Squeak-dev November 1998

squeak-dev@lists.squeakfoundation.org

128 participants
265 discussions

Re: Diff code available?
by William O. Dargel 19 Nov '98

19 Nov '98

Dwight Hughes wrote: > The FileContentsBrowser for Squeak, written by Bill Dargel, contains a > source code diff for comparing code in external files to code in the > image - quite nice it is too: > > http://isgwww.cs.uni-magdeburg.de/~raab/fcb/FileContentsBrowser-Squeak2.0.st I'm afraid that I can't take credit for writing it. That honor goes to Andreas Raab (among his many other contributions to Squeakdom). I merely ported it from 1.23 to 2.0 and got it to open up within Morphic in addition to MVC. ------------------------------------------- Bill Dargel wdargel(a)shoshana.com Shoshana Technologies 100 West Joy Road, Ann Arbor, MI 48105 USA

3 2

Re: Diff code available?
by Pat Caudill 19 Nov '98

19 Nov '98

I asked Brian and he said it was originally developed by Roxie Rochet for Tektronix Smalltalk. Juanita ( and/or Steve Messick) and then Brian rewrote it for use by Team/V. Brian was the final author. I think the code is still property of Objectshare though. Perhaps someone there (hint, hint Stephan B.) could get permission to release it. Pat Caudill Stephan B. Wessels intimated >I liked the one they had in Team too. but I think the primary author may >have been Brian Wilkerson. > >Allen could tell us for sure. > > - Steve > >JArchibald(a)aol.com wrote: > >> In a message dated 11/18/98 2:15:24 PM Eastern Standard Time, >> DanI(a)wdi.disney.com writes: >> << I would do this myself, but if anyone has such code around, or knows of a >> goodie that does it, it would save some time. >> Please respond directly to me, not to the list. >> >> >> Lot's of us would probably like this, Dan. I would encourage whoever knows >> about such to post to the list. >> >> The best that I recall was in Team/V, and might have been done by Juanita >> Ewing. >> >> Jerry. > > >

2 1

Interstitial Interfaces for Squeak
by Marcel Weiher 19 Nov '98

19 Nov '98

Coming back to 'ma', I'd say interfaces are a first step to making the interstitial aspects of our systems more explicit. I agree that Squeak should not follow examples such as Java interfaces blindly, but rather take the lead by providing mechanisms for making system architecture more flexible by making it explicit. The problems with interfaces in Java are really the problems of required static typing, I think. My experience with optional interface typing in Objective-C has been very positive. Protocols are optional, so they don't get in the way initially. Once interfaces have stabilized, protocols then allow major / recurring communication patterns to be formally documented and verified. This makes it much easier to read code on a basic level, because the reader can 'chunk' information and thus handle more of it. The benefits of forml verification and well-defined interfaces for reasonabl stable components is also without doubt, I think, because it allows for easier variation of the non-stable parts. With a statically typed system, of course, the latter benefit is lost and so the first one not really worth that much. Marcel > From: Laurence Rozier <lrozier(a)thepattern.com> > > I agree that the benefits you list are desirable, but perhaps should be > implemented as Roles or better yet, class behaviors associated with a > prototype. Interfaces exist in part in Java because classes in Java are not > real objects which makes it impossible to provide the capabilities you've > described. Interfaces can and in my view of the Java world, often do lead to > layers of abstraction and infrastructure which tend to make the executing > system brittle and difficult to comprehend. It seems to me that Squeak, > especially from a Morphic perspective, is leading us towards more > concrete > and flexible representations. While there are some valuable lessons to > learn > from Java, I think that Squeak should lead rather than follow in the case of > interfaces. > > Regards, > Laurence > > GTE/staypufd wrote:

1 0

Re: Diff code available?
by JArchibald＠aol.com 19 Nov '98

19 Nov '98

In a message dated 11/18/98 2:15:24 PM Eastern Standard Time, DanI(a)wdi.disney.com writes: << I would do this myself, but if anyone has such code around, or knows of a goodie that does it, it would save some time. Please respond directly to me, not to the list. >> Lot's of us would probably like this, Dan. I would encourage whoever knows about such to post to the list. The best that I recall was in Team/V, and might have been done by Juanita Ewing. Jerry.

3 2

Re: Interfaces ala Java for Squeak
by GTE/staypufd 19 Nov '98

19 Nov '98

>What would >they buy us in Squeak? Some of what I thought they'd give us would be: 1) A very trace-able view of how we are using polymorphism. 2) A better understanding of how the system is partitioned. 3) They might also lead to a nice way of verifing that a ST package that you are loading supports a complete interface. (We would need to implement something like SSL's in VW) 4) They are also kinda like a typing mechanism. Not for individual variables, but for groups of functions. Like the object is declaring I am one of these things that can do 'X". Where 'X' is the interface. 5) They are kind of like a 'contract' in the way I was thinging of implementing them. Because when you declared it in the class defintion, and then accepted it, the methods in that interface would then be auto-generated for you. Thus you would fullfill the 'Contract'. and finally, the one that Java really uses them for: 6) In addition to allowing multiple inheritence of interface. Any more thoughts? Sam Griffith Jr. staypufed(a)aol.com

1 0

Diff code available?
by Dan Ingalls 18 Nov '98

18 Nov '98

Folks - I find myself wanting a feature in ChangeLists, that will show the "diffs" between the method on file being browsed, and the current method in the system, if there. By "diff" I mean a new text with insertions and deletions cleverly identified and displayed (usually in color, and with insertions underlined and deletions struck out). I would do this myself, but if anyone has such code around, or knows of a goodie that does it, it would save some time. Please respond directly to me, not to the list. Thanks in advance. - Dan

1 0

Bug: disableSounds preference spontaniously becomes false
by Chris Reuter 18 Nov '98

18 Nov '98

Hi all, I just installed Squeak 2.2 on a Solaris PC and then installed updates from the web up to "393ChangeSetMsgs-di.cs". I then switched off sound (set disableSounds to true on the new preferences window), opened "Play With Me-4" and clicked the next-page button. The system proceeded to set disableSounds to false and then hang inside SoundPlayer. It seems that AbstractSound>>play ends up calling SoundPlayer class>>startPlayerProcessBufferSize:rate:stereo:, which enables sound. The complete stack trace can easily be found by putting a breakpoint in SoundPlayer class>>startUp. I'd post a patch but I don't know the sound system well enough to know _where_ the fix ought to go. --Chris

4 3

Re: Units package (was Method names)
by stan＠stanheckman.com 18 Nov '98

18 Nov '98

A units package for squeak is available at http://stanheckman.com/a/98h/glimpsedotter/ASqueakUnitsPackage.st This package defines a bunch of objects in the new method category "Numeric-Units", adds one method to Number (adaptToDimensionedNumber:andSend:), and adds one dictionary to the global namespace (BasisUnits). Here's an example cut and pasted from a workspace. Line's with a space at the beginning are the output of "printIt" on the preceding line. "Put units into global namespace" BasisUnit loadBasisUnitsIntoSmalltalk. "SpeedOfSound" boltzmannConstant := 1.38e-23 * Joule / Kelvin. 1.38e-23 joules per kelvin nitrogenAtomMass := 14 * AtomicMassUnit. 14.00000000000004 atomic mass units oxygenAtomMass := 16 * AtomicMassUnit. nitrogenMoleculeMass := 2 * nitrogenAtomMass. oxygenMoleculeMass := 2 * oxygenAtomMass. airMoleculeMass := 0.8 * nitrogenMoleculeMass + (0.2 * oxygenMoleculeMass). 28.8 atomic mass units airTemperature := 75 * Fahrenheit. 75.0 fahrenheit soundSpeed := ((7/5) * boltzmannConstant * airTemperature / airMoleculeMass) sqrt 346.465447053216 meters per second "height of mountains" pi := Float pi. vacuumPermittivity := 8.85419e-12 * Farad / Meter. planckConstant := 6.626e-34 * Joule * Second. protonMass := 1.00728 * AtomicMassUnit. electronMass := protonMass / 1836.15. electronCharge := 1.60218e-19 * Coulomb. hbar := planckConstant / (2 * pi). e2 := electronCharge squared / (4 * pi * vacuumPermittivity). hydrogenRadius := 4 * pi * vacuumPermittivity * hbar squared / (electronMass * electronCharge squared). 5.29160925619361e-11 meters hydrogenIonizationEnergy := electronMass * e2 squared / (2 * hbar squared). 2.17994806928187e-18 joules protonsPerRockMolecule := 50. 50 massOfMolecule := protonsPerRockMolecule * protonMass. 50.36400000000015 atomic mass units gravitationalAcceleration := 9.8 * Meter / Second squared. 9.8 meters per second squared meltingEnergyPerMolecule := hydrogenIonizationEnergy / 100. 2.17994806928187e-20 joules maximumMountainHeight := meltingEnergyPerMolecule / (gravitationalAcceleration * massOfMolecule). 26598.09072963288 meters maximumMountainHeight useUnits: Mile. 16.52728734790877 miles BasisUnit unloadBasisUnitsFromSmalltalk This package consists of my own work, which I'm giving to the public domain, and one routine based on Kurt Hebel's (hebel(a)uinova.cerl.uiuc.edu) arithmetic package from the smalltalk archive at ftp://st.cs.uiuc.edu/pub/Smalltalk/st80_r4/arithmetic.st. Kurt's package has no copyright notice, so he presumably still has copyright; if you want to do anything commercial with this package, you should either talk to Kurt or replace the gaussian elimination routine. -- Stan

1 0

More Palatino Fonts ?
by Stephen Travis Pope 17 Nov '98

17 Nov '98

Hello all, In some recent version of Squeak, there were a few extra strike font files (e.g., Palatino12.sf...). Does anyone have a complete set of any one serif variable-pitch font for sizes from 6 to 24 points as strike font files? -- stp _ Stephen Travis Pope _ stp(a)create.ucsb.edu, http://www.create.ucsb.edu/~stp/

1 0

Re: [sqvm] squeak vm startup (interpreter?)
by Mark Wai 17 Nov '98

17 Nov '98

Dan Ingalls wrote: >-snipped- > > (InterpreterSimulator new openOn: 'mini.image') test > >-snipped- I've tried 2.2 (both the SqueakVM and Jitter) under Windows but it didn't work. It complains an improper store in the Bitmap object (primitive 61). But the problem I believe started in >>initializeInterpreter:. Should the bytesToShift contains a negative value? I have also tried Tim's patch and I got the same result. I will poke around more tomorrow morning. Thanks. -- Mark Wai Wator Innovision mailto: mwai(a)ibm.net or:[ mwai(a)frontiersa.com ] __

2 1

← Newer
1
...
5
6
7
8
9
10
11
...
27
Older →

Jump to page: