Squeak-dev January 2001

squeak-dev@lists.squeakfoundation.org

112 participants
318 discussions

RE: Celeste poll
by Norton, Chris 26 Jan '01

26 Jan '01

Hi Daniel & co. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ danielv(a)netvision.net.il is rumored to have written: "But I still wonder what's a user to actually DO with 10,000 mail items in a list..." ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I don't use Celeste, but I will point out that I have over 10,000 messages in my Squeak folder... And I don't keep them all, just the important ones. :-) Cheers, ---==> Chris

1 0

[FIX] array literals
by Yoshiki Ohshima 26 Jan '01

26 Jan '01

Hello, I found that I can't create a menu that has only one menu item with its label string and the action selector are equal, in terms of String>>=. For example, a menu created by the following expression doesn't work. SelectionMenu labelList: #('suspend') lines: #() selections: #(suspend) This is because that the Compiler (Encoder) trys to merge the "same" literals, the selector to be #perform'ed is 'suspend' not #suspend. I ended up with a mutually recursive equality check in LiteralArray. Attached is the code. -- Yoshiki

1 0

Re: Swiki questions/suggestions
by Jochen F. Rick 26 Jan '01

26 Jan '01

> I just had a chat with a PHD-er from the psychology faculty of the Uni of > Groningen Holland. He was interested in using Swiki to publish and organize > their research projects on the net. Cool > I showed him how to use and set up a swiki. After the demo he saw for their > situation some limitations and gave some interesting suggestions: > a. What if you could put in xml like tags so we can have a better search > engine. Tags like '<author>' '<article><title>' would make it possible to > search more effectively. The Swiki search engine is pretty bad. I don't think anyone is working on making a better one though. > b. What if the edit-page has buttons to have an texteditor in a wysiwyg > style. Psychology students are a bit afraid of bare html editors. I have > seen this wysiwyg button option in a custom swiki during the demo of > NetUnify at OOPSLA'00. It's not bare HTML. Swiki has its own mark-up. You can think of it more like e-mail if you want. I'm not a big fan of wysiwyg, especially on the net, so I am not going to be creating this. > There are already so many swikis in use within universities, did one of them > already implement these functionalities? Nope. > Unfortunately, I'm only able to forward these suggestions and not to > implement them, cause of priority list limitations. Thanks for the suggestions though. You may want to join the PWS list, where we discuss Swikis. To subscribe, send a message to majordomo(a)cc.gatech.edu. In the body, put "subscribe pws". Peace and Luck! Je77

2 1

Re: [ENH] alternativeDrop-ls (was: Re: [Annoyance] Script editors accept drop on playfields)
by Sean McGrath 26 Jan '01

26 Jan '01

I agree that people don't look at the mouse position. They look at what they pick up and what they drop it on. The following change makes dropping much more intuitive in my experience (that means extensive user testing using FreeCell :). This ignores mouse position; instead it uses the center of the dropped morph to determine the drop target. -- Sean McGrath sean(a)manybits.net -- 'From Squeak2.9alpha of 13 June 2000 [latest update: #3193] on 26 January 2001 at 11:57:16 am'! !HandMorph methodsFor: 'grabbing/dropping' stamp: 'sm 12/25/2000 09:16'! dropMorph: aMorph event: anEvent "Drop the given morph which was carried by the hand" "Use dropped center instead of hand position for event position. This means the dropped morph is accepted if its center, not the hand position, is over the target. sm - 12/25/2000" | event dropped | dropped _ aMorph. (dropped hasProperty: #addedFlexAtGrab) ifTrue:[dropped _ aMorph removeFlexShell]. event _ DropEvent new setPosition: dropped center contents: dropped hand: self. owner processEvent: event. event wasHandled ifFalse:[aMorph rejectDropMorphEvent: event]. aMorph owner == self ifTrue:[aMorph delete]. self mouseOverHandler processMouseOver: anEvent. ! ! -- On Fri, 26 Jan 2001, Lex Spoon wrote: > > Karl Ramberg <karl.ramberg(a)chello.se> wrote: > > Script editors could follow > > the same rules as system windows and have embedding turned off by default. > > > > Definitely annoying. Note, though, the system windows solution has its > own annoyances. System windows and tile scripts are still morphs, and > sometimes you want to treat them as so, yet this preference puts up a > wall. The most frequent time I get bit by this preference, is when I > try to drop a window into a flap -- it doesn't work if the preference is > turned on. If nothing else, the unnecessary discrimination makes it > hard to demonstrate flaps to people. I've ended up a couple of times > saying "Check it out! Flaps let you drop morphs into them -- oh, well, > not any of *these* morphs, they're special morphs called > 'SystemWindows'". > > But then again, it still sucks when a browser falls into a BookMorph. > Philosophize all one wants, it seems we have all bumped into this. > > > So what to do? Thinking this through again, here's a solution that > seems to work. It has two parts. > > First, don't look at the mouse position. *People* often don't look at > the mouse position, and even programmers get confused by too much > reliance on the mouse. More than once I've seen someone accidentally > drop something into a playfield, and then go to extreme measures trying > to avoid it happening it again. The worst is the cringe they make as > they try the next drop -- they really don't know whether the system is > going to eat their window again. They don't realize that the system is > simply looking at where the mouse is. > > Second, don't accept a drop unless the dropped morph will visibly fit in > the underlying morph. At the very least, don't allow three pixels of > overlap to be enough reason to accept a drop -- the visual effect is > that the window disappears. But why stop there? Let's only allow drops > if the *entire* morph will fit intothe playfield. One might call this a > "pegs and holes" approach: don't accept a peg, unless the hole can > visibly accomodate it. Even adults understand pegs and holes. > > The main argument against this approach is that, occasionaly, you *do* > want to drop a large morph into a small playfield. However, these cases > seem to be rare. Furthermore, the functionality is still available, > albeit more difficult: you can either increase the size of the playfield > temporarily, or you can forcibly use the "embed" menu. > > > Cheers, > > -Lex > > > > ===== > 'From Squeak2.9alpha of 13 June 2000 [latest update: #3193] on 26 > January 2001 at 11:42:10 am'! > "Change Set: alternativeDrop-ls > Date: 26 January 2001 > Author: Lex Spoon > > Change the algorithm playfields and books use for deciding whether to > accept a dropped morph: ignore the mouse position, and drop iff the > morph's bounds fit entirely inside the playfield or book. > > "! > > > !BookMorph methodsFor: 'dropping/grabbing' stamp: 'ls 1/26/2001 11:39'! > wantsDroppedMorph: aMorph event: evt > (currentPage bounds containsRect: aMorph bounds) ifFalse: [^ false]. > ^ super wantsDroppedMorph: aMorph event: evt! ! > > > !PasteUpMorph methodsFor: 'dropping/grabbing' stamp: 'ls 1/26/2001 > 11:37'! > wantsDroppedMorph: aMorph event: evt > self isWorldMorph ifTrue:[^true]. "always" > self visible ifFalse: [^ false]. "will be a call to #hidden again very > soon" > self dropEnabled ifFalse: [^ false]. > ^self bounds containsRect: aMorph bounds > > ! ! > >

1 0

Re: Convention regarding exceptions
by Bob Arning 26 Jan '01

26 Jan '01

Tim, On Fri, 26 Jan 2001 11:08:03 -0800 Tim Rowledge <tim(a)sumeru.stanford.edu> wrote: >So can anyone explain to me where I can get a linuxPPC jitter vm(etc) ? I suspect that would almost have to be Ian or Marcus. >Should the latest 'plain' 2.9 image (I have 3193 to test) run on some >version of the j3 vm? I just started a completely current image with a Mac j3 VM from 1 December and it runs fine. Unfortunately, I don't know whether or where there is a version for linux that corresponds. Cheers, Bob

1 0

Portable Sound/MIDI API
by Stephen T. Pope 26 Jan '01

26 Jan '01

Hey VM hackers -- check this out! http://www.cs.cmu.edu/~music/portmusic/ -- stp Stephen Travis Pope http://www.create.ucsb.edu/~stp

2 1

Re: Fwd: have Alpha Linux VM, where/how to share it
by Bruce ONeel 26 Jan '01

26 Jan '01

Hi, This is on the ftp site at: ftp://st.cs.uiuc.edu/Smalltalk/Squeak/2.8/Squeak2.8-alpha-linux.tar.gz cheers bruce "Robert M. Riches Jr." <richesr1(a)inetarena.com> wrote: > > Date: Tue, 23 Jan 2001 10:54:34 +0100 > > From: Bruce ONeel <beoneel(a)mindspring.com> > > > > If you put it somewhere I can grab it (ftp/http) I'll build a 2.8 > > archive on the ftp site. If this isn't possible you can email it but > > I'd > > prefer not (slow line) > > Sorry for the delay, but it's been a bit hectic with > installing Redhat Linux 7 twice and trying to get the > XFree86 version 4 drivers to actually work... Anyway, > the Squeak VM that works for me on Alpha Linux with > both Redhat 6.1 and 7 is available via HTTP at > > http://www.cse.ogi.edu/~rriches/squeak > > I looked at my notes, and I couldn't find anything > about whether they support anonymous ftp, so HTTP is > probably the best bet. > > One issue I had that I forgot to mention in my bug > report message to the mailing list was the configure > script somehow thought it was big endian, so I hacked > the .h file to change that before compiling. > > Another issue is when I do fileIn and/or fileOut, it > uses CR (^M) as the line delimiter rather than LF > (^J). I don't know whether that is a configuration > issue or whether it is hardcoded in the VM. (I am > using a shell script that calls 'tr' as a workaround > for those few times I will be affected by this.) If > it's a configuration option, I would be glad to > recompile for the benefit of anyone else affected. > > Thanks. > > Robert Riches > richesr1(a)inetarena.com

1 0

RE: Squeak as an NT service
by Peter Crowther 26 Jan '01

26 Jan '01

> From: Raab, Andreas [mailto:Andreas.Raab@disney.com] > As I said in my previous post you cannot interact > (graphically) with Squeak > when it's running as a service. While this may be > theoretically possible it > would also imply that everyone who is logged on could do > whatever she wants > with Squeak and that is most certainly not what you want. The same's true for anything set up that way, but it's still useful on standalone server in order to start the service as the machine boots. If you don't do this, you need to configure auto-login on an NT box, which has a nasty side-effect of leaving an unencrypted password in a public area of the registry that can be read remotely. Just make sure you put the server in a safe place :-). - Peter

1 0

RE: Squeak as an NT service
by Peter Crowther 26 Jan '01

26 Jan '01

> From: Tim Dawes [mailto:tdawes@soliloquy.com] > "Starting the service is failed. > Error 5:Access is denied." Go to http://www.sysinternals.com, download and run their excellent regmon and filemon tools, try to start the service, and watch the regmon/filemon results. They'll tell you exactly which registry key open or write failed, or which file operation failed. Then look at the permissions on that key or file. Andreas, w.r.t. running as a service under LocalSystem but allowing that service to interact with the desktop: many native Win32 applications can do this with no problem (Performance Monitor is an old favorite of mine). It's certainly possible to do, and SRVANY.EXE in the NT4 Resource Kit doesn't seem to do anything special other than set up the appropriate registry entries. Does the Win32 VM do anything especially unusual when trying to access the screen, or is it all standard GDI calls? - Peter

1 0

RE: Squeak as an NT service
by Raab, Andreas 26 Jan '01

26 Jan '01

Tim, > Is it possible to interact with Squeak running as a service? As I said in my previous post you cannot interact (graphically) with Squeak when it's running as a service. While this may be theoretically possible it would also imply that everyone who is logged on could do whatever she wants with Squeak and that is most certainly not what you want. For applying updates/patches etc. I'd rather suggest that you stop the service, then run the image interactively, save it, and start the service again. Cheers, - Andreas

2 1

← Newer
1
...
4
5
6
7
8
9
10
...
32
Older →

Jump to page: