Squeak-dev April 2004

squeak-dev@lists.squeakfoundation.org

172 participants
488 discussions

Mac VM 3.7.1b2 (for testing)
by John M McIntosh 04 Aug '04

04 Aug '04

I've pushed a Mac VM 3.7.1b2 for testing to http://homepage.mac.com/johnmci/ Squeak 3.7.1Beta2.app.sit It contains: 3.7.1b2 a) JPEG READER WRITER Usage of Apple '-fast' uncovered a source code issue, initialization routines for internal jpeg plugin were not returning true or false, before this was kinda of considered as returning true? But when compiled with -fast it's zero, meaning the inialization failed, but that is bogus. b) Use of "-fast" apple gcc 3.3 compiler option, improves sends by significant % c) Broke high bit ascii macroman in 3.7.0bx. On EventKeyUp/EventKeyDown you get the UniCode On EventKeyChar you now get the UniCode->MacRoman translated character. For multi-byte character input you get EventKeyDown/EventKeyChar/EventKeyUp for each unicode character coming from the input sequence. Please note that I've a 3.7.1b3 pending due to window open issue in os-x 10.3.2, but please try this VM for speed/performance and correctness. OS-9 Users please note I did not rev the os-9 binary because these are os-x specific issues being fixed at this time. -- ======================================================================== === John M. McIntosh <johnmci(a)smalltalkconsulting.com> 1-800-477-2659 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===

3 2

[BUG][FIX] WeakGCFix-wbk
by bryce＠kampjes.demon.co.uk 03 Aug '04

03 Aug '04

from preamble: "Change Set: WeakGCFix-wbk Date: 21 March 2004 Author: Bryce Kampjes Weak references can leave dangling pointers if the weak object is in old space and it points to a new space object. The problem occurs when the new space object is incrementally collected but the old space weak object remains. The weak object now has a dangling pointer to where the new space object was. This change set adds a little code to ObjectMemory>>sweepPhase so that it will finalise any weak rootTable objects. There is still the chance that an interpreter variable may contain a pointer to a weak object. However interpreter variables have very specific uses. I'm not sure if they should be checked for weak references. "!

5 14

[ENH] ExceptionDialogs-avi
by avi＠beta4.com 28 Jul '04

28 Jul '04

from preamble: "Change Set: ExceptionDialogs-avi Date: 31 August 2003 Author: Avi Bryant Adds Notifications for #inform:, #confirm:, and #confirm:orCancel:. By default these still use PopUpMenu, but they could be caught and dealt with more appropiately in different UI contexts."!

2 1

[ENH] ReUseViewerFlaps-nk
by ned＠bike-nomad.com 28 Jul '04

28 Jul '04

from preamble: "Change Set: ReUseViewerFlaps-nk Date: 21 September 2003 Author: Ned Konz This is a minor change for the eToys environment that will re-use an existing viewer flap when you open a viewer from an object halo. "!

2 1

[BUG] Cannot open files in a Mac volume's root
by Bert Freudenberg 28 Jul '04

28 Jul '04

The culprit seems to be "MacFileDirectory class>>isAbsolute:" which answers false if no colon is found inside the file name - which is the case for a volume name. To reproduce: Open a file list, select a volume (scroll all the way up in the directory pane), and select a file in that volume (not in a subdirectory). You may have to buy a Mac first and install Squeak. - Bert -

6 8

[bug] stack overflow crashes Squeak.
by Michael van der Gulik 06 Jul '04

06 Jul '04

Hi all. If you make a naive class that calls itself: Forkbomb>>forkbomb ^ self forkbomb. Squeak 3.5 on Debian linux will eventually crash, leaving a huge stack dump in the terminal. Okay; it's really my fault for making such a stupid class, but I'm playing around with proxied objects, message sends etc and it's happening quite often while I'm debugging. Is there a way to limit the invocation stack to a sane level of deepness, after which a process/thread (whichever is Squeak terminology) gets suspended and an error thrown? Same probably applies to number of processes made by a thread. mikevdg.

