Squeak-dev December 2008

squeak-dev@lists.squeakfoundation.org

91 participants
114 discussions

worst crash yet trying to load stuff from Packages Universe in 3.10.2....
by Greg A. Woods; Planix, Inc. 16 Dec '08

16 Dec '08

OK, one more try I thought..... I took a fresh 3.10.2-7179 image, set my preferences, opened up the Packages Universe, carefully selected the packages I wanted, being extra careful to avoid the ones that I've had problems with already, and started the installs...... Things looked like they had finished fine, and so I tried clicking on the Packages Universe window again to see what it might say and KABOOM! (this all retyped by hand) ***System error handling failed*** Original error: MessageNotUnderstood: LabelMorph class>>contents:. Debugger error: MessageNotUnderstood: LabelMorph class>>contensts:: [] in Debugger class>>openOn:context:label:contents:fullView: {[:ex | self primitiveError: 'Original error: ', title asString, ', Debuge...]} BlockContext>>valueWithPossibleArgs: [] in MethodContext(ContextPart)>>handleSignal: {[(self tempAt: 2) valueWithPossibleArgs: {exception}]} blockContext>>ensure: MethodContext(ContextPart)>>handleSignal: MessageNotUnderstood(Exception)>>signal LabelMorph class(Object)>>doesNotUnderstand: #contents: UIThemeSoftSqueak(UITheme)>>buttonLabelForText: UIThemeSoftSqueak>>buttonLabelForText: UIThemeSoftSqueak(UITheme)>>buttonLabelFor: PluggableButtonMorph>>newLabel PluggableButtonMorph>>label: [] in Debugger>>buttonRowForPreDebugWindow: {[:quad | aButton := PluggableButtonMorph on: aDebugWindow getState:...]} OrderedCollection>>do: Debugger>>buttonRowForPreDebugWindow: Debugger>>buildMorphicNotifierLabelled:message: Debugger>>openNotifierContents:label: [] in Debugger class>>openOn:context:label:contents:fullView: {[debugger := self new process: process controller: controller co...]} BlockContext>>on:do: [] in Debugger class>>openOn:context:label:contents:fullView: {[[debugger := self new process: process controller: controller co...]} --------------------------- Type CR to enter an emergency evaluator. Type any other character to restart. Sadly it doesn't look like there's even any obvious hint as to what package might have caused the problem. I'm beginning to think this is a hopeless cause. Clearly not even a majority of packages in the currently published universe for 3.10.2 have ever actually been tested. -- Greg A. Woods; Planix, Inc. <woods(a)planix.ca>

11 30

Re: [squeak-dev] I would like to create a new team (actually "old" project because I'm already working on it :) )
by Eliot Miranda 16 Dec '08

16 Dec '08

On Tue, Dec 16, 2008 at 12:53 PM, Jecel Assumpcao Jr <jecel(a)merlintec.com>wrote: > Eliot Miranda wrote: > > in response to Janko message about teams I would like to create two > new new teams called: > > Faster Execution Engine > > I am interested in joining this one. > Welcome aboard! > > and Faster Interoperability Engine > > (suggestions as to what FUM could stand for gratefully received). > > Fast Unloadable Modules > > > Yes, you smell the blood of an Englishman. > > Bryce? I am sure he could help with this :-) Yes, antipodeans have a particularly sharp nose for Englishmen ;) But yes, Bryce would be very welcome. > > > -- Jecel > > >

1 0

3.11 Process A) Small Fixes
by Keith Hodges 16 Dec '08

16 Dec '08

Contributions A) Small fixes ========== 1. Bugs are reported, discussed and fixes proposed on mantis. 2. When a fix is ready for harvesting a small installer script is appended as a note to the bug report. This script indicates exactly which files are required in order to fix this issue. example script: "fix begin" Installer mantis bug: 1234 fix: 'SomeFile.1.cs.gz'. "(.gz support is fixed)" "fix end" 3. Bugs will be nominated by those providing the fix for inclusion in the unstable build. We are still working out the details as to how this will be carried out. But it will be possible within Mantis. 3b. The build process tasks will need to be informed of any explicit dependencies between fixes if there are any. As long as it knows about them it will honour them. 4. The Unstable build will be automatically built (and tested) on a periodic basis. The fixes are loaded into changesets with the bug fix number prepended to the changeset name. Also the full mantis report is copied into the image for documentation as to what has changed. 5. All packages effected by the most recently added fixes are written to Monticello, in such a way as the history for any one Monticello package makes it look like that package was incrementally fixed. In reality the build process starts from scratch every time. By doing it this way we successfully combine the sensible way of fixing via changesets with the desired way of managing code in packages. Keith

3 5

I would like to create a new team (actually "old"
by Eliot Miranda 16 Dec '08

16 Dec '08

Hi All, in response to Goran's message about teams I would like to create two new new teams called: Faster Execution Engine and Faster Interoperability Engine (suggestions as to what FUM could stand for gratefully received). Yes, you smell the blood of an Englishman. I would like to be team leader. The tasks are: Faster Execution Engine: produce an execution engine that supports closures and provides performance broadly comparable with other Smalltalk JIT virtual machines, of the order of 10 to 20 times faster than the current VM for benchFib. Faster Interoperability Engine: produce a more flexible and faster FFI based on and integrated with the execution technology in FEE. best Eliot

1 0

SoC and DOMParser?
by tjohnson＠iwu.edu 16 Dec '08

16 Dec '08

Hi, Did anything ever happen with the Summer of Code project to create a rendering UI on top of the HTML DOMParser? Thanks, TimJ

2 1

FractalMorph 1.2 won't load into 3.10.2-7179
by Greg A. Woods; Planix, Inc. 16 Dec '08

16 Dec '08

I get this while trying to load FractalMorph 1.2 from the Package Universe browser: PluggableTextMorph(Object)>>doesNotUnderstand: #hideScrollBarIndefinitely FractalMorph>>initButtons FractalMorph>>initialize FractalMorph class(Behavior)>>new UndefinedObject>>DoIt Compiler>>evaluate:in:to:notifying:ifFail:logged: Compiler class>>evaluate:for:notifying:logged: Compiler class>>evaluate:for:logged: Compiler class>>evaluate:logged: [] in MultiByteFileStream(PositionableStream)>>fileInAnnouncing: {[val := (self peekFor: $!) ifTrue: [(Compiler evaluate: self nextChunk l...]} BlockContext>>on:do: [] in MultiByteFileStream(PositionableStream)>>fileInAnnouncing: {[:bar | [self atEnd] whileFalse: [bar value: self position. self skipS...]} [] in ProgressInitiationException>>defaultMorphicAction {[result := workBlock value: progress]} BlockContext>>ensure: ProgressInitiationException>>defaultMorphicAction ProgressInitiationException>>defaultAction UndefinedObject>>handleSignal: MethodContext(ContextPart)>>handleSignal: ProgressInitiationException(Exception)>>signal ProgressInitiationException>>display:at:from:to:during: -- Greg A. Woods; Planix, Inc. <woods(a)planix.ca>

19 53

#saveOnFile won't do
by Aidan Gauland 15 Dec '08

15 Dec '08

Hello, I'm working on a game engine, and I found the #saveOnFile method in the Object class, which I can use for saving the game state, but I also want to do exactly the same thing for storing the initial game state (loaded when the user starts a new game) with one exception: instead of writing to a file, I want to store it internally (in the image). I can't set up the game programatically, because I use the Object Explorer for building the game (which works very well for this purpose, I might add; you don't get that with other programming languages). Is there a way I can do this? Thanks, Aidan P.S. Please let me know if I should have asked this on the newbies list; I've only been posting there, until now.

3 4

what assembler should I use?
by Eliot Miranda 14 Dec '08

14 Dec '08

Hi All, I now have an x86 simulator integrated into Squeak that can execute code form a ByteArray, Bitmap etc. So I'm now ready to write the JIT code generator. I want a simple framework that can convert from an abstract register-transfer-level instruction set to a concrete machine code, x86/IA32 in the first instance. I wonder what frameworks there are that I can, um, exploit. The end result must be Smalltalk code that can generate machine code and can be translated via Slang to C. I've already extended Slang somewhat for the Stack VM so that simple classes can be converted to structs. But the Slang technology I have is still a long way from translating arbitrary Smalltalk to C, so when you think "Smalltalk framework" think "Limited Smalltalk/C hybrid framework". I know that Exupery has an assembler in it but it is quite high-level and oriented to textual assembly whereas my JIT will have no assembler level in it, just stack contents abstractions (to model receiver, arguments, temporaries, intermediates, constants, etc), register transfer level and concrete machine code. I know Ian has written the exact framework I need but in C. Are there any other candidates? Ian, how difficult would it be to translate your framework (ccg?) into Smalltalk so that it can be translated back into C via Slang? My guess is not too hard. Have you got a pointer to an up-to-date ccg? Also, what other ISAs does it target? best Eliot

3 2

Smalltalk, the Free Software spirit in action
by Hilaire Fernandes 14 Dec '08

14 Dec '08

Some advocacy for Smalltalk... http://blog.ofset.org/hilaire/index.php?post/2008/12/12/Smalltalk-the-free-… In this article, I explain how Smalltalk is since 30 years a practical social implementation of the freedom #3 and #4 of the Free Software philosophy: freedom to study, to modify and to redistribute a software. -- http://blog.ofset.org/hilaire

2 2

Re: [squeak-dev] Kay's Little Blue Book, or, Let's Just Do It! (was Re: The Gospel of Alan)
by Ties Stuij 14 Dec '08

14 Dec '08

On Thu, Nov 27, 2008 at 3:54 AM, Jecel Assumpcao Jr <jecel(a)merlintec.com> wrote: > Tony Garnock-Jones wrote: >> Bert Freudenberg wrote: >> > Just a meta remark - I find it highly amusing how people dissect the >> > Gospel of Alan, even interpreting it literally. He must get quite a >> > chuckle from that ;) >> >> Well I hope so! :-) > > Given how happy he said he was to not have any disciples, he might be > less than amused to find out otherwise ;-) > > Going to the other extreme, where everyone's opinion is equally valid, > leads to situations like a guy explaining to me on comp.lang.lisp that > multiple dispatch is the most important feature of OOP and so CLOS and > C++ are true OO languages but Smalltalk is not. > >>http://groups.google.com/group/comp.lang.lisp/browse_thread/thread/b2aa1842… I know I shouldn't react, but could you just dissect for me how Alan's post, either the one before or the one after the post of yourself you linked here, makes the point you're suggesting him to be making? Is it the 'I've no idea where I'm going with this.' part perhaps? /Ties

3 3

← Newer
1
...
4
5
6
7
8
9
10
11
12
Older →