Date   : Tue, 23 Aug 1994 11:57:58 +0100
From   : jfid@... (James Fidell)
Subject: Re: xbeeb updated on ftp.pipex.net

Stephen Quan wrote:

> JF writes :
> 
> > Would you send me a list of the sizes of all the basic data-types on
> > your machine ?  Perhaps I'll see a potential problem.
> 
> On the Sun we have
> 
>   sizeof(char)  = 1
>   sizeof(short) = 2
>   sizeof(int)   = 4      /* on a MS-DOS machine this is 2. */
>   sizeof(long)  = 4

These are exactly what I would expect, so I don't see that there
should be a problem here.

> I did try to globally replace all your unsigned int to unsigned short
> but the problem previals.  It is unusual that the software crashes
> at the macro ReadWordAtPC, but when I typed in the expression in you
> macro into the debugger it gave me a value without crashing.  Weird.

Changing int to short will break some things, I'm sure.

I can't see any obvious problems with ReadWordAtPC once the patches
have been applied -- can you send me the expression being evaluated,
and the values of Mem and EmulatorPC ?

> > JB writes :
> > > With LITTLE_ENDIAN undefined, and optimisation turned off, I've now
> > > got xbeeb working on the Alpha, but with some odd caps lock problems.
> > > (I'm not even sure how to generate caps lock on this shrunken keyboard
> > > - all the obvious key combinations no longer work since I upgraded the
> > > OS.)
> 
> Same here.

Obviously LITTLE_ENDIAN shouldn't be defined for the Sun, so are you
saying that xbeeb functions OK with optimisation off, but gives a
problem in ReadWordAtPC with the optimiser on ?

> > For some reason, it seems that the CAPS lock key has to be pressed twice
> > to get it to work correctly.  I don't know why, but I'll look into it
> > ASAP.
> 
> Couldn't after n clock cycles you just say the CAPs lock was released?

That won't work if CAPS LOCK is meant to be checked in a keyboard scan --
say if a game uses for some function.

I was thinking of modifying the X server keyboard map so that CAPS lock
mapped to some other obscure X keysym which was treated as a pushbutton.
Then it could be handled normally.

> > > Also, the palette swapping seems somewhat random - sometimes no new
> > > colourmap is allocated at all.
> >
> > I have to admit to being confused by this.  Looks like a little more
> > verbose error handling is required...
> 
> Goodluck!  I know X programming is a real dog.  I have 3 good manuals
> with me but after trying it several months ago I haven't went back!
> Hopefully when most of the other things have bowled over I'll get back
> into the gist of things.

The real problem with X programming seems to be that if you can get the
program to work on one display, there seems to be no guarantee that it
will work on any other :-(

James.

-- 
 "Yield to temptation --             |
  it may not pass your way again"    |     jfid@...        
                                     |
        - Lazarus Long               |              James Fidell