Squeak-dev March 2000

squeak-dev@lists.squeakfoundation.org

131 participants
580 discussions

Re: [ENH] Layouts (Fixed size buttons, etc.)
by Doug Way 29 Mar '00

29 Mar '00

On Fri, 24 Mar 2000, Vassili Bykov wrote: > Here is an implementation of what was discussed but not done at SqueakEnd on > Sunday morning: Layouts (an analog of VW's LayoutFrames) and a few base > methods fixed to make use of them. They make buttons and some other UI > elements of standard browsers and the debugger fixed size. > > The changes are relative to 2.7 image. Vassili, thanks for the changeset. The inability to create fixed-height subpanes in SystemWindows has been a long annoyance for me. (e.g. The instance/?/class buttons on the Browser needlessly adjust their heights when the window is resized... especially annoying if you size the window smaller and then the text in the buttons is obscured. Same goes for the optionalButtons panes, annotation panes, category panes in hierarchy browsers, etc.) I would like to see this functionality included in the base image at some point. Your implementation seems reasonable to me... it still leaves room for defining subpanes the old way. I did notice a minor bug, though. After loading the changeset and opening a System Browser, I noticed that fairly often (maybe 50% of the time), the black borders around the window and between the subpanes are 3 pixels thick, instead of 2 pixels thick. I'm guessing this is some kind of rounding error... it even occurs with windows such as the Method Finder, which don't have any fixed-size subpanes. Oh wait, I was just testing with 2.8alpha. Things look better (mostly) with 2.7. Sorry. Anyway, I thought I'd also mention a couple of sort-of-related wish list items... 1. It would be nice to have a simple single-line pluggable text morph which is not scrollable, when a fixed-height (single-line) subpane is desired. This could be used for annotation panes, category panes in hierarchy browsers, the template ("*") pane in the file list window, the text entry pane in the method finder, etc., etc. Scrollbars serve no purpose in these text panes, and they usually look clunky (especially inboard) because they stretch down below the lower border of the pane, and can interfere with resizing the lower border (a bug). This probably wouldn't be too hard to implement... I guess one could add a little functionality to PluggableTextMorph (or maybe just TextMorph?) to make it non-scrollable, prevent wrapping, etc. 2. Another nice thing would be the ability to drag all associated fixed-sized panes when resizing a pane border. For example, dragging the horizontal border above the method text pane in the System Browser would also drag the instance/class buttons up and down with it, since they are of "fixed" height anyway. This capability would be a thing of wondrous beauty... although there would be some complications. You would probably need to prevent resizing of the fixed-size panes in general for this to work. (Otherwise, e.g. if you had both an annotation pane and the optionalButton pane open, there would be confusion over what to resize versus what to move.) Arguably, you don't really need to be able to resize these fixed-size panes anyway, though. Implementing this would not be as difficult as it sounds... I think the only changes would be in SystemWindow. (To see if a subpane is of fixed height, you'd simply test for topFraction = bottomFraction...) I may try coding this. Of course, you could also take a completely different approach to fixed-size subpanes, by having the SystemWindow not register them as an official subpane (paneRect). Instead, you'd simply implement them as a submorph within an existing subpane. It would then be visually obvious that they were not resizable, too. (I guess this would only work for Morphic.) Anyway, enough rambling for now... - Doug Way EAI/Transom Technologies, Ann Arbor, MI http://www.transom.com dway(a)mat.net, @eai.com

1 0

[GOODIE] Client Server Framework (very basic)
by Robert Withers 29 Mar '00

29 Mar '00

Hi everyone, Here is a base cut at a client server framework. I cut out chatter (distributed object framework) so I could focus on stabilizing the core subsystems. I have spent the last week refactoring the locators, resolvers, listeners, and connecters. I would absolutely love it if folks picked this up and started using it alpha so we can shake the problems out and get it to production quality. It should port to other dialects fairly easily as well. It's open source, of course. The rule seems to be that bugs get discovered in direct proportion to the number of eyes looking at the code. This is my pitch - I will diligently coordinate any assistance that folks can give to this effort, in the hopes that my passion is rewarded with a valient user base; please try and break it - that makes me happy. :-) (and you most certainly will!) MOTIVATION: -Distributed Objects between dialects FEATURES: -Uses the Camp Smalltalk BinaryInterop ObjectStream, which runs on 5-6 dialects. -Provides for a reliable stream interface to sockets. -The endpoint hierarchy has a stackable protocol stream -Uses a connection manager to pump accepts -There is an InMemoryTransport, InMemoryListener, InMemoryConnecter, and ReliableInMemoryStream to implement an inmemory transport system. Locators should look like: 'comm://inmemory/4501/ObjectService' asLocator -(others to be sure...) BUGS: -Dangling Service/Manager connection -Can't always save the image -Not very robust examples or applications (well, there is one...hey, what time is it over there?) -Spotty UnitTests -Needs work in the exceptions arena. They are thrown but not as helpful as they could be. -UDP??? INSTALL: Load the BinaryInteroperability changeset. Load the ClientServerFramework changeset. YOu should have SUnit to run the testCase. PROBLEM REPORTS: 1) running the Unit test twice will give you a reliable stream error the second time. Basically the Service doesn't shutdown cleanly. The third time it will work again so it seems either enough time has passed for shutdown or the second run through will cause the manager/service to close. 2) If you save the image after running the test, then when you reopen it the ObjectService is spasming. This may occur if you save after running the test once or thrice, .. and, thus, it is the same problem as above. I need to try and run it 2 times then save. 3) The binary interopt stuff is probably not the latest and greatest that Paul Baumann has. It's in ENVY so it is tricky getting it filed into Squeak. I am accepting questions, advice, criticisms and fixes, perhaps even complaints, but not until I've had my lunch... thanks and enjoy, Rob -- -------------------------------------------------- Smalltalking by choice. Isn't it nice to have one!

1 1

Re: Test
by Tim Cuthbertson 29 Mar '00

29 Mar '00

>Is this e-mail in normal text mode now or still in that funky multi- part >MIME mode? > >-david > > Looks like plain text, to me. Tim

1 0

[Project][Morphic][Drag&Drop] Drag&Drop for PluggableListMorph (usable for standard tools)
by Stephan Rudlof 28 Mar '00

28 Mar '00

Dear Squeakers, I'm planning to implement some Drag&Drop functionality for PluggableListMorph, which could be used to implement Drag&Drop functionality for standard tools in Morphic. Are there any efforts in this direction already? Greetings, Stephan -- Stephan Rudlof (sr(a)evolgo.de) "Genius doesn't work on an assembly line basis. You can't simply say, 'Today I will be brilliant.'" -- Kirk, "The Ultimate Computer", stardate 4731.3

2 1

Test
by dpennell＠quallaby.com 28 Mar '00

28 Mar '00

Is this e-mail in normal text mode now or still in that funky multi-part MIME mode? -david

3 2

RE: zoomable morph?
by Torsten.Bergmann＠phaidros.com 28 Mar '00

28 Mar '00

An interesting representation for trees is Hyperbolic Tree http://www3.inxight.com/htdbdemo/htdocs/index.html It's patented :( Torsten

1 0

RE: Test
by dpennell＠quallaby.com 28 Mar '00

28 Mar '00

Hooray! BTW - it turned out to be a config issue on the Exchange SERVER, not on the client. -david > -----Original Message----- > From: Bob Arning [mailto:arning@charm.net] > Sent: Tuesday, March 28, 2000 9:14 AM > To: squeak(a)cs.uiuc.edu > Subject: Re: Test > > > On Tue, 28 Mar 2000 08:27:48 -0500 dpennell(a)quallaby.com wrote: > >Is this e-mail in normal text mode now or still in that > funky multi-part > >MIME mode? > > plain text. > > Cheers, > Bob >

1 0

Re: zoomable morph?
by Dino 28 Mar '00

28 Mar '00

So where is Ivan's ZUI morph? Hsin-hao Yu wrote: > > Hi, the amazing Squeak community, > > I am doing cluster analysis on natural languge corpus. > I couldn't find a free program that can draw a large > tree, and let me zoom in and out regions of that tree. > So my questions are: > 1. Do you know this kind of software? (free I mean) > 2. If not, do you think it is easy to write such > a morph easily? I am totally new to Squeak, so > please give me some idea where to start. > 3. Are there existing code that I could reuse? > I found something very similiar to what I need > in Siren's DisplayList, but that was written in > MVC, which I really don't want to learn.

1 0

Re: Test
by Bob Arning 28 Mar '00

28 Mar '00

On Tue, 28 Mar 2000 08:27:48 -0500 dpennell(a)quallaby.com wrote: >Is this e-mail in normal text mode now or still in that funky multi-part >MIME mode? plain text. Cheers, Bob

1 0

RE: Morphic Documentation -- Eye Candy
by Jarvis, Robert P. (Contingent) 28 Mar '00

28 Mar '00

> -----Original Message----- > From: Bert Freudenberg [SMTP:bert@isgnw.CS.Uni-Magdeburg.De] > Sent: Tuesday, March 28, 2000 3:23 AM > To: squeak(a)cs.uiuc.edu > Subject: Re: Morphic Documentation -- Eye Candy > > On Mon, 27 Mar 2000, Jecel Assumpcao Jr. wrote: > > When a lot of companies (led by DEC, if I remember correctly), wanted > > to kill Sun's OpenLook GUI (not happy to have made us adopt the icky > > X11 infrastructure instead of NeWS), they decided to make their Motif > > system visually compatible with the Windows look and feel in order to > > get Microsoft to join their little group. > > The ultimate reference for this is here > http://catalog.com/hopkins/unix-haters/login.html > > Login, ls x-windows, cat disaster :-)) > ROFL! I especially liked the part about Obviously it is Allah's will that I throw the unix box out the window. I submit to the will of Allah. Motif'll do that to you... :-) ObSqueak: Thankfully, we have Morphic instead of Motif. Bob Jarvis Compuware @ Timken

1 0

← Newer
1
2
3
4
5
6
7
8
9
...
58
Older →

Jump to page: