Squeak-dev December 2000

squeak-dev@lists.squeakfoundation.org

96 participants
272 discussions

RE: [BUG] 2.9a, global flaps got whacked?
by Norton, Chris 17 Dec '00

17 Dec '00

Hi Bob & company, Unfortunately, I don't have that image handy at home and recently I learned of a death in my family, so I may not be at work any time soon. But I can tell you that I was working from one of Bruce O'Neil's releases of the Windows image (maybe his last 2.9a release?). And I regularly update my changes when Dan Ingalls releases them. So working back from Dan's update notes will probably pinpoint where/when this problem got into the build. I believe the problem is related to the AlignmentMorph changes, because they happened shortly before I noticed the problems. I can approximate how I found the problem. Fortunately, since this is my test pilot image, most of the code is baseline stuff. This image does not include much code of my own that was not posted to the list. The changes I made to this image are trivial, like setting the browser colors to white. 1. Create your basic Morphic project. 2. Turn on Global Flaps. 3. Save and exit 4. Launch again and update from server (up to the revs in Dan's e-mails) 5. After each update, save and exit. Sooner or later, you'll find that the menu tab in your Morphic project gets all funny looking (the menu locations within the tab get screwed up). Hope that was helpful! ---==> Chris

1 0

Re: Squeak on iPaq..success!
by Bob Arning 17 Dec '00

17 Dec '00

Hi Kevin, On Sun, 17 Dec 2000 10:27:59 -0500 (EST) Kevin Fisher <kgf(a)golden.net> wrote: >A couple notes: > >The majorShrink seems to have had some unpleasant side-effects. When I >save the shrunken image I get an error box complaining about a missing >seperator. Still, it saves the image. That would seem to be an effect of changes you have made to the shrink process. I don't get it in a normal image. >As I briefly mentioned, using "Smalltalk lastRemove" seems to hang the VM. >If you cmd-. you can break out of it..it seems to be hung in a Semaphore. Did you do "Smalltalk lastRemoval" manually? Or is it in modified code you have? It does not seem to be run as a part of the normal process in 2.9a. I did note that the rehashing of MethodDictionaries can take a VERY long time (on the order of 10 minutes @ 400 MHz). I have a fix that deals with that and will put it out after some further testing. Cheers, Bob

1 0

Re: The Puzzle Page
by Bob Arning 17 Dec '00

17 Dec '00

On Sun, 17 Dec 2000 12:44:40 +0100 Karl Ramberg <karl.ramberg(a)chello.se> wrote: >Another bug is that when painting with the paint tols >it allways paints on unlimited >canvas overriding the preference. Hi Karl, I'm not seeing this one. How big is your display and what does the following method look like on your system? Preferences class>>defaultPaintingExtent "Answer the preferred size for the onion-skin paint area when launching a new painting within a paste-up morph. Feel free to change the parameters to suit your configuration." ^ 800 @ 600 Cheers, Bob

1 0

Re: The Puzzle Page
by Bob Arning 17 Dec '00

17 Dec '00

Hi Kar, On Sat, 16 Dec 2000 22:42:18 +0100 Karl Ramberg <karl.ramberg(a)chello.se> wrote: >If I have a viewer on the world the little world flap blinks every second. That's more or less to be expected. The flap is trying to shot an up-to-date picture of what the viewee looks like. Since it can't do that in a more general way, it just updates periodically. >When I open a viewer flap, sometimes 2/3 of the screen flashes, instead >of just the flap part. Is there anything else in the flashing area? One thing the DamageRecorder does is to merge intersecting rectangles and sometimes, when there are too many, it even merges non-intersecting ones to keep the number down. Knowing what else is in that area may point us to ways to cut down the flashing. >When I pull a script off to edit it, the whole flap area flashes >when I move the script around. While you are moving the script? Or when you drop the script on the world? I see the latter, but not the former. The latter is a reasonable need on the part of the flap to make sure its contents are drawn correctly since dropping phrases and scripts in the world can cause changes in the flap. Cheers, Bob

2 1

RE: The Puzzle Page
by Raab, Andreas 17 Dec '00

17 Dec '00

Bob, > >When I pull a script off to edit it, the whole flap area flashes > >when I move the script around. > > While you are moving the script? Or when you drop the script > on the world? I see the latter, but not the former. The > latter is a reasonable need on the part of the flap to make > sure its contents are drawn correctly since dropping phrases > and scripts in the world can cause changes in the flap. I think Karl is talking about something else. After you've dropped something in the world (it doesn't really matter what) you'll get a flash for all of the flap (tab) areas. This happens because of the attempt of moving all those flaps and tabs to the front. So how about implementing PasteUpMorph>>addMorphFront: aMorph ^self addMorphInLayer: aMorph. This should prevent many of these (almost always unnecessary) redraws. Cheers, - Andreas