12 19

The Magic Book - a design draft
by goran.hultgren＠bluefish.se 30 Jun '04

30 Jun '04

The idea of a "Magic Book" inside Squeak acting as both an introductory documentation for Squeak and as a reference manual is quite intriguing. A few basic ideas that got this started: - Tying unit tests to documentation in order to detect stale documentation - Not bloating the image - Easily maintain the book cooperatively (like swiki) Ok, what would be the STTCPW? (For all you non XP aware: "Simplest Thing That Could Possibly Work"). By applying reuse - what do we already have? We already have: - Projects: A very good format for "active content". I am not sure how easily they are backwardscompatible though. Could anyone post some insight in that matter? - StarBrowser: A universal UI kindof thing. We could easily let StarBrowser be the main UI of the Magic Book! It lends itself easily for the task - a treeview on the left, a dynamic area to the left etc. If we somehow find a good way to integrate a little search form then it would be quite perfect. - SqueakMap: In whichever form we choose to have the Chapters of the book (Projects) we could register them on SM (a few new categories would be nice) and thus get a distributed always up to date model of chapters for free. And when SM1.1 gets completed it will track versions of the chapters for us, etc. And in SM1.1 it will have a sensible smart diskcache for the packages (read Chapters) so we get that too. - Free text engine: Scott Crosby wrote a neat little engine for maintaining keyword indices that the book obviously should use. I have also been intending to integrate that into SM. We should be able to add it to SM in such a way that the Magic Book can simple get that "for free". So... we already have a bunch of cool technology to implement the Magic Book IMHO. We just need to knit it all together in a smart way. If we split the book in the following pieces: 1. The TOC. This is essentially a list of chapters organized in a hierarchy with names and a link to the corresponding SM package that is the chapter content - a Project. 2. The chapters themselves. These are simply Projects published as packages on SM. Perhaps we need to follow a few conventions for these Projects so that we easily can "jump into" them in a proper manner, extract keywords for indices etc. 3. The search indices and the index (as books have last). If 1 and 2 above also are packages on SM (in whichever suitable form, doesn't matter) it will be even more simplified. So the book *contents* are then managed as packages on SM. We get the possibility of having different owners for different chapters, release/version tracking etc. It is obvious when you think about it - why shouldn't they be packages? The StarBrowser is the obvious choice of a UI - it already has navigation mechanisms and it can also show contents from the book along with other things in a highly dynamic fashion. Why reinvent that? And SqueakMap gives us a lot of stuff like efficiently always updated content, a smart cache etc. And there will be no image bloat whatsoever. Well, ok, what do you all think? Darn, this sounds like real fun... I need to get cranking at my SM1.1 ... :-) regards, Göran

2 1

[ENH][FIX] StringMorphFromTextFix-nk
by ned＠squeakland.org 25 Jun '04

25 Jun '04

"Change Set: StringMorphFromTextFix-nk Date: 26 February 2004 Author: Ned Konz Does these things: - lets StringMorph absorb text color, font size, and font style from Text instances passed to #contents: (in addition to emphasis). - makes Text>>asMorph return an unwrapped TextMorph instead of a StringMorph. - adds asStringMorph and asTextMorph for when you want to be more specific about what you get. " <This post brought to you by BFAV2>

2 1

[FIX] decompFixes-ls
by Lex Spoon 25 Jun '04

25 Jun '04

from preamble: "Change Set: decompFixes-ls Date: 30 January 2004 Author: Lex Spoon Fixes the decompiler bug posted by Ned Konz on 11 February 2003. The issue is that if the last case in a case statement returns from the method, the decompiler gets confused trying to parse the otherwise: branch. This changeset has the decompiler track all non-return exits from the case statement, so that it has enough information to do a right thing. The changeset also adds a couple of methods that the decompiler is relying on. After loading this changeset, Ned's example decompiles correctly. Also, decompiling and recompiling everything leaves a running system. The test is included for future use as [Decompiler recompileAllTest]. s0: evt evt caseOf: { [ #Entry ] -> [ self log: 's0-ENTRY;'. ^nil ]. [ #Exit ] -> [ self log: 's0-EXIT;'. ^nil ]. [ #Init ] -> [ self log: 's0-INIT;'. self initProtected: #s1:. ^nil ]. [ #e ] -> [ self log: 's0-E;'. self tran: #s211:. ^nil ]. } otherwise: [ ^#top: ]"!

2 1

[FIX] RefStrmObsolCls-tk
by ducasse＠iam.unibe.ch 25 Jun '04

25 Jun '04

"Change Set: RefStrmObsolCls-tk Date: 19 September 2003 Author: Ted Kaehler V2: stef: Changed Smalltalk by SystemNavigation default at the right place please check. It is really too late. SmartReferenceStrm is in charge of converting reading in old .pr files and taking care of instances of classes that have changed shape. It also handles obsolete classes. It gives the user a chance to name a new class that old instances will be converted to. In 2000 Bob Arning made major changes, but these were never extended to converting the instances of obsolete classes. This changeSet fixes that."

1 1

Jump to page: