Re: [NTLK] AW: AW: AW: MP2100 Problem

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Wed Dec 12 2001 - 18:15:38 EST


>Paul? Could you give a short comment?

May I give a long one?

First: the 710031 doesn't do any harm. It has been tested by many
users, and its difference with teh 717260 is minimal. Please don't
say again that it could be the cause of problems without proving what
your are saying. If you looked at what I changed (I think I posted
this info on csnp), you would know that it's not the cause of any
mysterious problem of yours or of others. It is not the cause of the
jaggies and could not be. It is not the cause of the 10061 bug or
other bugs you might find in the NewtonOS, on the contrary, it fixes
an additional bug compared to Apple's latest patch (717260).

Second: the 710031 doesn't install itself on a German MP. It will
install itself only on a MP2x00 US with a system patch lower than it
(i.e. 717260 and lower). The level of a system patch is not indicated
by the number (710031 < 717260 BTW) but by a version number. I don't
recall what I did put exactly for the 710031, but 717260 has the same
version number than system patch 717246 (which is rather rare). They
fix the same things, hence the same version number, but 717246 says
"Newton, Inc" (hence the fact that it's rare).
If you weren't able to install latest patch on your German MP, there
are three possible explanations:
- the patch you had fixes the same bugs as the one you wanted to
install, although it had a different visible number (look at the
number of days between them, cf the FAQ for what the system patch
number means), and therefore Apple put the same internal version
number in it.
- the patch you wanted to install wasn't suitable for your Newton
(710031 will typically not install to your Newton) or older than the
one installed.
- the patch you had on your Newton or the one you wanted to install
were corrupted. The fact that the patch on your Newton was corrupted
is highly unprobable, writing the first blocks of the flash is hard
to do for an application as it doesn't have a direct access to the
drivers. And if you install a corrupted patch, you can nuke your
Newton (trust me, I experienced it).

The error message would tell you about all this (except case 3).

Third: the system patch is saved in the internal flash memory. There
is a hardware way to get rid of it, the way you're describing
(putting another Voyager ROM Board, i.e. eMate or MP2x00 Fr/En/De).
It also means losing all the internal store data. And there is an
additional risk. I don't recall how this method was discovered (I
recall losing my system patch when the accelerator was too fast), but
I wouldn't advise it. I'm grateful to Carsten (I think this is him)
or whomever did the test between an eMate and a MP2x00, but I
wouldn't advise it. I don't advise to open a Newton, neither.

The problem is that I'm not sure the first blocks could be entirely
regenerated. They include some calibration information, and this
could only be added by Apple. You've been warned.

There are also software ways:
a/ use a special system patch remover published by Apple to German
system update testers. I won't use it if I were you (and it's hard to
obtain anyway).
b/ modify a system patch installer part to force it to install it
whatever the internal version of the current patch is. Problem is
that installing a modified system patch is dangerous.
c/ I think that this could be achieved by just patching some NS
function, but I don't recall exactly if the installer code calls the
NS function or the native function or even looks directly in memory.
In these two last cases, this solution doesn't work.

Anyway, you shouldn't need to remove a system patch. It could do harm
to your Newton by unleashing bugs you probably don't want to have
(typically the 4 MB not recognized on MP2100s).

If you don't trust the 710031 update, I understand. I'm not an Apple
developer. You could ask here who is using it, though. It was
published nearly two years ago and nobody seriously complained about
it except as you're doing; the last time someone did, the 710031 was
quickly discharged. In Michael Whitten's case, Jaggies existed before
the 710031.

With reactions like yours I doubt I'll look further to fix deeper
bugs such as the -10061, or rather that I'll publish patches. You
will keep saying my patches are responsible for all the problems you
or others have (even if as a MP2100D user you don't use them). BTW,
please don't use my NewtWorks find memory leak patch or NIE Patch.

I won't further comment on this point, I said all I knew about system
patches on this list, please look at the archives.

Paul

-- 
Home page: http://www.kallisys.com/
Newton-powered WebServer: http://newt.dyndns.org:8080/

-- This is the Newtontalk mailinglist - http://www.newtontalk.net To unsubscribe or manage: visit the above link or mailto:newtontalk-request_at_newtontalk.net?Subject=unsubscribe



This archive was generated by hypermail 2.1.2 : Wed Jan 02 2002 - 12:01:42 EST