info-mac@uw-beaver (05/31/85)
From: Moderator John Mark Agosta <INFO-MAC-REQUEST@SUMEX-AIM.ARPA> INFO-MAC Digest Thursday, 30 May 1985 Volume 3 : Issue 2 Today's Topics: macget & macput for DG machines SKEL.C bug non-CPU intensive TIME Desk Acc. Lint library for Mac ToolBox calls? More harmonic series bench results Switcher 2.6 problem Objective-C Getting Mazewars to run ? Broken disk as a result of DIY upgrade. ---------------------------------------------------------------------- Date: Thu, 30 May 85 01:05:31 EST From: "James Lewis Bean, Jr." <BEAN@MIT-MC.ARPA> Subject: macget & macput for DG machines I managed to get macget and macput running this weekend. I have not transfered many files with them, but they seem to be stable. If anyone is interested in getting them drop me a line. lewis bean bean@mit-mc.arpa ------------------------------ Date: Thu, 30 May 85 02:03:24 EST From: Leigh L. Klotz <KLOTZ@MIT-MC.ARPA> Subject: SKEL bug I noticed this bug a few months ago but I thought it had been corrected in a subsequent release. I noticed today that it hadn't. In the doMouse procedure inContent case and inGrow case there's a statement like the one below if (whichwindow = mywindow) which should be if (whichwindow == mywindow) [ This change has been made in UTILITY-SKEL.SHAR -jma ] ------------------------------ Date: Tue, 28 May 85 05:14:27 cdt From: werner@ut-ngp.ARPA (Werner Uhrig) Subject: non-CPU intensive TIME Desk Acc. [] Following is a newer version of my "Time" Desk Accessory I previously posted a few months back. There was a problem with Finder 4.1. The format below is in BinHex4.0 format, using Apple's new Font/DA Mover. The purpose was twofold.. 1) teach myself how to create DA's and 2) provide a very small and non CPU intensive clock to be used with VersaTerm. Since "Time" only tickles SystemTask() every 60 seconds, the effective speed of a terminal emulator is not reduced. "Time" takes only about 1K in the heap. I fudged a little on the GoAway... hope it doesn't bother anyone. "Time" was created almost all in 'C' using Manx's Aztec C68k. "Time" may be freely distrubuted, but not sold, as it is a Copyright 1985 by Lonnie R. Abelbeck. Please report any "bugs" and comments to: Lonnie Abelbeck ihnp4!aluxe!lra [ please find this d.a. in DA-TIME.HQX , with length 1259(7) -jma ] ------------------------------ From: stew%lhasa@harvard.ARPA Date: 30 May 85 03:34 EDT Subject: Lint library for Mac ToolBox calls? Does there exist a file containing function headers for the Mac ToolBox calls suitable for use as a lint library? It seems to me that this would be very useful. If such a file does not exist, I would be interested in collaborating on typing it in. The MegaMax manual contains what I want as an appendix; perhaps they would be willing to release it in machine readable version (it appears to have been printed on a dot-matrix). It is only 37 pages long, though, and if I get half a dozen helpers, that means only 6 pages of typing each. The result would be of use to anyone with access to any Macintosh C and any unix. Any takers? Stew ------------------------------ From: stew%lhasa@harvard.ARPA Date: 29 May 85 23:18 EDT Subject: More harmonic series bench results I ran the harmonic series benchmark and got 16.50 seconds using asm calls to sane within megamax C, and 39.5 using only megamax C. Why are we all getting such different numbers? In particular, how did someone claim a figure of 8-9 seconds for Consulair 80 bit floats when the tightest assembly language loop I get is twice that? The report of a bogus answer from Megamax was due to a bug in the compiler having to do with casting a constant to a float. To get around this, always cast to double; all floats are converted to double anyway by C, so this costs nothing. Would someone explain why a double can't be 80 bits long? That would make things a lot easier \and/ faster, wouldn't it? Then C's and SANE's idea of internal precision would mesh nicely. Incidentally, I get times of 0.94 secs on my 750 without FPA, 0.136 secs on a 780 with FPA, and 1.24 on an IBM-PC/AT with an 80287 and a souped up crystal. I expected more out of the AT -- I guess the math chip isn't quite up to the speed of the mpu. All of these calculations were using 64 bit floats, except the asm version for the Mac, which used 80 bits. All got the same answer to 15 significant figures: 9.787 606 036 044 38. One bug to check for if you get a different answer: when evaluating f = f + 1.0 / i, is the division done in 32 or 64 bit floats? If it's C, it should be done in double. If fortran (or any other language with reasonable arithmetic rules), then you have to say 1.0D0 to get double precision division. Stew ------------------------------ Date: Wed, 29 May 85 09:29 EST From: Daniel Warner <warner%clemson.csnet@csnet-relay.arpa> Subject: Switcher 2.6 problem OverVUE (version 1.0c) locks up when running under Switcher 2.6. Once OverVUE starts working on a file (even a new file) any command to the switcher (clicking on the double arrow or typing the keyboard equivalents) causes OverVUE to lock up. Option-Command-Shift-Period does get back to the switcher, but returning to OverVUE produces an empty bomb box. I have OverVUE version 1.0(d) on order and will report on its' behaviour when I get it. On the good news side I have been routinely using Edit and VersaTerm together under the switcher. This provides an Email facility second to none with enormous flexibility. In fact I'm using it rigth now. The desk accessories DiskInfo and MockPrinter also work fine in this environment. Dan Warner WARNER@CLEMSON (CSNET) ------------------------------ Date: 23-May-85 23:26 PDT From: Kirk Kelley <KIRK.TYM@OFFICE-2> Subject: Objective-C Apple announced MacApp (set of classes for writting an application) will be available in Objective-C. June BYTE has an article by the Objective-C folks about programming methodology called "Software-ICs". It is really about object oriented programming. "The notion of a Software-IC, in which reusable code is built and tested by a supplier and then delivered to consumers in binary form, is not possible without "dynamic binding." They claim that this use of dynamic binding will superceed Ada, Modula-2, and CLUE which bind statically at compile time. It seems their argument would apply to C++, the "other" object oriented C as well, I guess. -- kirk ------------------------------ Date: Thu, 30 May 85 00:19:13 EDT From: kevin@harvard.ARPA (Kevin Crowston) Subject: Getting Mazewars to run ? I'm trying to get the MazeWars game programme to work, but am having some trouble doing so. If anyone has gotten the programme to work, I'd appreciate some advice as to what I am doing wrong. The programme runs fine, but seperately; the two programmes come up but they don't seem to notice each other. I don't have AppleTalk, so I followed the advice in a previous letter on how to hook up two Macs without it. The letter suggested using the cable that came with the development system, or "two imagewriter cables back to back, connecting pins 7 to 7, 2 to 3 and 3 to 2" (my paraphrase). Looking at an imagewriter cable, I noted that it had (going from the 9-pin Mac side to the 23-pin ImageWriter side) pins 5 to 2, 9 to 3 and 3 and 8 (shorted together) to 7, so I made a cable with two 9-pin connectors, connecting 3 and 8 (shorted together) to 3 and 8, 5 to 9 and 9 to 5. Question 1: Is this a reasonable cable? Where do I plug it in (the printer or external device plug)? I then went to a local dealer and copied the system file that comes with the LaserWriter to get the various resources (two drivers, some init code, etc.) As the previous letter indicated, these resources are already in the system file. There was also a desk accessory, Choose Printer, also refered to in that letter. However, I wasn't sure what I wanted to choose using this DA. Also, on the 128K Mac (I have a 512K Mac and a 128K Mac) the AppleTalk Connect button was greyed out, but not on the 512K Mac. In neither case was there any choice in the list of printers besides ImageWriter. I didn't run any of the other software on the disk (there were some applications like install printer, etc., because I accidentally deleted the disk right after I copied the system files. Question 2: What do I want to choose with the Choose Printer DA? Is there support for AppleTalk on a 128K Mac or do I need two 512K Macs for this game? Should I run some other application to initialize the AppleTalk stuff? Any advice anyone can offer will be most appreciated. As I said, I've gotten the programme to run; it's frustrating not being able to get two people into the maze. Kevin Crowston MIT Sloan School ------------------------------ Date: Thu, 30 May 85 10:14:00 EST From: Daniel Weise <DANIEL@MIT-MC.ARPA> Subject: Broken disk as a result of DIY upgrade. Last night I tried to upgrade my MAC myself. I am no novice to making board level modifications. After installing the new chips but before configuring for 512K I powered up the machine. All memory diagnostics ran fine, but when it come time for it to read the disk it said the disk was bad, ejected it, and then kept trying to eject it (kerchunk, kerchunk, kerchuck). This happens on all disks. So I seem to have broken the disk drive. Questions: Could the two actions be related? I see no lines for the disk drive running anywhere near the memory chips. I checked for solder splashes, broken traces, etc. If they are not related, has anyone seen their disk do this? Is this a common problem with a common solution? Daniel Weise ------------------------------ End of INFO-MAC Digest **********************