# =========== Header ===========# File:					Changes# Project:				ATA Support# Written by:			Paul Guyot (pguyot@kallisys.net)## Created on:			02/27/2001# Internal version:		1## Copyright:			© 2001 by Paul Guyot.# 						All rights reserved worldwide.# ===========# =========== Change History ===========# 02/27/2001	v1	[PG]	Creation of the file# ===========ATA Support 1.0b8* There was a bug in the driver which I fixed. It does solve some weird time outs (which may explain the abnormalslow transfer with some cards), but it doesn't fix the impossibility to write to some card which has been reported.* The installer has been changed quite a lot to be used with the final non demo version (it asks the registrationcode if compiled with kDemo set to NIL).* I've been mostly working on the documentation, and I'm doing it slower than expected. Hence the fact that thisbeta will expire in two weeks to give me the time to fix the bugs and finish the documentation. At least, I'll try.ATA Support 1.0b7* The cards are now put to sleep if they're not used for a given time (which you can set in the preferences)* The bug of the log version under NewtonOS 2.0 has been fixed.* The bug in the Set Auto Mount button has been fixed.* The log can now be deleted by the regular version.* Various improvements in the installer.* I added more debug code (in the log version only) to detect and fix the bugs I don't experience (see below).	As usual, I need your help for those.* The software now says 'CF or compatible' instead of CompactFlash when it encounters cards which behaves like	CompactFlash cards. Indeed, some users reported that they had regular ATA PC-Cards which were said to be	CompactFlash cards while they hadn't the size of those. The log showed that their firmware indeed said they	were CF compatible hence the mistake.ATA Support 1.0b6* Fixed the aesthetic bug in the list of partitions (which ate the name of the partitions).* Fixed a bad bug in the power management. The cards (including non ATA cards) were never powered off	and this lead to problems with some 3Com cards.* Fixed a problem in the management of separated transactions. Separated transactions weren't handled	as the system actually expected and this lead to various problems including a weird -10061 when the	ATA Store was declared as the default store and you surfed the internet with Newtscape.* Added notify icon (blinking star) actions to unmount the stores on the ATA cards. There is also	a preferences checkbox to disable this feature.* Added the version string (which, when tapped, gives the build string) in the about box.* Fixed some elements in the NewtonScript interface.ATA Support 1.0b5* Fixed a bug in transaction abortion (which corrupted data). *Therefore it's	highly recommanded to re-initialize any store that was used with a previous version*.* Fixed a timing problem in the order the various modules are loaded. This leaded to a -10059 when	rebooting with a card inserted.* Implemented a workaround for Avi's Backdrop card icon to work as expected even with ATA cards.ATA Support 1.0b4* The compilation of the native code is now controlled by a makefile instead of the Newton C++ Toolbox menus.	The difference for the end-user is that the code in NTK native modules will be smaller and the developer	will last longer as he drinks less coffee because a full compilation takes less time.* Fixed a bug in the interface: the card slip can now be moved.* Little optimization when commiting transactions.* In order to prevent data loss/corruption, the Newton is restarted when an ATA card	which was still in used is brutally removed.* The software now comes in an installer package.* The stores can now be erased (in fact, the system erase function only works with linear cards - it plays	with XIP which isn't possible with ATA cards, so erasing a store is strictly equivalent to unmounting it,	initializing it, remounting it and confirm erase twice).* There is an alert (which can be disabled) when a store that was just mounted wasn't	in a clean transaction state.* When a store gets full, the system sometimes call a function which wasn't implemented. This function	has been implemented and apparently cards getting full are no longer a problem.* The menu item for the backup feature was removed (the backup feature will not be implemented for 1.0 release).* When partitionning, there is now no alert to confirm the initialization of the stores which	have the auto-mount flag set.* There is now a confirmation for Initialization and Erase.ATA Support 1.0b3* The software now supports 16 (1 x 16) and 8 (1 x 8) bits transfer width for the memory mode in addition to 32 (4 x 8) bits.	I think that most incompatibility problems came from the fact that the cards didn't support the PCMCIA controller's 32 bits	mode.* A bug in the management of installed ATA partitions was fixed. This bug had several symptoms: -8007 when tapping the card	icon in the extras drawer and maybe the error -10059 at startup.* The software now has a enhanced log mechanism which is kept between restarts (to help to fix the emate sleep bug).	Besides, it can be sent by mail directly. Now, most exceptions generated by bugs in the native code are logged as well.* ATA Support now better guesses the size written on the card for the card slip (the card info slip shows the real size).* The arrow keys in the button bar now let you scroll in the list of partitions. Additionnally, there are now	arrow buttons which appear in the card slip if there are more than 4 partitions.* Changed the fake serial ID computed for 2.0 devices to something more satisfying.ATA Support 1.0b2* The bug "view missing viewClass" was fixed.* The bus error bug when the store was full was fixed - *because of this bug you would better	initialize stores used with release 1.0b1** The used size is now computed in a more pessimistic way, but when the card will run out of	space for journalization, it will say that the card is full.* A bug in the use of the sector cache was fixed.* A bug in the identification of the card was fixed.* Changes in the cache sizes are now taken into account when the cards are inserted or when	the stores are mounted.* The name of stores are no longer read when the system preferences are open (to prevent an	error if the system prefs are open to force initialization of the stores).* Since the MP1x0 newtons don't have a serial number accessible from software, registration	is now based on the user name and the kind of Newton for them.* The ATA compatibility verification is now more strict.* The card infos dialog gives more informations concerning the card.ATA Support 1.0b1* The software can now unmount stores (and the whole mounted store registrar has been redesigned)* The graphical interface is now useful and includes icons by Michael Vac’k.* Several bugs in the driver were fixed. I hope it will work with every card now.* The low level layer of the store has been revamped. This should fix the data corruption bug.* The cache (or rather the caches) have been implemented. There is an interface to change their size.* ATA Support now has preferences (basically for the cache sizes).ATA Support 0.1a2* Seven bugs in the store model (the database engine) have been fixed. * The name of the stores now appear properly (instead of [n/a: error]). * Tapping on the "Card" icon now shows all the card slips and generates an error message   only if there is no recognized card. * The ATA card slip can now be moved. * The graphical interface no longer throws -8007 errors (nil view). * The machine type wasn't tested because of a bug (that's why the previous version could   be installed on any 2.x machine) * The software should run on every Newton with NOS 2.x (thanks to Frank GrŸndel, Harri   Hohteri, Carsten Lemmen, Brian McEwen and Konstantinos Xonis) * The core was simplified and is now in one part instead of two (since the core is copied   in RAM, this reduces the both the storage space and the memory space required for the   software). * The interface between the native code called from NewtonScript and the core is now done   via a P-Class (a protocol). This reduces the (rather small) risk of a semaphore   conflict with the virtual memory. * The recognition phase generates a log which can be exported to the notes. ## ============================================================================== #### Fortune suggests uses for YOUR favorite UNIX commands!                         ####                                                                                #### Try:                                                                           ####         ar t "God"                                                             ####         drink < bottle; opener                  (Bourne Shell)                 ####         cat "food in tin cans"                  (all but 4.[23]BSD)            ####         Hey UNIX!  Got a match?                 (V6 or C shell)                ####         mkdir matter; cat > matter              (Bourne Shell)                 ####         rm God                                                                 ####         man: Why did you get a divorce?         (C shell)                      ####         date me                                 (anything up to 4.3BSD)        ####         make "heads or tails of all this"                                      ####         who is smart                                                           ####                                                 (C shell)                      ####         If I had a ) for every dollar of the national debt, what would I have? ####         sleep with me                           (anything up to 4.3BSD)        #### ============================================================================== ##