Squeak-dev January 2002

squeak-dev@lists.squeakfoundation.org

176 participants
612 discussions

[VM] unstability
by Stephane Ducasse 01 Feb '02

01 Feb '02

Hi I'm using Squeak 3.2.1Beta6MT under mac Os 9.2 with the module image. And Squeak crashes every 10 minutes in average. Am I the only one? Stef

5 5

Re: Persistence & DTSTTCPW: ZODB clone?
by Avi Bryant 01 Feb '02

01 Feb '02

On Thu, 31 Jan 2002, Cees de Groot wrote: > The structure (now I'm going to tread dangerous territory here - this is from > the back of my head, I really should refresh my knowledge from the docs) is > basically: magic, transaction header, object, object, object, ..., transaction > header, object, object, object, ... etcetera. Objects have backpointers to > parent versions, transactions too - this mean you can 'timetravel' to older > transactions or older object versions (guess how trivial it is to build a Wiki > on top of that...). A few questions - Does anything without an oid get serialized in full to the transaction log, as part of its parent object? How does it recover the oids from the objects when writing to disk - does it keep a reverse map object->oid? Or does every business object have to have an oid field? Which brings me to - what can be used in business objects? In Squeak, will standard collections be usable, or will you have to use an ODBDictionary, ODBOrderedCollection, etc, that have oids, and can associate themselves with transactions? And how do most objects (that don't have an easy at:put: protocol) know they've been modified? Compiler hacks like kats uses? VM hacks? Or do your business objects have to send "self changed" all the time? I ask all of this because I've been thinking about similar designs myself without coming up with satisfactory answers. What I've been looking at instead, recently, is generalizing my hacked together little O/R mapping framework (sort of a GLORP-lite) so that it will work with simple in-memory tables as well as RDBs. Once you map an object structure down to simple relational tables (or arrays of dictionaries of strings, which is what a relational table looks like to my framework), transactions become very easy to serialize. Of course, the disadvantage is that you're constrained by the O/R model and have to have all of this metadata about relationships, etc. One upside is that you can freely mix your "object" (in-memory) DB and any relational DBs you need to use, in terms of having relationships between them, since they all use the same model. The other is that all of the DB-related stuff is totally external to your business objects, rather than forcing them to inherit from anything in particular (although you still have the "self changed" problem to some degree). I don't particularly like this approach, so if you manage to get a ZODB lookalike working, I'll be very pleased. I'm also happy to help out if you need it. Cheers, Avi

2 1

[ANN]Squeaker of the year award nominations
by Editor - Squeak News 01 Feb '02

01 Feb '02

Announcing: The first ever SqueakNews Squeak Developer of the year Award Nominations in by: 10th February, 2002 Announcement: 16th February, 2002 Squeak News is introducing a yearly award for the person/s who contributed to Squeak significantly. This year we'll have a very informal and rush selection. We will exclude SqC members in order to save the awards for the rest of us mortals (They are only eligible for lifetime achievement awards). Also SqueakNews employees and management are excluded :(. However excluded people can still nominate others. (our authors and volunteers are independent of Squeak News and they CAN be nominated). The panel again will be formed very informally from among SqueakNews employees and management, interested SqC members, members of the Squeak community and other prominent people who have maintaned interest in Squeak. In order to be eligible a person or persons and their work must be nominated by an individual or a group of individuals. You can nominate yourself and your own work. We do not accept anonymous nominations. Please send your nominations including the following information to: Person(s) nominated: The Work(s) that the person(s) has(have) contributed before February 2002: Why is the work significant: Any other reasons for the nomination: The Nominator: I, the nominator nominate the following person(s) for the above reasons. Name: Email: (a valid email is required, we will verify the nomination with the given email). Contact Number and best times: (will be used only in the case that we can't contact you via email, or we need to clarify something quickly) the editor at: mailto:editor@squeaknews.com?subject=2001_SqueakNews_Award or, print and fax your nomination to: +1 212 208 4454 The information collected here will strictly be used for the nomination purposes and will be destroyed after the selection is over. Nominations from more than one person/group may increase chances of the nominee(s) but will not guarantee a selection. If we receive a significant number of nominations for one person/group, we may introduce a "Most Popular Person/Achievement" award for them. In addition the *first* 3 nominators of the award winner will receive a 1 year SqueakNews subscription. The Reward: The person(s) will receive a groovy certificate of award and a small statue which will include a real piece of antique core memory (signifying their core contributions to Squeak and our rememberance of their contributions) that they can display prominently. They will also receive a lifetime subscription to SqueakNews, an article and a permanent web page describing the achievements and the award, and links to their work(s). If you would like to be in the selection panel please email me. We can not include people nominated as panel members. Tansel Ersavas Editor, Squeak News http://www.squeaknews.com

