Squeak-dev December 1999

squeak-dev@lists.squeakfoundation.org

146 participants
451 discussions

TestCodeGenerator: coercion issues
by Stephan Rudlof 31 Dec '99

31 Dec '99

Dear Andrew, "Andrew C. Greenberg" wrote: > > Looking at your code gave me some thoughts: > > I think you pointed out a great issue that I might address in the > next version. What happens when loading and coercing a variable in > the prolog may cause a garbagecollection? It seems to me that if you > generate: > > foo: a with: b: with: c with: d with: e > > self primitive: 'primFoo' > parameters: #(Oop LargeOrSmallInteger > Oop LargeOrSmallInteger Oop) > receiver: #Oop. > > ^self internalFoo: a with: b with: c with: d with: e > > then the generated code ought to be something like: > > /* code to load a from stack: 4 after coercing */ > interpreterProxy->pushRemappableOop(a); > /* code to load b from stack: 3 after coercing, can cause GC */ > /* code to load c from stack: 2 after coercing */ > interpreterProxy->pushRemappableOop(b); > interpreterProxy->pushRemappableOop(c); > /* code to load d from stack: 1 afer coercing, can cause GC */ > /* code to load e from stack: 0 after coercing */ > c = interpreterProxy->popRemappableOop; > b = interpreterProxy->popRemappableOop; > a = interpreterProxy->popRemappableOop; > interpreterProxy->popthenPush(6, internalFoowithwithwith(a,b,c,d)) > > If I add a coercion function to let the codegenerator know whether a > prolog is "dangerous," this can all be automatic. Thus, you could > define a coercion, say, LargeOrSmallInteger, that does all the work > of loading values, checking if they are SmallIntegers and coercing to > large integers if they are. The GC pointer saves can be efficiently > handled, and you need never worry about it again. > > What do you think? I want to argument from a newby (what I was a few weeks or days before) in plugin programming view. First I think it is important to know, which coercion functions are available and with which semantics at all. In case of a LargeOrSmallInteger conversion a user could think: 'If it is an SmallInteger it becomes a C integer, if a LargeInteger a LargeInteger oop', what isn't intended. It takes some time for a newby in plugin programming to grasp the semantics, that some objects like e.g. SmallIntegers are converted in C ints automatically, but normally oops remain oops. I think you cannot protect the plugin programmer from the effort to get a full understanding of the possible problems originating from GC of oops. As more implicit conversion there are as later these problems occur, but they occur though! > The GC pointer saves can be efficiently > handled, and you need never worry about it again. This is true for the calling process of primitives for some kind of object. But normally you aren't able to avoid GC issues at all, e.g. for generating ST objects as results. As a result I think it is a good idea to give some more conversion functions - with implicit GC safe behavior. But only with precise documentation regarding their behavior: 'What do they convert, which preconditions have been met, what are the results?'. Some documentation concerning the GC originating problems, in which a plugin programmer runs into, would also be nice (I know that this is work, what isn't so funny as improving code... So this shouldn't be read as a request to me or you for something what has to be made. But it would be nice...). 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

1 0

[FIX] Bold binary selectors in method text panes
by Doug Way 31 Dec '99

31 Dec '99

--0-696666604-946607354=:99317 Content-Type: TEXT/PLAIN; charset=US-ASCII There is a bug in which binary selectors do not appear in bold in the top line of method text panes in browsers, debuggers, etc. (Unlike unary and keyword selectors, which do appear in bold, of course.) This changeset fixes the bug by commenting out a line in Text>>makeSelectorBoldIn:. (I have no idea why the line was there... things seem to work fine without it.) Changeset is attached. - Doug Way dway(a)mat.net --0-696666604-946607354=:99317 Content-Type: text/plain; name="BoldBinarySelectors-dew.cs" ; x-mac-type="54455854" ; x-mac-creator="522A6368" Content-Disposition: attachment; filename="BoldBinarySelectors-dew.cs" Content-Transfer-Encoding: imap_stub 0,2757,2,1110,17, --0-696666604-946607354=:99317--

1 0

RE: [Q][WinCE] How to send Jot/Keyboard input to Squeak?
by Dean_Swan＠Mitel.COM 31 Dec '99

31 Dec '99

From: Dean Swan@MITEL on 12/30/99 08:12 PM Brian, Since you are using a Casio E-100, which has a pen interface, I would like to point out that "keyboard" input in Squeak will be directed to the window and/or pane which currently contains the cursor. When you pop up the WinCE keyboard or Jot, WinCE will relocate the cursor to a point which is probably not in the pane where you want your input to go. The solution is simple. After popping up the WinCE text entry pane (keyboard or Jot), tap the screen in the Squeak pane where you want your input to end up. I have been using Squeak on a Casio E-105 since July, and the on-screen keyboard definitely works with Squeak. Also, regarding you comment about sluggishness of the UI, this is **NOT** a drawing speed issue. Assuming that you are using Andreas's WinCE VM, look at the source code for the VM, where the VM preferences are implemented. They are currently not included in the WinCE build, but if you rebuild the VM with "reduce CPU usage" set to FALSE, you will find Squeak to be quite snappy on the Casio. If you don't know how to do this, keep watching the list. Andreas and several others at Squeak Central now have E-105s, so Andreas is updating the VM for WinCE. -Dean Swan dean_swan(a)mitel.com

1 0

RE: [Q][WinCE] How to send Jot/Keyboard input to Squeak?
by Raab, Andreas 30 Dec '99

30 Dec '99

Brian, > I've been using Squeak on CE for a little while, mostly > toying with enhancements for handhelds that I've been coding > up, but I haven't found a way to link the string output of > Jot or the keyboard to Squeak. I'm still getting familiar > with WinCE programming, so not all of the VM sources make > much sense to me. At any rate, is there a simple fix to > this, either at the user level or the VM-level? I'm not sure what you mean by 'linking the string output of Jot or the keyboard to Squeak'. Could you explain what the problem is?! (I'm assuming that you have a workspace or something alike open in Squeak so that you can actually *see* if there is some input ;-) You should try 'Sensor kbdTest' to see if there is actually some input generated. > Also, if possible, it'd be great to know if there's a way to > get the VM to 'force' its way to the top of the window stack > on start-up, so that I didn't have to use the Task-Switcher > to get to it. I have no idea why Squeak isn't coming up in the foreground - I suspect that reading the image takes so long (couple of seconds) that CE gets confused (the window is opened after the image is read). If anyone knows the reason for this (admittedly strange) behavior, please let me know. Andreas

2 2

[GOODIE; UPDATE] PluggableColumnListMorph
by Jason McVay 30 Dec '99

30 Dec '99

http://users2.ipa.net/~jmcvay/squeak/2.6/PCLMorph.cs from the preamble: this class give you a multi-column, sortable pluggable list morph. use DaProcessesMorph (http://users2.ipa.net/~jmcvay/squeak/2.6/DaProcessMorph.cs) as an example of its usage. __________________________________________________ Do You Yahoo!? Talk to your friends online with Yahoo! Messenger. http://messenger.yahoo.com

1 0

[ENH] SubmorphHalo
by Jason McVay 30 Dec '99

30 Dec '99

http://users2.ipa.net/~jmcvay/squeak/2.6/SubmorphHalo.cs from the preamble: this changeset provides a magenta halo that allows you to quickly select the currently selected morph's owners or submorphs. i will beef this up and make it more useful at a later time. selecting the halo with shift automatically selects the morph's main owner--the morph highest in the selected morph's owner chain just below the WorldMorph __________________________________________________ Do You Yahoo!? Talk to your friends online with Yahoo! Messenger. http://messenger.yahoo.com

1 0

Re: MajorShrink & SystemTracer
by JArchibald＠aol.com 30 Dec '99

30 Dec '99

Hernan, Do you have a fileIn of the changes you made to 2.6 to get to the state you're in? Jerry. Re: => 12/30/99 7:56:34 AM EST, hwilkinson(a)bitx.com.ar => << I'm doing some experiments with the VM and I need to add some instance variables to classes like Symbol and Behavior, To be able to test the changes, I'm using the InterpreterSimulator. To be able to use the InterpreterSimulator, I have to run it with a small image (let's say, no bigger than 1.5 megs.). Everything works great up to here. Now, to be able to change the definition of Symbol & Behavior, I have to use SystemTracer to generate a clone of the image with the changes that I want. So I started using SystemTracer to see how it worked, and I found a problem. If I generate a clone of the image that gets shipped with version 2.6, it works fine, I have no detected any problem. (But I need to shrink the image...) If I generate a clone of a shrinked image (after using majorShrink with some changes so it does not delete the class SystemTracer), Squeak crashes everytime I want to inspect something. The problem is that Squeak tries to access to a memory position that it is out of range... If I run the image with the InterpreterSimulator, I get the error "subscript is out of bounds". If I run the image with the compiled VM, it generates the dump that I'm attaching to this note. >>

1 0

More about StrikeFonts
by Stefan Matthias Aust 30 Dec '99

30 Dec '99

This text contains some rambling about needed/done refactorings/improvements to the StrikeFont and FontSet classes. The StrikeFont class has three access methods (in version 2.6) familyName:size: familyName:size:emphasized: familyNames to access certain fonts. The first two are for requesting a specific font in a specific size (or anything closeby). Currently, the size argument is the height of the font in pixels. I'd recommend to change the behavior into requesting the point size. This allows one to switch fonts more easily because the point size is always 8 or 10 while the pixel size is - depending on the font - 13, 15 or even more pixel. The change needs to be done in TextStyle>>fontIndexOfSize:. This method is also used by TextMorph>>string:fontName:size: and BookMorph>>insertPageShowingString:fontName:fontSize:, but fortunately, these methods aren't used at all, so a change in behavior should be possible. I'd also recommend to alphabetically sort the answer of familyNames. This makes the Cmd+K fonts menu of paragraph editors much more useful. The method is only used here. The method StrikeFont>>familySizeFace should probably also answer point sizes instead of heights, but fortunately, this method seems to be unused and could probably also be deleted. The only usage is in TextStyle>>printOn: and can easily replaced by calling name directly. Please explain the different between StrikeFont>>lineGrid and StrikeFont>>height to me. IMHO one method can be removed. What's the meaning of raster in StrikeFont? It seems completely unused. I dislike that StrikeFont>>readFromStrike2Stream: closes it file stream. This is against the rule that you should be touch the resources of other persons. It also requires the FontSet class to use a ReadWriteStream to read strike fonts because that's the only non-file stream that understands close. I recommend to move the close command to readFromStrike2named: With the exception of the height -> pointSize convertion, I did the proposed changes in an 2.6 squeak. As the changeset became a little too big, I will not attach it. But I'll be happy to send it to anybody who's interested. Especially when working with new fonts like Tahoma, the synthetic fonts Squeak automatically generates for bold type faces look very ugly compared to the real font. Therefore I extended the system to allow more fonts in a fontArray. I hacked StrikeFont>>reset which will now add the other variants directly to the derivativeFonts array instead of nilling out everything. To allow these fonts to be saved, I had also change the FontSet class to use a new method "sizeNNN" where NNN is not only the point size but the complete name. It's still called size because FontSet is scanning for selectors starting with size. This change allows also to unify ComicPlain and ComicBold to a single Comic family. I created a new font called ComicSans which does this and also contains the squeak specific ^ and _ glyphs. I noticed that the Squeak font is also based on Microsoft's Comic Sans MS and it seems that the Squeak central thought that including that font isn't against the licence. Is this right? bye -- Stefan Matthias Aust // Bevor wir fallen, fallen wir lieber auf.

1 0

MajorShrink & SystemTracer
by Hernan Wilkinson 30 Dec '99

30 Dec '99

------=_NextPart_000_0060_01BF52AC.09AABC20 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 8bit Hi everybody, I'm doing some experiments with the VM and I need to add some instance variables to classes like Symbol and Behavior, To be able to test the changes, I'm using the InterpreterSimulator. To be able to use the InterpreterSimulator, I have to run it with a small image (let's say, no bigger than 1.5 megs.). Everything works great up to here. Now, to be able to change the definition of Symbol & Behavior, I have to use SystemTracer to generate a clone of the image with the changes that I want. So I started using SystemTracer to see how it worked, and I found a problem. If I generate a clone of the image that gets shipped with version 2.6, it works fine, I have no detected any problem. (But I need to shrink the image...) If I generate a clone of a shrinked image (after using majorShrink with some changes so it does not delete the class SystemTracer), Squeak crashes everytime I want to inspect something. The problem is that Squeak tries to access to a memory position that it is out of range... If I run the image with the InterpreterSimulator, I get the error "subscript is out of bounds". If I run the image with the compiled VM, it generates the dump that I'm attaching to this note. I will appreciate if somebody can give me a hint to solve the problem (or solve it!). I'm using Squeak 2.6, no fixed applied on a Windows environment. Thank you, Hernan. _____________________________________________ Hern·n A. Wilkinson BitX S.A. Argentina Phone: (54-11) 4574-1544 Fax: (54-11)4573-3368 Franco 3419 . 1er piso - (1419)- Buenos Aires - Argentina Web: www.bitx.com.ar ------=_NextPart_000_0060_01BF52AC.09AABC20 Content-Type: text/plain; name="crash.dmp" ; x-mac-type="65417070" ; x-mac-creator="43534F6D" Content-Disposition: attachment; filename="crash.dmp" Content-Transfer-Encoding: imap_stub 0,2742,2,1138,0, ------=_NextPart_000_0060_01BF52AC.09AABC20--

1 0

SELF-style accept buttons
by Stefan Matthias Aust 30 Dec '99

30 Dec '99

--=====================_946541677==_ Content-Type: text/plain; charset="us-ascii" Hi, I found tha attached changeset in one of my old images and ported it to 2.6. Perhaps somebody likes this too... >From the changeset: A small hack to add SELF-style accept and cancel buttons to PluggableTextMorphs. Each morph gets two small left-aligned push buttons colored green for accept and red for cancel. IMHO the buttons look best with static scrollbars to the left. bye --=====================_946541677==_ Content-Type: text/plain; name="SelfStyleAcceptButtons.30Dec100" ; x-mac-type="65417070" ; x-mac-creator="43534F6D" Content-Disposition: attachment; filename="SelfStyleAcceptButtons.30Dec100" Content-Transfer-Encoding: imap_stub 0,2741,2,2372,0, --=====================_946541677==_ Content-Type: text/plain; charset="us-ascii" -- Stefan Matthias Aust // Bevor wir fallen, fallen wir lieber auf. --=====================_946541677==_--

1 0

Jump to page: