Squeak-dev May 2002

squeak-dev@lists.squeakfoundation.org

160 participants
453 discussions

Re: Squeak lockup
by Bill Schwab 31 May '02

31 May '02

Peter, Have you read the info on the Wiki about the CPU monitor etc.? The only remotely long-duration sessions I've done with Squeak have been (3.2 gamma, I think) on Win2k for a couple of days at a time, and it's done well. I have an intermittent (but persistent) problem with a lockup on Win95 in a MIDI app; that can take seconds or hours to show up, but, the VM seems to keep running, the Morphic UI freezes. I'm slowly shifting hardware around in hopes of trying the same thing on another computer and OS without "destroying" the machine with the reproduceable problem. The short version is that I'm very interested in seeing how one debugs such things in Squeak. To date, I've gotten crash dumps from it by enabling 3D hardware acceleration on a particular machine; the results leave me confident that if something similar happened in my code, I'd at least know where to start looking for the problem. So far though, I've not been able to wear down the lockup. Bill Wilhelm K. Schwab, Ph.D. University of Florida Department of Anesthesiology PO Box 100254 Gainesville, FL 32610-0254 Email: bills(a)anest4.anest.ufl.edu Tel: (352) 846-1285 FAX: (352) 392-7029

3 2

Squeak Attack image
by Torsten.Bergmann＠phaidros.com 31 May '02

31 May '02

Hi, I've found this "Squeak Attack" image on the web: http://www.celebritydesktop.com/movies/link.cgi?ID=516 Have fun Torsten

3 3

[BUG] FileUrl
by olmeca＠netscape.net 31 May '02

31 May '02

I ran into the following bug. If you evaluate the following code: | dir file | dir := ServerDirectory new fullPath: (FileDirectory default url). file := dir asServerFileNamed: 'test.txt'. Transcript show: (file fileNameRelativeTo: dir); cr. You'd expect to get on Win32: D:\Users\angela\Squeak\test.txt and on MacOS: Mars:Users:angela:Squeak:test.txt But in both cases there is an additional leading '/'. The code that causes this is in class FileUrl, method #pathForDirectory: ^ String streamContents: [ :s | isAbsolute ifTrue: [ s nextPut: $/ ]. 1 to: self path size - 1 do: [ :ii | s nextPutAll: (path at: ii); nextPut: FileDirectory default pathNameDelimiter ] ] If the FileUrl object is absolute then a '/' is prepended, independent of the computer platform. I'd the slash should only be prepended on platforms that start an absolute path with a slash. I don't know how to perform platform tests in Squeak, so I can't propose a solution. Anyone? Rudi Angela __________________________________________________________________ Your favorite stores, helpful shopping tools and great gift ideas. Experience the convenience of buying online with Shop@Netscape! http://shopnow.netscape.com/ Get your own FREE, personal Netscape Mail account today at http://webmail.netscape.com/

1 0

[ENH] ClassDefinitionChangeList
by Kazuki MINAMITANI 31 May '02

31 May '02

Hi, I developed "ClassDefinitionChangeList", which keeps every versions of class definition. If you choose "versions" from SystemBrowser's class pane, you can see previous versions of selected class. ScreenShot is here: http://www5.ocn.ne.jp/~minami/squeak/classDefinitionChangeList/index. html Enjoy! -- Kazuki MINAMITANI e-mail: k-minami(a)joy.ocn.ne.jp URL: http://www5.ocn.ne.jp/~minami

1 0

Re: [ENH] workspaceDropPref-ls
by danielv＠netvision.net.il 31 May '02

31 May '02

Aahh... that is nice. Joshua 'Schwa' Gargus <schwa(a)cc.gatech.edu> wrote: > On Fri, May 31, 2002 at 11:38:02AM +0300, danielv(a)netvision.net.il wrote: > > What exactly do you mean by "perform operations on multiple morphs from > > the same code"? > > > > You mean you can address the morphs as variable from the workspace, like > > - > > myCelesteSystemWindow model close > > ? > > Yes. Although it gave me the broken name 'system browser1973'. > I will provide a fix. > > > > > If so, Cool, how do I do this? > > Click on the menu icon of a Workspace (next to the 'X' close button). > Select 'create textual reference to dropped morphs'. Drop a morph on > the workspace. > > If you want to do it with SystemWindows, you should first ensure that > the preference 'systemWindowEmbedOK' is on (SystemWindows are not > usually considered as drop candidates). Also, if you have the > preference 'fastDragWindowForMorphic' turned on, you will have to drop > the window using the black 'grab' halo handle. > > Wait for that fix... > > Joshua > > > > > > Daniel > > > > Lex Spoon <lex(a)cc.gatech.edu> wrote: > > > > > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A) > > > Content-type: TEXT/PLAIN > > > Content-transfer-encoding: 7BIT > > > > > > Workspaces accepting drops is very nice. It lets you perform operations > > > on multiple morphs from the same code window, unlike with an inspector. > > > It's not as nice as Morphic Wrappers, but it's pretty good and it's > > > already present in the image. This preference makes the feature more > > > convenient for those who like it. > > > > > > In fact, I wonder if anyone would want to turn this preference back off > > > after turning it on? The only thing you lose is the ability to easily > > > place morphs over workspaces. Perhaps we could just turn this on and > > > not make it a preference any longer? > > > > > > > > > from preamble: > > > > > > "Change Set: workspaceDropPref-ls > > > Date: 29 May 2002 > > > Author: Lex Spoon > > > > > > Add a preference for the acceptDroppedMorphs property of workspaces. If > > > you like this feature, then you will probably like to have it enabled > > > for all workspaces by default." > > > ! > > > > > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A) > > > Content-type: application/octet-stream; NAME=workspaceDropPref-ls.cs.gz > > > Content-transfer-encoding: base64 > > > Content-disposition: attachment; filename=workspaceDropPref-ls.cs.gz > > > > > > H4sIAAAAAAAAAHVSTW/bMAw9+6D/wGSHrMDi5qMBWt+CFcU6tECBDNhhGAbapmMtsqRKctP0 > > > 149y3cSH7GaTT+89PnJy50wDm+eWcLdMl6hsjWAqWM7gO+oW3QEWs9kCfikM5AO0tuSPDD5d > > > XV/Pf4PRsLiBR+xRGGCerVbZ1QpsMxmJ8dca9ZZgQyFLkr1xO2+xoFtn7JOjaqq8uI10SZIM > > > aMS6DbVxsfpAr7Cxxmgh1mUJCJafkSNdEFTGQagJsCjIhshpqXw0ztaeYcaSC4c4ylHWpwD3 > > > FRxMC0ruiB9LDxVhaB19iVS66+2lUpEgx1wdeqSBGl8IZADSXKayU0cGntghP0BJFbYqpGMx > > > EkKMfn40oSEeqfR3PBZMpJZBopJvNHWkCD1NwAdsLPeUh9Xl4uayy3O+zOYrzvH0QIjEtzwa > > > nEqpSM5F8Aeejln5gc31EevTEbDNMf4v2V4jegsUozyjMxZDGaY6/fKVnJWFgum2xknyDPmc > > > O7P3Um8vRJ9eBhUqT5Bzvrz6b6RiMLy5Pvt+V4Pku5UVjqLNQK+hZdMDV7y+8t0zNJ1pPoQf > > > tWGJAahA/c6bE7SeoVJDaaYyRHo+la4LcV8vqMNJPZ0IPpa/VIQHU6D6MF90l3+vOcuGOAot > > > VV/bMEUR4iFwLeUF/APFmQ9chAMAAA== > > > > > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A)--

1 0

[FIX] DroppedMorphNaming
by schwa＠cc.gatech.edu 31 May '02

31 May '02

from preamble: "Change Set: DroppedMorphNaming Date: 31 May 2002 Author: Joshua Gargus Morphs with names containing non-alphanumeric characters were given broken names when dropped on workspaces. This makes sure that the name that morphs are bound to contains only alphanumeric characters."!

1 0

Re: [ENH] workspaceDropPref-ls
by danielv＠netvision.net.il 31 May '02

31 May '02

What exactly do you mean by "perform operations on multiple morphs from the same code"? You mean you can address the morphs as variable from the workspace, like - myCelesteSystemWindow model close ? If so, Cool, how do I do this? Daniel Lex Spoon <lex(a)cc.gatech.edu> wrote: > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A) > Content-type: TEXT/PLAIN > Content-transfer-encoding: 7BIT > > Workspaces accepting drops is very nice. It lets you perform operations > on multiple morphs from the same code window, unlike with an inspector. > It's not as nice as Morphic Wrappers, but it's pretty good and it's > already present in the image. This preference makes the feature more > convenient for those who like it. > > In fact, I wonder if anyone would want to turn this preference back off > after turning it on? The only thing you lose is the ability to easily > place morphs over workspaces. Perhaps we could just turn this on and > not make it a preference any longer? > > > from preamble: > > "Change Set: workspaceDropPref-ls > Date: 29 May 2002 > Author: Lex Spoon > > Add a preference for the acceptDroppedMorphs property of workspaces. If > you like this feature, then you will probably like to have it enabled > for all workspaces by default." > ! > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A) > Content-type: application/octet-stream; NAME=workspaceDropPref-ls.cs.gz > Content-transfer-encoding: base64 > Content-disposition: attachment; filename=workspaceDropPref-ls.cs.gz > > H4sIAAAAAAAAAHVSTW/bMAw9+6D/wGSHrMDi5qMBWt+CFcU6tECBDNhhGAbapmMtsqRKctP0 > 149y3cSH7GaTT+89PnJy50wDm+eWcLdMl6hsjWAqWM7gO+oW3QEWs9kCfikM5AO0tuSPDD5d > XV/Pf4PRsLiBR+xRGGCerVbZ1QpsMxmJ8dca9ZZgQyFLkr1xO2+xoFtn7JOjaqq8uI10SZIM > aMS6DbVxsfpAr7Cxxmgh1mUJCJafkSNdEFTGQagJsCjIhshpqXw0ztaeYcaSC4c4ylHWpwD3 > FRxMC0ruiB9LDxVhaB19iVS66+2lUpEgx1wdeqSBGl8IZADSXKayU0cGntghP0BJFbYqpGMx > EkKMfn40oSEeqfR3PBZMpJZBopJvNHWkCD1NwAdsLPeUh9Xl4uayy3O+zOYrzvH0QIjEtzwa > nEqpSM5F8Aeejln5gc31EevTEbDNMf4v2V4jegsUozyjMxZDGaY6/fKVnJWFgum2xknyDPmc > O7P3Um8vRJ9eBhUqT5Bzvrz6b6RiMLy5Pvt+V4Pku5UVjqLNQK+hZdMDV7y+8t0zNJ1pPoQf > tWGJAahA/c6bE7SeoVJDaaYyRHo+la4LcV8vqMNJPZ0IPpa/VIQHU6D6MF90l3+vOcuGOAot > VV/bMEUR4iFwLeUF/APFmQ9chAMAAA== > > --Boundary_(ID_7GDOUavaFQH+11h1rxUO+A)--

2 1

Re: Sensible dynamic menus?
by danielv＠netvision.net.il 31 May '02

31 May '02

That sounds good. We'd also need something a bit more complicated to fix the problem of menus on lists (like all the menus of lists in the browsers). stephane ducasse <ducasse(a)iam.unibe.ch> (home) wrote: > Hi > > I was thinking about the problem of the world menu as this introduces a > lot of dependencies we will have to deal with. Here is an idea that I did > not have the time to implement. > > We could have a kind of description based on symbol of how the menus > should look like. These descriptions could be used by central registration > mechanism to produce menus that are dynamic but can follow a certain > structure described by the menu description. The menu does not really have > to be fully dynamic but just recreated every times a new module is loaded > or unloaded. > > Stef > > > > On Monday, May 27, 2002, at 06:47 PM, Lex Spoon wrote: > > > danielv(a)netvision.net.il wrote: > >> you get the idea. This simply depends on the load order, which is > >> simple, but butt-ugly and inconsistent to the user. > >> > > > > Well, alphabetizing will help with the former. I don't know about the > > rest. Surely, however, "changes sorter" and "email reader" should be > > in separate areas of the menu. > > > > Really, which morphs are special enough to be "apps", anyway? As > > opposed to "tools" or just regular "morphs"? If the question is > > unanswerable, perhaps we could move everything into the "objects" > > browser and simplify things a bit. > > > > I dunno. How to arrange everything seems to lead back to the question > > of what Squeak is, exactly. > > > > > > Lex > >

1 0

[GOODIE] Befunge'93 now supported in Squeak!
by Brian Keefer 31 May '02

31 May '02

It now wastes a perfectly good sheet of html at http://minnow.cc.gatech.edu/squeak/2488. I tried clicking the recycle button on my Web Browser, but it only gave me another useless page. What a wasteful system! The null pointer hits you! --more-- Do you wish to have your stack identified? [yn]

1 0

[VM] Linking and external libs
by Stephen Pair 31 May '02

31 May '02

I want to create a plugin that links against an external library using the existing VMMaker/make scheme. Does the existing configure/make system allow for this? I can modify the existing makefiles and such easy enough, but I'd like to not touch those if possible. I would imagine that it should be possible to list the libs that are needed by a given plugin in a special file included in the plugin's directory (either the "Cross" or platform specific directory within "platforms") and have the build scripts detect it and generate the Makefile with the appropriate -l options. Does something like this currently exist, or do I need to modify the build scripts? - Stephen

5 6

Jump to page: