Squeak & end-of-line

Tim Rowledge rowledge at interval.com
Sun Dec 6 22:39:15 UTC 1998


On Sun 06 Dec, Colin Sarsfield wrote:
> A few questions:
> 
> 1.  Is their a fix for Squeak that will make it recognize a lone LF (10
> ascii) as an end of line character?
Don't know about that - Acorn is very happy with CRs
> 
> 2.  In porting to the BeOS, I've had to map the keyboard enter key to 13
> (which Squeak expects), instead of 10 (which the BeOS) uses.  Are their
> any other typical key mappings I should know about (aside from the
> command keys of course)?
All the key mappings are very Mac biased. You will almost certainly need a
mapping array somewhere in your vm code, possibly with complicated rues as
to which set of mappings is needed at any time. Look at the way Acorn has
to do it, W95 etc. All the source file are out on the web for you to look
at.
It would be much better if the key mappings were done in the image with
platform-savvy classes. Eliot invented a really neat mechanism a _long_
time ago for this, the "SystemDescription" classes.

> 
> This is more of a C/C++ question:
> 
> 3.  Is their any cross-platform way to get the address of a line within
> a function in C/C++?  If there is, we could probably make a
> cross-platform JITer easily.  (with a couple more defines in sqConfig.h,
> of course)
The Jitter code worked ok on any gcc platform, which covers alomst every
machine known. As Ian said "if it won't run gcc, it's a calculator not a
computer".
On the Acorn I had to knock up a couple of sneaky assembler routines to
work out the addresses needed, but it only took a couple of hours. And I
could have installed gcc if I had a couple of weeks to spare.
 

-- 
Useful random insult:- Always in the right place, but at the wrong time.
Tim Rowledge:  rowledge at interval.com (w)  +1 (650) 842-6110 (w)
 tim at sumeru.stanford.edu (h)  <http://sumeru.stanford.edu/tim>





More information about the Squeak-dev mailing list