Oct 1,1993

- fixed problem that occurred when reading some control.dat's causing error
26's

- Program will now operate correctly on both Normal and Small control.dat
files

- Altered the method of reading in Message base names. If CONTROL.DAT
creator (the door) breaks the rules and has names longer then 13 characters
MMST will now truncate them. This will fix a display problem in the buttons
to access the message bases.

- Fixed problem where Taglines couldn't be selected by registered users
 
- Fixed problems with word wrap on the macro file. Now works correctly.

- Found problem when GDOS (NVDI,G+Plus) were loaded with no fonts that caused
a 2 bomb error.

- Altered the fontsize detection routine so that Double and quadruple spacing
should not occur

- Worked a bit more on the GDOS font routines. Cleaned them up substantially
but nothing will really be noticeable to the user

- Added a CLOSE channel to the show file routine. Erratically it would report 
an error 22 so I placed a close right before the open (never could find where
the error was originating).

- Somewhere in the tweaks and pokes of this version I eliminated bug that 
would cause an error 0 for some users when reading a single line message.

- Fixed Major problem with attempting to access negative array elements if
you selected Next and Previous to often from the conference selection dialog
box

- Refined the memory requirement a little bit more. This version has been 
running stably on my system since September 7 (day after last release).



September 6, 1993

- Found and fixed a few more areas where MMST would run out of memory.
The program is getting tight on a 1 meg machine (but will still run with
not much else in the way of ACC/Auto folder programs).

- Added a Search button to the conference selection dialog. This will allow
you to enter the name (or part of a name) for the base you wish to go to. If
it finds a base matching then it will be placed in the button at the top
left corner. You can select Search again to continue your search from the
last found base.

- Added some status dialogs as I found there were times when even I thought
the program had hung. Program now keeps you better informed of where you are
in the process.

One note I would like to make a general note to all, I spent many hours 
debugging a routine the other night that kept giving me a "not an integer" 
errors. Turns out after much trouble shooting and hair pulling, it was a
bad IDX file causing the problem. ZIPJR had missed on the extraction and 
had not reported the error. This is just a warning to all so that you can
maybe check something else before you report the problem as a bug in mmst.


August 31, 1993

- Best laid plans of Mice and Men Category - Seems folks that my set-up 
was satisfactory to 2.95 but mine was the only one. Everyone else that 
attempted to run the program got two bombs or worse. I was able (through
a 1.5 meg cache) to reduce the memory on my system to 793K at which point 
I started to see all the errors people have been reporting for quite a 
while regarding the reliability of MMST. I have made an interim patch that
did allow me to run MMST effectively under this set-up and this is basically
what 2.95 is all about. I will say that the memory management still needs
some work and I am going through the code to make certain of efficient use
of RAM and there definetly will be at least a 2.97 soon but I know you 
all wanted to get using the new release soon so here is 2.96. Let me know
of any oddities you encounter so that I can optimize the memory use of MMST.

- Also found a little problem with REP's that would cause some IBM doors to
reject the packet. This has been fixed. (BTW The program no longer sends the
deleted messages. They are purged prior to compression).

- Couple of cosmetic changes to the windows. There are a few more but like
I said I wanted to get this version out to you.

August 29, 1993

- GDOS fonts now work through proper VDI calls instead of GFA's DEFTEXT
and should now be more stable and accurate.

- added the ability to Qwikscan while reading (option J) as I found I
needed a quicker way to get around the message bases.

- tweaked the window routines when reading so that the window no longer
closes and opens between each message. (This was giving me a headache).

- Provided info about the packet under the "About MMST" main menu option

- Found a few mathmatical errors in the text output routines that would
result in scrambled window information.

- Fixed the error 10 reported with 2.90

- Got rid of the use of the PTR file which has dramatically reduced the
disk i/o when reading

- Now buffers the NDX file again eliminating a lot of disk i/o

- Added the ability to use the +/- keys when reading a message from
qwikscan. Example, you select a message to read from the qwikscan window
you can now use the +/- (or left/right arrows) to view the next/previous
messages. You do not return to the qwikscan menu until you select Quit.

- When reading messages you can now go backwards to the previous base. 
Example, you are reading message #1 in conference 2, selecting - (or back
arrow) you will be taken to the last message in conference 1.

- Added a long time wish list item so that both backspace and Delete will 
pull up text from the previous line within the editor.

- Changed the conference selection dialog so that you only have to click
on the conference you wish to read and not the conference and the Ok button.

- Designed a proper dialog box for the progress bar when searching. (This
was the last of the no RSC dialogs in the program).

- Worked on the routines that allow you to drop a QWK on the MaxiMiser
icon in TOS 2 and higher (same basic routines allow you to install Maximiser
as an application) and these should work now.

- Eliminated a few more variables which should free up about 50K in memory.

- Changed the Dialog that informs you of a non zero (error) result from the
archiver. I have found the STZIP will return errors even though there are none
so I now give you the option to ignore the error.

- Worked on some of the GEM message handling and the program should be more
stable now in handling mouse clicks and keyboard input.

- Program now can dynamically allocate room for the number of conferences.
Previous versions were alway hard coded at an upper limit (originally 2500,
then 3500 and most recently 5000) conferences but I have found a couple of 
doors that added even more then was needed and this was causing errors.

- Fixed problems in the editor that would not show you what was on a line
when down arrowing until after you had arrowed beyond the line.

- Program previously could only handle a message up to 32K and/or 1000 lines
this is now unlimited on K and 5000 lines.

- Search routine output now properly Highlights the text matching the search
criteria. Previously it would only highlight on an exact match even though
it would should you the message without and exact match.

- Program now tells you how many messages are in a conference through the 
select conference dialog box. The total # of messages in the packet is
shown under the About MMST menu dialog.



