanks ahead of time, Phuong Bui

3 2

Re: NEWBIE Question: == vs =?
by Bob Arning 24 Jan '01

24 Jan '01

On Wed, 24 Jan 2001 12:04:13 -0800 "Phuong Bui" <pbui(a)mail.worldkey.net> wrote: >What exactly is the difference between the two boolean mesages: #== and #= ? > >Can someone give an example where #== results in true but not #= ? >(and vice versa?) #== answers true if the receiver and argument are the same object. #= answers true if they are the same object OR they are equivalent, the definition of equivalent being one that can vary from class to class. Example. | a b | a _ 'hello, world'. b _ a. a == b. "<==== true" a = b. "<==== true" b _ a copy. a == b. "<==== false" a = b. "<==== true" The above answers your vice-versa case. The case of #== being true and #= being false *could* happen if you defined a class where #= returned false when presented with the same object. Why you would want to do that, however, I can't imagine. So the short answer is #== false and #= true happens a lot. #== true and #= false probably never happens in real life. Cheers, Bob

1 0

Meaningful Swiki URLs?
by Henrik Gedenryd 24 Jan '01

24 Jan '01

Would it be possible to augment the Swikis to allow some pages to be converted from numbered names ("1235.html") to meaningful names, and/or adding named subdirectories with default index pages. This would be useful for certain more significant topics. In that way we could refer questions to meaningful URLs instead of essentially telling everyone to go to the swiki and search for a given topic? e.g. http://minnow.cc.gatech.edu/swiki/15 might become http://minnow.cc.gatech.edu/swiki/pws or something. others could be swiki/projects, swiki/download, swiki/faq, swiki/screenshots and so on. Whenever someone asks me about the type stuff, I can't remember the URL (I think one of them is 1241.html), or even what links to follow. Henrik

2 1

Re: Plugin Security (Was Re: How do I create a SqueakPlugin.imagefrom a 2.9a ?)
by Bert Freudenberg 24 Jan '01

24 Jan '01

On Tue, 23 Jan 2001, John M McIntosh wrote: > >You completely disallow it? How about only disallowing it when the > >original squeaklet was not a "file:"-URL? > > We disallow the ability for a Smalltalk application to use the > browser interface to parse a file: URL. There is a interface in the > Smalltalk code that lets you give a URL to the browser and fetch back > the stream of data, or post a stream of data to the supplied URL. I know, I implemented that stuff for Unix ;-) My point is, when I locally test my plugin I may want to use "file:" URLs. OTOH, it may well work when I use relative URLs only ... I'll put disabling file: URLs onto my "to do" list, then. Do you fail silently or display a notifier? -- Bert

2 1

ConnectionQueue improvement (changes)
by Peter Schuller 24 Jan '01

24 Jan '01

Hello, attached is a set of changes that adds blocking I/O to the ConnectionQueue class. Specifically, it adds a waitForConnection method. It uses a semaphore to achieve the blocking functionality. listenLoop and oldStyleListelLoop has been changed to signal that semaphore when a connection is accepted. This is something I personally want. Is it something that might get added to "standard" Squeak? If not, I'll re-write the functionality in a separate class. I don't want to rely on non-standard modifications to standard classes. -- / Peter Schuller, InfiDyne Technologies HB PGP userID: 0x5584BD98 or 'Peter Schuller <peter.schuller(a)infidyne.com>' Key retrival: Send an E-Mail to getpgpkey(a)scode.infidyne.com E-Mail: peter.schuller(a)infidyne.com Web: http://scode.infidyne.com

4 6

← Newer
1
...
8
9
10
11
12
13
14
...
32
Older →

Jump to page: