[comp.sys.atari.st] MWC problems update

bob@wiley.UUCP (Bob Amstadt) (05/13/87)

Since so many people were kind enough to answer my plea for help I
thought I would post an update. I did some experiments:

My system is a 1040ST (really new), SM124 monitor, and a SH204 hard drive.

Experiment 1: I disconnected the hard drive and installed MWC for
	a single double-sided system. This worked just fine. No problems.

Experiment 2: Reconnect the hard disk, but continue to use only the ram disk
	and floppy. No problems.

Experiment 3: I used the hard disk in the following trials (the table
	indicates the location of the specified entity):

			Trial 1     Trial 2     Trial 3     Trial 4

C compiler and		floppy      hard disk   floppy      floppy
libraries

Temporary files		hard disk   ram disk    ram disk    hard disk

Source			ram disk    ram disk    hard disk   hard disk
			---------------------------------------------
Results of trial:	worked      bombed      worked      worked


I'm going to experiment further. I'd be interested in hearing the
opinion of Atari. What part of my system is failing?

		"Thank you for your support."


		---Bob Amstadt

		bob@wiley.uucp
		{csvax.caltech.edu,trwrb.uucp}!wiley!bob

jdn@homxc.UUCP (J.NAGY) (05/14/87)

In article <528@elmer.UUCP>, bob@wiley.UUCP (Bob Amstadt) writes:
[results of some experiments - if he keeps the C compiler and
libraries on his hard disk, and temporary files and sources on
RAM disk, then his system bombs]

I have tried this configuration and it works fine for me.  I should
qualify this by saying, I've tried it with MWC version 1.2, using
an older 1040ST, SH204, and the Bob Locanto RAM disk configured
for 512K.  I have not tried it with the newer MWC version 2.0.


		Jonathan Nagy
		{ihnp4|harvard|allegra}homxc!jdn

john@viper.UUCP (John Stanley) (05/15/87)

In article <528@elmer.UUCP> bob@wiley.UUCP (Bob Amstadt) writes:
 >
 >			Trial 1     Trial 2     Trial 3     Trial 4
 >
 >C compiler and	floppy      hard disk   floppy      floppy
 >libraries
 >
 >Temporary files	hard disk   ram disk    ram disk    hard disk
 >
 >Source		ram disk    ram disk    hard disk   hard disk
 >			---------------------------------------------
 >Results of trial:	worked      bombed      worked      worked
 >
 >
 >I'm going to experiment further. I'd be interested in hearing the
 >opinion of Atari. What part of my system is failing?
 >

  Well Bob, I'm not "Atari", but just from looking at your tests, I'd
guess the problem(s) is/are that a) your ramdisk is too small for both
the sources and the temporary files and b) that the compiler you're
using doesn't test for read/write errors on the temporary files so it
gets fried without telling you about it when it trys to interpret half 
written files.

  You didn't say how large a program you were compiling or how big your
ramdisk was.  Many people are suprised to discover just how much room is
necessary for temporary files even on small programs....

--- 
John Stanley (john@viper.UUCP)
Software Consultant - DynaSoft Systems
UUCP: ...{amdahl,ihnp4,rutgers}!{meccts,dayton}!viper!john

trb@stag.UUCP (05/15/87)

I use MWC with the following configuration all the time:
  Compiler/Libraries: Hard disk
  Source and final code: ram disk
  Temporary files: ram disk
I use a 400K ram disk, and compile programs like HDSCAN (1800 lines of
code, 46K executable) with no trouble. I have had problems using a
512K ram disk (even though my program will compile on a 512K system
fine!) I also have Megamax 2.0, but still prefer MWC 2.0 for some
bizarre reason (probably the better docs).
  -Todd Burkey
  ...ihnp4!meccts!zeke!stag!trb

bob@wiley.UUCP (Bob Amstadt) (05/18/87)

Posting-Front-End: GNU Emacs 18.36.2 of Mon Feb 23 1987 on acme (berkeley-unix)


In article <986@viper.UUCP> john@viper.UUCP (John Stanley) writes:
>     Well Bob, I'm not "Atari", but just from looking at your tests, I'd
>   guess the problem(s) is/are that a) your ramdisk is too small for both
>   the sources and the temporary files and b) that the compiler you're
>   using doesn't test for read/write errors on the temporary files so it
>   gets fried without telling you about it when it trys to interpret half 
>   written files.

>     You didn't say how large a program you were compiling or how big your
>   ramdisk was.  Many people are suprised to discover just how much room is
>   necessary for temporary files even on small programs....


Thanks for the suggestion, but unfortunately not correct. This weekend I
was able to try my hard disk and compiler with a different 1040ST.
Guess what! The compiler ran flawlessly. Unfortunately, this does
not tell me why the compiler dies on my system when everything else
SEEMS to run fine.

So this is a plea for diagnostic software. Has anyone seen software written
for the specific purpose exercising the computer? I'd write my own, but
I'm a little upset with my machine.

		---Bob Amstadt

		bob@wiley.uucp
		{csvax.caltech.edu,trwrb.uucp}!wiley!bob