1 1

[BUG]HandMorph(Object)>>doesNotUnderstand:
by jhinsley＠telinco.co.uk 31 Jan '02

31 Jan '02

I filed Rolodex.morph (essentially a glorified book morph) into my 3.3a #4170 image. Amazingly, it all works apart from the search button I've scripted by using "do menu item" and "hand me a button to fire this script". The "find" menu item works just fine. As both worked fine in the changeset prior to the latest updates, looks like something in E toys has got clobbered. 31 January 2002 10:22:46 pm VM: unix - Squeak3.1alpha of 7 March 2001 [latest update: #4173] Image: Squeak3.3alpha [latest update: #4730] HandMorph(Object)>>doesNotUnderstand: Receiver: a HandMorph(3216) Arguments and temporary variables: aMessage: a Message with selector: #buildMorphHandleMenuFor: and arguments: #(a...etc... Receiver's instance variables: bounds: 323@404 corner: 339@420 owner: a PasteUpMorph<world>(1622) [world] submorphs: #() fullBounds: 323@404 corner: 339@420 color: Color blue extension: a MorphExtension (539) mouseFocus: nil keyboardFocus: a MenuMorph(345) eventListeners: nil mouseListeners: nil keyboardListeners: nil mouseClickState: nil mouseOverHandler: a MouseOverHandler lastMouseEvent: [323@404 mouseUp 4138562] targetOffset: 32@5 damageRecorder: a DamageRecorder cacheCanvas: nil cachedCanvasHasHoles: false temporaryCursor: nil temporaryCursorOffset: nil hasChanged: true savedPatch: nil userInitials: '' lastEventBuffer: #(1 4138562 323 404 0 0 0 0) genieGestureProcessor: nil BookMorph(Morph)>>doMenuItem: Receiver: a BookMorph<Book>(2257) Arguments and temporary variables: menuString: 'search for text' aMenu: nil anItem: nil aNominalEvent: nil aHand: a HandMorph(3216) m: nil Receiver's instance variables: bounds: 167@157 corner: 646@623 owner: a PasteUpMorph<world>(1622) [world] submorphs: #(an AlignmentMorph<Page Controls>(2581) a Morph(4058) a PasteUpMorp...etc... fullBounds: 167@157 corner: 646@623 color: (Color r: 0.9 g: 0.9 b: 0.9) extension: a MorphExtension (3113) [sticky] [externalName = Book ] [player = ...etc... borderWidth: 1 borderColor: Color black pageSize: 360@228 newPagePrototype: a PasteUpMorph<page>(1124) pages: an OrderedCollection(a PasteUpMorph<page>(2435) a PasteUpMorph<page>(149...etc... currentPage: a PasteUpMorph<page>(1494) Player96(Player)>>doMenuItem: Receiver: a Player96 (3346) named Book Arguments and temporary variables: menuString: 'search for text' Receiver's instance variables: dependents: nil costume: a BookMorph<Book>(2257) costumes: nil Player96>>script4 Receiver: a Player96 (3346) named Book Arguments and temporary variables: Receiver's instance variables: dependents: nil costume: a BookMorph<Book>(2257) costumes: nil --- The rest of the stack --- Player96(Player)>>runScript: [] in ScriptActivationButton(SimpleButtonMorph)>>doButtonAction BlockContext>>ensure: CursorWithMask(Cursor)>>showWhile: ScriptActivationButton(SimpleButtonMorph)>>doButtonAction ScriptActivationButton(SimpleButtonMorph)>>mouseUp: ScriptActivationButton(Morph)>>handleMouseUp: MouseButtonEvent>>sentTo: ScriptActivationButton(Morph)>>handleEvent: ScriptActivationButton(Morph)>>handleFocusEvent: [] in HandMorph>>sendFocusEvent:to:clear: PasteUpMorph>>becomeActiveDuring: HandMorph>>sendFocusEvent:to:clear: HandMorph>>sendEvent:focus:clear: HandMorph>>sendMouseEvent: HandMorph>>handleEvent: HandMorph>>processEvents [] in WorldState>>doOneCycleNowFor: Array(SequenceableCollection)>>do: WorldState>>handsDo: WorldState>>doOneCycleNowFor: WorldState>>doOneCycleFor: PasteUpMorph>>doOneCycle [] in Project class>>spawnNewProcess [] in BlockContext>>newProcess

1 0

Re: Iterators? (Was: Squeak practical use? ...)
by Richard A. O'Keefe 31 Jan '02

31 Jan '02

"Andrew C. Greenberg" <werdna(a)mucow.com> wrote: Not only that -- how, exactly, is this continuation example an example of a tail recursion? Certainly, its been decades since my A-exams, but as I recall, a tail recursion is generally defined as a recursive call in a tail context. No. TRO is a misnomer, it refers to tail CALL optimisation. The technique turns _any_ tail call into a jump, whether it is recursive or not. It's "procedure call = assignment + GOTO".

1 0

[ANN] call for participants: near-term Flow development
by Craig Latta 31 Jan '02

31 Jan '02

Hi all-- So, I'm away from home at the moment, but I'll be back on Monday. Let's see then how we might coordinate the next release of Flow (http://netjam.org/flow)... Please let me know if you'd like to be on a list for talking about this (and if you plan to actively discuss or lurk, either is fine :). Also please feel free to describe things you'd like to do or see addressed. thanks! -C -- Craig Latta composer and computer scientist craig.latta(a)netjam.org www.netjam.org crl(a)watson.ibm.com Smalltalkers do: [:it | All with: Class, (And love: it)]

3 2

Re: Tail recursion.
by Scott A Crosby 31 Jan '02

31 Jan '02

On Thu, 31 Jan 2002, Andrew C. Greenberg wrote: > > On Thursday, January 31, 2002, at 02:19 AM, Scott A Crosby wrote: > > >> How, exactly, can I safely and in straightforward manner optimize this > >> for any Class, when the definitions of foo1 and foo2 may well depend > >> upon definitions in subclasses, which in turn may change? Moreover, > >> since these kinds of expressions can be horribly complex, I would HATE > >> if the debugger lost contexts when showing me the problem. > > > > Actually, its not all that bad to implement. Tail recursion can be done > > by > > replacing your own context with a new one.. > > > > IE: > > > > foo > > blah blah > > ^self bar: 12 > > > > -- > > > > What happens is when foo sees the `^bar: 12' at the end, it replaces its > > method context with the one for `^bar: 12'. (or, equivalently, it pops > > its > > stack frame. and pushes the new one for ^bar.) > > How, exactly, do you intend to do this, with optimization, without > massively rewriting the compiler? Assuming bar is inherited, we can't One bytecode. call it a #tailSend:. Basically, it sets up the methodcontext, then 'moves' it down to replace the current methodcontext. As such, it doesn't care about the function being invoked, just the invoker. > optimize this code without recompiling foo everytime we change a bar, > whether in this class or a superclass. You don't need to change the function you're invoking. the function #bar: that gets invoked doesn't need to change and can be written as normal. IE: Let slook at the stack, normally, it'd go: PARENT >> foo PARENT >> foo >> bar PARENT >> foo >> bar >> (stuff in Bar) ..... PARENT >> foo >> bar >> (stuff in Bar) PARENT >> foo >> bar PARENT >> foo PARENT Here, we make the tailcall of bar instead: PARENT >> foo PARENT >> bar PARENT >> bar >> (stuff in Bar) .... PARENT >> bar >> (stuff in Bar) PARENT >> bar PARENT >> bar PARENT Which we do by replacing foo's methodcontext *completely* with bar's methodcontext, thus, for all intents and purposes, bar will think that its been invoked by PARENT, and return its result there. This change only requires a new bytecode that replaces: send: returnTop with a single one called tailSend: which does the above optimization. foo ^self bar: self. -- 9 <70> self 10 <70> self 11 <E0> send: bar: 12 <7C> returnTop In *either* case, the code for #bar: remains completely unchanged. What changes is the last two bytecodes in #foo. Since I don't care about #bar when doing this, it doesn't matter if its inherited or where it came from. It can be a primitive, inherited, or even changed. BTW, assume we have something like: foo: n ^ self bar: 2*n bar: n ^ self baz: n+1 baz: n ^ n Then, using tailcall, the stack looks like: PARENT PARENT>>foo PARENT>>foo>>* PARENT>>foo PARENT>>bar PARENT>>bar>>+ PARENT>>bar PARENT>>baz PARENT -- All in all, it should be an easy optimization. The only complexity there might be would depend on exactly how it builds methodcontexts... I don't know; I'm not planning on learning that till I see the new BC VM. Scott

1 0

Re: A Tool for managing modules
by Bergel Alexandre 31 Jan '02

31 Jan '02

On Thu, Jan 31, 2002 at 12:37:43PM -0800, Ned Konz wrote: > On Thursday 31 January 2002 08:20 am, Bergel Alexandre wrote: > > > And I tried to import my Connectors change set (after changing all the > > > category names to be 'People-NedKonz-Connectors'). > > > > > > And I imported the change set. When it executes the postscript of the > > > change set, I get a walkback on a doesNotUnderstand:. > > > > I have tried to reproduce this behavior, but I couldn't. Where can I get > > your Connectors change ? > > You can get the change set I was using at: > http://nedkonz.dhs.org:8080/Ned/31 I have tried to install it, on 4653 (3.2gamma) and 4689 (module so), neither is succesful for fileIn it. By the way, what is StarSqueakPlugin on your wiki ? (-: Regards, Alexandre

2 1

RE: [INFO] Chess on 3.2
by Torsten.Bergmann＠phaidros.com 31 Jan '02

31 Jan '02

>I have just downloaded the Squeak 3,2 alpha #4646 image and I found a Chess Board in it!! >Fantastic!! Where I can find more information about the algorithms used for >programming it? >Who is the author? Andreas Raab did it.

1 0

RE: A Tool for managing modules
by Torsten.Bergmann＠phaidros.com 31 Jan '02

31 Jan '02

>By the way, what is StarSqueakPlugin on your wiki ? (-: It's a plugin for running the ant colony and other StarSqueak simulations faster. See StarSqueakPlugin class and see the objects tool in your morphic world menu (cmd-O, then click on star squeak and grab one of the star squeak morphs). Use the morphs halo menu (the red one) to influence/start/stop the simulation. I actually play with the ModuleTool on Win32. Thx for it. Is it possible to send a short note to the list if you update the changeset on your website ? Has someone be able to track down the error of evaluating the code twice before getting it into the image. Keep on talking small Torsten

1 0

← Newer
1
2
3
4
5
6
7
...
62
Older →

Jump to page: