[NTLK] [ANN] ATA Support 1.0RC6

From: Paul Guyot (pguyot_at_kallisys.net)
Date: Thu Oct 16 2003 - 15:23:32 PDT


Hello all,

I'm pleased to announce ATA Support 1.0RC6.

I planned more features, but I will have little free time (if any) in
the next days, so I don't know when I'll be able to finally test
these (mostly backups).

ATA Support is a software to use ATA cards in your Newton like linear
cards. It works on NewtonOS 2.x (MP120/2.0, MP130, MP2000, MP2100,
eMate 300). There is a problem with eMate 300s, though, but it works
for some users.

How to update to RC6?
=====================

ATA Support 1.0RC6 introduces fixes to the store engine. This
basically means there were bugs in the store engine in 1.0RC5. These
mostly applied to circumstances where the Newton would crash or be
restarted during a storage operation (like installing a package). But
still, your stores can be corrupted by these bugs.

Consequently, I highly suggest to reformat your stores. This is not
mandatory. The format did not change since RC5.

What is new since ATA Support RC5?
==================================

A lot of things. In a nutshell: there was a lot of fixes for the
store engine, and, except eMate power problem, the famous bugs were
fixed (like MP1x0s incompatibility, impossibility to format large
stores).

* Interface
[CHANGE] changes in the interface for preferences, now using the
overview-like list (as in alt.rec....).
[NEW] added a preference to not open the card slip
[NEW] added a preference to disable auto-mounting (this can be done
on a card basis since earlier versions by opening ATA Support
preferences)
[NEW] added a preference to disable package activation (this can be
done since first release by pressing on the screen with your pen like
you can do it with linear cards)
[FIX] fixed a bug where pressing the [i] button in some circumstances
[NEW] pushed the PIO mode to the NewtonScript world in order to
display it in the card information slip (apparently, I don't display
it yet).

* Installer
[NEW] installer can now be beamed/moved to a card (if you don't
install ATA Support, you'll be asked if you want to keep it to
install ATA Support on another Newton)

* Native code changes
[NEW] new event from the driver towards the NewtonScript world to
know when a card is inserted. I think I updated a little bit the
interface between NewtonScript and native code, although I'm not
exactly sure.

* Store engine & caches
[FIX] fixed a bug relative to transaction abortion and the cache
[FIX] fixed a bug in the handling the status of cached pages (they
always stayed in the transaction state)
[CHANGE] optimized the update of size of pages after changes by
storing only the size in the map instead of storing the exact size
(this avoids rewriting the same information in the map)
[FIX] fixed a bug relative to deleted pages and abortion of transactions
[FIX] SeparatelyAbort and AddToCurrentTransaction no longer fail when
the object is not in transaction (this actually fixes a bug
introduced by an optimization of the deletion of objects created
within separate transactions)
[FIX] fixed a bug in the resizing of objects that could occur when
the objects are small and the page full.
[FIX] fixed a double object deletion in AbortSeparateTranny (this
would in fact yield to a leak).
[FIX] fixed a bug in separate transactions where some objects would
be deleted while it was the original copy.
[FIX] fixed the famous bug where stores larger than 128 MB
(approximatively) couldn't be formatted.

* Driver
[FIX] fixed a problem relative to the sleep code where the driver
would try to put the card to sleep while performing another operation
with it.
[FIX] fixed a bug in the power voltage level negotiation (only affects MP1x0s)
[FIX] fixed the famous MP1x0s incompatibility (namely a bug with the
8bits/16bits negotiation)
[FIX] fixed a bug in the engine where some page could be marked as
free without being actually released by the store cache. The sector
cache could then grab it and this would actually yield to a conflict,
leading to a data corruption for sure.

* Misc, source code.
[CHANGE] major update of the tests suite, including a store engine in
RAM to supervise the actual store engine.
[CHANGE] esthetic changes in the code.
[CHANGE] completely updated the debugging code

and probably other things I forgot.

Known issues
============

The eMate 300 power problem is still not fixed (I think). This is the
only main problem that prevents me from releasing ATA Support as 1.0
(unless I completely forgot something).

Roadmap towards 1.0 and beyond
==============================

The roadmap towards 1.0 is to fix the eMate 300 power bug. I also
have to check that some issues with applications such as NewtsCape or
Inkspot are fixed or to fix them. Most importantly, I want to enable
backups for 1.0 release.

Also, I need to rewrite the documentation.

1.1 will either include FAT support or asynchronous driver (i.e.
speed). I haven't decided yet. Please tell me what you prefer (the
replied I got so far were more in favor of FAT support since ATA
Support can now compete with linear cards for most operations).

Where to download ATA Support RC6?
==================================

You'll find it on my website.
http://www.kallisys.com/newton/ata/

Regards,

Paul

-- 
Philosophie de baignoire - consultations sur rendez-vous.
NPDS/NewtonOS: http://newton.kallisys.net:8080/
Apache/FreeBSD: http://www.kallisys.com/
-- 
This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries
List FAQ/Etiquette/Terms: http://www.newtontalk.net/faq.html
Official Newton FAQ: http://www.chuma.org/newton/faq/


This archive was generated by hypermail 2.1.5 : Fri Oct 17 2003 - 15:00:03 PDT