2 2

Re: [updates] A few more tweaks
by Bruce ONeel 17 Dec '00

17 Dec '00

Hi, These and the previous batches of changes are in: ftp://st.cs.uiuc.edu/Smalltalk/Squeak/2.9_alpha/Squeak2.9-3125.tar.gz Tell me if the .tar.gz file causes problems in Windows systems. WinZip does work ok with .tar.gz, right? I'll check it at work tomorrow. I'll be on vacation later this week until after the new year so this is likely to be the last upload for this year :-) cheers bruce Dan Ingalls <Dan.Ingalls(a)disney.com> wrote: > ... as follows... > ----------------------------- > 3122fixStories-raa -- Bob Arning -- 16 December 2000 > - make page transitions between camera marks in paino roll a property of the camera mark. Use the red menu on the *ending* mark to set the transition > - shift click on camera mark to go directly to that mark > > 3123FlapsFix-ccn -- Chris Norton -- 15 December 2000 > Fixes a small bug in the construction of global flaps. If the morph's world is nil, this method used to result in a walkback. This change seems to fix it." > > 3124moreLayers-raa -- Bob Arning -- 16 December 2000 > Enhance morphic layering with the addition of Morph>>addMorphInFrontOfLayer: which will insure that the morph in question appears in front of any morphs of the same priority. This will reduce unnecessary competition between layers since we will use this in PasteUpMorph>>addMorphFront:. The original Morph>>addMorphInLayer: will retain its current behavior of allowing morphs to be behind morphs of the same priority since this is necessary to prevent endless reshuffling by morphs which periodically reassert their position in the pecking order. > > 3125menuFixups-raa -- Bob Arning -- 16 December 2000 > improve the menu flap by > - setting appropriate layout parameters to keep things from getting jumbled > - making the menus sticky again so the don't come out accidentally. > Still pending: how to best consume any pending events on the mouseEnter: flop out so that they don't end up activating a menu. > There is a 'Utilities replaceMenuFlap.' in the postscript"

1 0

Re: Building 2.9a VM on StrongARM Linux
by Bruce ONeel 17 Dec '00

17 Dec '00

Hi, You can always just not compile that bit of code and then write little dummy prims to solve the link errors you will get. Actually, since this is a plugin not building it should be ok though it's been a while since I played with unix builds. This sort of error message probably means that you code is too big, ie, you have a if/while/?? with too much code in the middle. You can either: - comment out the code - put the bit inside the if/while/etc in a function and call that function. cheers bruce Kevin Fisher <kgf(a)golden.net> wrote: > > Kevin Fisher <kgf(a)golden.net> is widely believed to have written: > > > > > > > -c -o SoundCodecPlugin.o ../src/generated/SoundCodecPlugin.c > > > /tmp/ccFTDeUu.s: Assembler messages: > > > /tmp/ccFTDeUu.s:1394: Error: Can not represent OFFSET_IMM8 relocation in this > > > object file format (1) > > > /tmp/ccFTDeUu.s:1396: Error: Can not represent OFFSET_IMM8 relocation in this > > > object file format (1) > > > > Yup. It says the same on NetWinders (usually anyway) and points to a bug > > in the ARM toolchain somewhere. I _think_ Phil Blundell (sp?) is the > > main maintainer, but I'm sure that a suitable email address must be in a > > help message orHOWTO somewhere. Failing that you might get some luck > > through www.armlinux.org etc > > Ah! Drat...well, thanks for the pointer, I'll go bug the armlinux folks about > this. :) > > Alternately, is there any way to build the VM without sound? Configuring with > --disable-sound doesn't seem to work (nor does commenting out the OSS section > in sqUnixConfig.h).

1 0

Re: 2.9a majorShrink bug & shrunken Morphic
by Bob Arning 17 Dec '00

17 Dec '00

On Fri, 15 Dec 2000 22:35:00 -0500 Kevin Fisher <kgf(a)golden.net> wrote: >The image size is about 13megs, which isn't too bad, but the changes file is >huge! Kevin Is that *after* shrinking? My normal image is around 13.5 meg, so this doesn't sound like much discaarding has happened. You might reclaim some more space with: ChangeSet allInstancesDo: [:cs | cs zapHistory] and the changes file can be compressed with: Smalltalk condenseChanges Beyond that, you may need to look at Morphic with a keen eye - which parts do you want and which can you do without? Cheers, Bob

3 3

[updates] A few more tweaks
by Dan Ingalls 17 Dec '00

17 Dec '00

... as follows... ----------------------------- 3122fixStories-raa -- Bob Arning -- 16 December 2000 - make page transitions between camera marks in paino roll a property of the camera mark. Use the red menu on the *ending* mark to set the transition - shift click on camera mark to go directly to that mark 3123FlapsFix-ccn -- Chris Norton -- 15 December 2000 Fixes a small bug in the construction of global flaps. If the morph's world is nil, this method used to result in a walkback. This change seems to fix it." 3124moreLayers-raa -- Bob Arning -- 16 December 2000 Enhance morphic layering with the addition of Morph>>addMorphInFrontOfLayer: which will insure that the morph in question appears in front of any morphs of the same priority. This will reduce unnecessary competition between layers since we will use this in PasteUpMorph>>addMorphFront:. The original Morph>>addMorphInLayer: will retain its current behavior of allowing morphs to be behind morphs of the same priority since this is necessary to prevent endless reshuffling by morphs which periodically reassert their position in the pecking order. 3125menuFixups-raa -- Bob Arning -- 16 December 2000 improve the menu flap by - setting appropriate layout parameters to keep things from getting jumbled - making the menus sticky again so the don't come out accidentally. Still pending: how to best consume any pending events on the mouseEnter: flop out so that they don't end up activating a menu. There is a 'Utilities replaceMenuFlap.' in the postscript"

1 0

Re: [BUG] 2.9a, global flaps got whacked?
by Bob Arning 16 Dec '00

16 Dec '00

On Fri, 15 Dec 2000 18:45:50 -0500 "Norton, Chris" <chrisn(a)Kronos.com> wrote: >I updated my test pilot image to the latest last night and my pre-existing >global flaps got wacky. Specifically, the menu flap seems to be quirky (all >of the menus got sort of mashed together into a new spot). Hi Chris, Could you tell me the latest change set in your image *before* you updated? I.e., what was the last change set where the menu flap was OK? Cheers, Bob

1 0

← Newer
1
...
7
8
9
10
11
12
13
...
28
Older →

Jump to page: