On 13 Feb 2009, at 10:33, matthiasm wrote:
> Avi's fix is a good approach, but it has two flaws IMHO:
>
> 1: it is too complicated. <snip>
>
> 2: it is programmed as an extension. It should however be a patch.
> <snip>
Agreed.
On 13 Feb 2009, at 15:56, Stephen Rudy wrote:
> Would it be easier to just change the "start counting from" date in
> the ROM?
Well, yes; adjusting the NewtonScript seconds time base from 1993 to
2003 would give us another ten years, for example. But this is not
trivial -- it means patching a couple of C functions because we need
the 32-bit unsigned integers only available in C.
Patching the NewtonScript functions doesn’t work because arithmetic
overflow in NewtonScript doesn’t wrap numbers around the available 30-
bit scale -- it crashes the interpreter.
Interestingly, the clock functions in ROM count seconds since 1904,
presumably being based on Mac clock functions of the time. This will
roll over some time in 2040, causing problems of its own.
The main problem remains -- whatever solution we adopt for fixing the
2010 problem we need a system patch to apply it persistently. I don’t
believe Apple has *any* Newton knowledge any more, so is in no
position to release it no matter how persuasive our appeals may be.
The last person to create a system patch, Paul Guyot, did so some nine
years ago, and achieved it by patching Apple’s previous system patch.
If he or any former Newton OS engineers can remember how it’s done, I
think they are our only hope.
Simon
====================================================================
The NewtonTalk Mailing List - http://www.newtontalk.net/
The Official Newton FAQ - http://www.splorp.com/newton/faq/
The Newton Glossary - http://www.splorp.com/newton/glossary/
WikiWikiNewt - http://tools.unna.org/wikiwikinewt/
====================================================================
Received on Fri Feb 13 15:38:29 2009
This archive was generated by hypermail 2.1.8 : Fri Feb 13 2009 - 16:30:01 EST