Squeak-dev December 1998

squeak-dev@lists.squeakfoundation.org

125 participants
236 discussions

Re: squeak program delivery, etc
by JArchibald＠aol.com 04 Dec '98

04 Dec '98

On 12/3/98 11:12:17 AM EST, alank(a)wdi.disney.com writes: << A "net browser" is now just a screen handler that requires almost no code >> Alan-- Are you referring to Scamper here? Jerry Archibald.

2 1

Re: squeak program delivery, etc
by Steve Dekorte 03 Dec '98

03 Dec '98

On Thu, 3 Dec 1998, Dwight Hughes wrote: > A comparison I like to use is that open source is to closed source as > science was to alchemy. Science is built by sharing knowledge and > discoveries openly so the work of anyone can be built on by everyone > else -- alchemists hoarded their knowledge and discoveries, each to > himself - his knowledge dying with him, so each generation started anew, > and no one got very far. When products or their companies die they take > many person-years of knowledge and discoveries down the drain with them, > which gets laborously reinvented to a greater or lesser degree of skill > again and again and again. That's the down side, the up side is that people will pay you to work on closed source software. Now if software was handled like drug patents... Steve

2 1

[VW]Reading Large Files
by Brett Taylor 03 Dec '98

03 Dec '98

I have encountered a limitation in VW 3.0 that I need to find a work around for. I am trying to parse large files that are in the 1.5GB range. I receive a walkback when sending the message #atEnd to aFileStream after a certain position in the file has been reached. The failure actually occurs in the following primitive method "IOAccessor>>primReadInto: buffer startingAt: index for: count". This occurs when the parameter "index" becomes a LargePositiveInteger (The number is 536870912 or SmallInteger class>>maxVal + 1). This only allows me to read the first half GB of the file. I have also experimented in Squeak which has a limit of 1073741823 bytes (also SmallInteger class>>maxVal + 1). VSE under NT uses an OS call which limit is based on the page size of the buffer so it has no problem with very large files (offsetIndex := filePosition // bufferSize). I mention these other two Smalltalks just as a side note, but I am only interested in a solution for VW 3.0 under NT at this time. Does anyone have a answer? Brett Taylor btaylor(a)quiknet.com

1 0

Re: How To Open Transcript?
by R. A. Harmon 03 Dec '98

03 Dec '98

At 06:21 PM 12/2/98 +0100, you wrote: [snip] >> I want to open a Transcript window, [snip] >> continue executing. > >That's true. Have a look at StandardSystemController>>open and >ControlManager>>scheduleActive: -- you'll see that the currently active >process is terminate when a window is scheduled. It's hard to get around >this in MVC. Thanks Andreas. I thought I might have missed what I've seen elsewhere called (I think) non-modal windows that do what I need (continue executing). How difficult would these windows be to implement in Squeak and could anyone point me where to start? -- Richard A. Harmon "The only good zombie is a dead zombie" harmonra(a)webname.com E. G. McCarthy

1 0

Slight strangeness in 8 bit CLUT support code
by Marcel Weiher 03 Dec '98

03 Dec '98

The code for generating the 8 bit color tables looks like this: for (r = 0; r < 6; r++) { for (g = 0; g < 6; g++) { for (b = 0; b < 6; b++) { i = 40 + ((36 * r) + (6 * b) + g); ^^^^^^^^^^^^ if (i > 255) error("index out of range in color table compuation"); SetColorEntry(i, (r * 65535) / 5, (g * 65535) / 5, (b * 65535) / 5); } } } Shouldn't that read '6*g+b' to give the usual layout of colors in the RGB cube? The code building the Squeak-side CLUT also does it like this as well as the code for finding colors (Color closestPixelValue8), so the colors seem to map correctly, but it still seems strange. Marcel

1 0

Re: squeak program delivery, etc
by Alan Kay 03 Dec '98

03 Dec '98

Dwight -- A very apt comparison! Cheers, Alan ----- At 6:42 PM -0000 12/3/98, Dwight Hughes wrote: >Mark Guzdial wrote: >> > >> > Dan Ingalls has pointed out that in an Open Source Software of modest >> >size -- and particularly one that tries to be self-disclosing (as we hope >> >to do a lot of in the next year) -- one or two savvy people are all it >> >takes to be independent. >> > >> >> I suggested that they build the consistencies and better ways and >> documentation themselves. I made a contrast similar to Alan's, but in >> terms of respect. Standards and closed languages implicitly say, "We have >> this covered. People smarter than you have already figured it out. Just >> take our way." And if bugs appear, the answer is, "We will fix it for you. >> Soon. Really." Squeak (and other Open Source efforts) are implicitly >> saying, "You are smart, too. You can figure this out and improve it. >> Please share what you have done because then we'll all be better off." The >> former path respects the programmer as a user, while the latter path >> respects the programmer as a peer. > >A comparison I like to use is that open source is to closed source as >science was to alchemy. Science is built by sharing knowledge and >discoveries openly so the work of anyone can be built on by everyone >else -- alchemists hoarded their knowledge and discoveries, each to >himself - his knowledge dying with him, so each generation started anew, >and no one got very far. When products or their companies die they take >many person-years of knowledge and discoveries down the drain with them, >which gets laborously reinvented to a greater or lesser degree of skill >again and again and again. > >-- Dwight

1 0

squeak program delivery, etc
by Alex Rice 03 Dec '98

03 Dec '98

Greetings, my name is Alex Rice and I'm a Java programmer. I came along looking for something faster and less bloated and I think I've found it in Squeak. I'm truly impressed by the philosophy of Squeak. A couple of newbie questions I still have after reading the FAQ and skimming the mailing list archives: 1) How are people delivering Squeak programs? For instance, if I wrote some interesting little program and wanted to give it to other people, what form would it take? Would they have to learn the Squeak desktop-browser paradigm as a prerequisite to using the Squeak program? Can someone give me an example of something written in Squeak that is intended for use by non-Squeaking persons? Can a Squeak program have a GUI without having the desktop and browser too? 2) Can I use C/C++ code and native libraries from Squeak progams? 3) Is the Microsoft licensing effecting the development of Squeak at all? Specifically, is the portability going to remain good, or will there be a Win32 MS-Smalltalk which the whole world starts using, which will be incompatible with the real Squeak? TIA, -- Alex Rice <alrice(a)swcp.com>

5 6

Morphic GUI Builder?
by Chris Reuter 03 Dec '98

03 Dec '98

Hi all, Lately, I've been getting really sick of hand-coding user-interface objects. The first couple were fun but now it's just tedious. I've been toying with the idea of writing a GUI-builder for Morphic. Actually, all that's required is 1) the ability to add submorphs to a particular Morph from the menu and 2) extend the "make own subclass" option to also generate an "initialize" method recreating the Morph's current state. I could then layout GUIs in the World and the code to recreate it would be automatically generated. The first is trivial to implement but the second is less-so. However, it looks like there's some support for this sort of thing in Morphic via the method #printConstructorOn:indent: and #constructorString. So, before I go reimplementing class Wheel, could somebody tell me if the code I want already exists or is in the works? --Chris

1 0

Re: New CompiledMethod format images available
by Tim Rowledge 03 Dec '98

03 Dec '98

I've just put a new set of image/changes/w95 vm on my website which adds the primitive error return value idea to methods calling primitives. Theres a trivial example in SimpleMIDIPort to illustrate the concept. 'Fixing' all the other prims would be simple if the idea is popular. See http://sumeru.stanford.EDU/tim/pooters/SqFiles/release/Sq23bNCM-ER.zip for the full pre-made set and http://sumeru.stanford.EDU/tim/pooters/SqFiles/deltas/PrimErrRet.zip for the change sets. tim -- Sure it's user-friendly...if you know what you're doing. Tim Rowledge: rowledge(a)interval.com (w) +1 (650) 842-6110 (w) tim(a)sumeru.stanford.edu (h) <http://sumeru.stanford.edu/tim>

2 1

Unidentified subject!
by Bill Cattey 03 Dec '98

03 Dec '98

, Ian Piumarta <Ian.Piumarta(a)inria.fr> Subject: Re: Unix sound bug not fixed in 2.3 beta. CC: squeak(a)cs.uiuc.edu, "Lex Spoon" <lex(a)cc.gatech.edu> In-Reply-To: <199812022113.QAA13245(a)cleon.cc.gatech.edu> References: <199812022113.QAA13245(a)cleon.cc.gatech.edu> Excerpts from pm.squeak: 2-Dec-98 Re: Unix sound bug not fixe.. "Lex Spoon"@cc.gatech.ed (1178) > Anyway, try it with asyncronous support and see if THAT works.... And > be sure to turn UseReadySemaphore back to 'true'. The problem with building the asynchronous version is that, since I'm running with so ancient a version of Linux as RedHat 5.0 (which came out before the dawn of time, April 1998), I have libc 5.3.12, which is older than the version you used to hand-integrate LinuxThreads 0.6. Quite frankly, I'm trying to build a squeak that won't hang when I turn the page in the demo, or hang when I plunk down a piece of a morphic script. Integrating threads feels like more work than I want to do. Debugging Synchronous mode also feels like more work than I want to do. Andreas: This all started with some change around 390 when "disable sound true" stopped disabling sound. Is there SOME relief we can get here that does not involve redoing the basics of how primitives are disabled, integrating thread packages into Linux, or hanging images? -wdc

1 0

Jump to page: