Squeak-dev February 1998

squeak-dev@lists.squeakfoundation.org

110 participants
320 discussions

Cheese w/Win support finally available!
by Boris G. Chr. Shingarov 02 Feb '98

02 Feb '98

Fellow Squeakers, I was finally able to paint a simple native window from Squeak 1.3. The source is in Cheese-1.3-v2-os2.zip. This archive supports Squeak as both PM and VIO app (i.e. you can switch PM support on or off). It's not full stuff I have in that it just paints a simple window and you don't have a native Transcript in which you can evaluate/ print expressions, inspect stuff, and so on - but I'll make this publically available pretty soon (tomorrow?). Best regards, Boris

1 0

Re: Squeak DLL callout support
by Boris G. Chr. Shingarov 02 Feb '98

02 Feb '98

Alejandro, > Is it for OS/2 ? Yes. I was able to paint native windows using this code back in October 97, but I still can't extract the code I wrote so that it can be filed into anyone else's image... I completely lost track of the changes I made :-( I can't make it work with 1.23, nor 1.30 (it's for 1.18), nor in another 1.18 image :-((( Now I gave up with this and am recreating the whole windowing stuff from scratch, this time in 1.30 and paying great attention to what I do... so I expect I'll be able to upload a copy of the full source modifications. Boris

1 0

Native windows
by Boris G. Chr. Shingarov 02 Feb '98

02 Feb '98

Fellow Squeakers, I am still unable to refactor my old-promised stuff in such a way that it could be loaded into another image :-( I promise I'll finish the thing in a day or two. For those who gave up believing there is anything working at all, I upload my 1.18 image and interpreter to create. (Yes I WILL send incorporatable source as soon as I have one myself). To see what this code can do: Evaluate the contents of the 'Play with me' window. You'll see a Transcript on your PM screen. Try resizing the Transcript (this should work). Enter '1 + 1' or '123','45', select it, and click print or do or inspect from the edit menu. File->Exit Squeak is also worth trying. Of course this all doesn't work at all but we are able to draw some native windows anyway :-) Boris

1 0

Re: Case-insensitive selectors
by Terry Raymond 02 Feb '98

02 Feb '98

On Thu, 29 Jan 1998 09:53:19 -0700, Dan Ingalls wrote: > >2. We already have a couple of tools that help a lot with the inconvenience >of matching case during programming. For one thing, the spelling corrector >will usually do the job for you. [You might be interested to know that it >is easy to change the corrector to take action without even asking if one >of the alternatives differs only in case]. Also, many people are unaware >of the cmd-Q feature that will complete a partially typed selector, saving >time to type the letters as well as to guess at their case. [If you haven't >typed enough to be unambiguous, you can keep hittinq cmd-Q to see more >alternatives. Also cmd-sh-A (we need a simpler pattern) will step forward >between keywords so inserted] > >3. One step which would help with the real problem, without any significant >compatibility issues, would be to recognize and discourage the definition of >a method whose pattern is equal to an existing pattern in letters but not in case. It seems to me the best approach would be to provide a tool that lists similarly named methods. That would allow case-sensitivity and assist the user in detecting mis-typed words. Terry Raymond Crafted Smalltalk (401) 846-6573 http://www.craftedsmalltalk.com

1 0

#include sqMachDep.h (is missing ?)
by Alejandro F. Reimondo 01 Feb '98

01 Feb '98

I am trying to compile the new interpreter (jitter) project under Windows95. We must include "sqMachDep.h" file to be compiled; but there is no reference to this inclusion in any file. Why not to add the line: #include "sqMachDep.h" when generating interp.c code? thanks. Ale.

2 1

Re: #include sqMachDep.h (is missing ?)
by Ian Piumarta 01 Feb '98

01 Feb '98

> I am trying to compile the new interpreter (jitter) project under Windows95. > We must include "sqMachDep.h" file to be compiled; > but there is no reference to thi s inclusion in any file. clotho$ head ~/squeak/devel/src/sq.h #include <math.h> #include <stdio.h> #include <stdlib.h> #include <string.h> #include <time.h> #include "sqMachDep.h" #define true 1 #define false 0 Ian

1 0

RE: Smalltalk Design Question
by Ian Bicking 01 Feb '98

01 Feb '98

Alejandro F. Reimondo writes: > Provably the solution to do Not-What-I-Say is not to say it. I > think the problem is to have a languaje (the power of Smalltalk is > not in the language, it is the environment and objects). The > problem of programming languages are extended to objects when we > interact with them sending written messages. On defined languages > (like Pascal, C, etc) the problem is reduced reducing the language > words. But with objects, each object has potentially a language of > communication. > How to solve it? Provably the better way is to programming with > gestures. (see Mark Guzdial's mail on how to write a line). Like > touching objects and programming by example. "The see-and-point principle states that users interact with the computer by pointing at the objects they can see on the screen. It's as if we have thrown away a million years of evolution, lost our facility with expressive language, and been reduced to pointing at objects in the immediate environment. Mouse buttons and modifier keys give us a vocabulary equivalent to a few different grunts. We have lost all the power of language, and can no longer talk about objects that are not immediately visible (all files more than one week old), objects that don't exist yet (future messages from my boss), or unknown objects (any guides to restaurants in Boston)." --The Anti-Mac Interface, http://www.acm.org/cacm/AUG96/antimac.htm This is talking about interface, not programming languages, but programming is really just user interface when the user is more experienced and more computationally ambitious, which is all the more reason that we need to be as expressive as we can (though I think non-progammer users deserve just as much expressiveness in their communication as well). *Sometimes* we can be most expressive using pictures (blueprints are a good example, Morphic being a good computer example). This doesn't mean we want to give up language, which seems to be the most powerful means of communicating with a computer (though the "language" is hardly as complex or powerful as human language). I'm curious how these two means of communication (language-based and picture-based) could best fit together. Pictures still clearly play second-fiddle to words, which might be innevitable because computers (unlike humans) are much less capable of understanding pictures, which possibly relegates pictures to being the programmer's means of communicating to the user, but not to the computer. -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- || Ian Bicking | bickiia(a)earlham.edu || || drawer #419 Earlham College | (765) 973-2537 || || Richmond, IN 47374 | || -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --

1 0

Re: Jitter - Segmentation fault on startup
by Ian Piumarta 01 Feb '98

01 Feb '98

Boris, > jitter always fails on startup: > error(s="opcode address is not an integer object") -> > 1) initOpcode should not get the error You should take a very close look at sqMachDep.h and make sure that the symbols JUMP_ALIGN_STRICT and JUMP_ALIGN_BYTE are defined correctly for your CPU (in sqConfig.h). JUMP_ALIGN_BYTE should be defined if instructions can be arbitrarily aligned (e.g. Pentium). JUMP_ALIGN_STRICT should be defined if the CPU does _not_ ignore the lowest bit(s) in jump destination addresses (e.g. PPC and Alpha, but not Sparc -- I don't know about MIPS without looking it up in the manual [which is at work, unlike me ;) ]). When you figure out the correct values, please send them to me. Thanks! > 2) printCallStack should not get SIGSEGV anyway... The opcode table is initialised _before_ the initial context is loaded (if I remember correctly), so any attempt to print a backtrace at that stage of the game is doomed to failure. Regards, Ian

2 1

Announce: Cheese for OS/2
by Boris G. Chr. Shingarov 01 Feb '98

01 Feb '98

Fellow Squeakers, A new port of 1.3 to OS/2 is on create. Also, a new version of DLL callout mechanism is also there. Boris

1 0

Re: Squeak DLL callout support
by Boris G. Chr. Shingarov 01 Feb '98

01 Feb '98

Tom, I've just uploaded my latest DLL stuff (modified to work with 1.30) to ftp.create.ucsb.edu. I call it Cheese as someone has recently proposed. I have a whole lot of code to deal with native windows and will devote all of Sunday to make it more or less transportable onto others' machines (I think I SHOULD manage to upload it before the next week begins - unless something serious like an earthquake happens). Btw, how we should call this part of Squeak? Squeak Windows? Squeak CheeseHoles (you know, those round openings in cheese, I don't know their English name)... Regards, Boris

1 0

Jump to page: