|
| |
Credits :
Download :
Look for downloads on external sites:
Pokefinder.org
Production Info
Submitted by iAN CooG on 27 April 2011
Summary of changes from 4.43 to 4.44
------------------------------------
* Changes to the Disassembler
- A major issue with how comments get unloaded before exiting the
Disassembler has finally been found. The database was not being updated
with the proper flags to indicate that the comment was now stored
externally.
- Attempting to disassemble a project for a second time would, under
certain circumstances, result in no output being generated even if the
assembly filename was changed.
- Changes to destination links of direct strings would not work and would
result in the removal of the string as well as cause a "data
corruption" message.
- My code to detect a duplicate label/variable name was broken. It would
pass some obviously duplicate names, and fail some unique ones.
- When changing a existing label or variable name, you could sometimes
change it to a duplicate name. I was only checking for a duplicate name
on new labels and variables.
- Re-wrote one core function (calculate memory array usage) for speed.
- When generating a disassembly to file, and specifying the output
formatting to be "just like it is on screen", the comments are now
output in their full length. They used to be cut off, just like on
screen.
- I've added a bunch more C64 zero page and low memory references (below
1Kb). I also added some debugging code to detect duplicate names and
addresses in these definitions. When I tried it, I indeed found some
duplicates in both the C64 and 1541 definitions and they have been
fixed.
- The "Src:" filename display was not being cleared when switching to
another source file, leaving parts of the previous names behind.
- A table being displayed would get broken up into multiple small tables
if a database entry existed within the table even if it had no label.
- Sometimes the disassembly display for a file will start right at the
end, at the "End of file." line. When this happens, the only cursor
movement that works immediately is HOME. I haven't figured out why this
happens, but the condition is now detected and corrected so the file
will display from the beginning of the file instead.
- The opening dialog when starting a new project used to ask about
skipping the load address on the files, and had a checkbox to select
it. It was confusing and has been redesigned. Also, under certain
circumstances, the load address would be lost.
- When starting a new project and you said "No" when asked to use the
first two bytes as the load address, subsequent calls to set a
destination variable or link would be flaky.
- The Fkeys displayed in the fkey bar in both the Project Manager and DB
Editor would revert back to the main fkeys if you called up the Opening
Window from either subfunction. They would still work correctly, but
the text labels assignments would be wrong.
- Fixed some storage issues with the various in-memory arrays for
label/variable names, comments and direct/variable strings. Much empty
space could result when many changes are made to these entries.
Hopefully this condition has been eased.
- Fixed some visual and operational issues in the Add/Edit tables dialog.
The tables are now all lined up underneath each other, and the
selection dot for them now works properly. The SIZE value was being
retained between dialogs and so I have defaulted it to zero (blank).
* Discovered several nasty bugs in the GCR conversion routines
- Conversion of D64+error bytes with error 20's was broken when going to
ZipCode6 and G64.
- Conversion of ZipCode6 and G64 with specific errors (typically 20's) or
missing sectors to D64+error bytes was broken. I wasn't checking for
missing sectors when reconstructing the D64, so some would get missed
messing up tracks.
- ZipCode6 sectors weren't being de-interleaved properly. This could
cause sectors in a track to be out of order.
- Conversion from G64 to ZipCode6 when tracks were missing sectors was
broken. The ZipCode6 wasn't being re-interleaved properly.
- The track offsets in the G64 headers were being written wrong
- F64/D64 files were not being read correctly
- Error 29 was not being generated properly and was coming out as error
27 instead
- Going directly between ZipCode6 and G64 didn't work right as I was not
re-ordering the sectors into linear order (0-x). Going to a D64 was
fine.
- Both panels will now get updated if they are pointing to the same
directories.
* Fixed some small bugs in ZipCode/D64 conversion, specifically in the
ZipCode code portion.
- During a multi-file conversion, if a ZipCode file was missing a file,
the previous converted D64 would be deleted.
- The checkbox button [ ] in the initial dialog box for ZipCode/D64
conversion was on a different line from the button text.
- If any error happens during conversion, the relevant filename will now
be shown in the error dialog box.
* Changes to the ROM/EPROM conversion
- Rearranged the main dialog for space. Now I can easily add more options
when needed.
- Added "Swap D3/D4 data lines" option to allow for encrytion of the ROM
for the Excelerator Plus FSD-2 and Oceanic 118 drives. They have
encrypted the ROM this way and swapped the D3/D4 lines on the logic
board to decrypt the data.
- Changed the extensions to proper ones. Motorola Exormax is now EXO and
Intel Intellec 8 is now MCS.
- Added many more known ROMs, mostly JiffyDOS, to the internal list
* Added EPROM (ASCII) format conversion to the Alt-F10 menu. This function
converts the common ASCII formats such as Intel Intellec 8, Motorola
Exormax and Hexadecimal ASCII back to binary. The source ASCII file type
is usually auto-detected correctly.
* Some LNX files weren't recognized because their signature was shorter
than expected. I've changed the way LNX signatures are parsed to
recognize shorter or longer ones. I've also changed the signature that
64Copy creates LNX files with, removing the version number and making it
more generic.
* Fixed a small bug in my CheckXXX code dealing with logging. A string
variable was defined too small and could be populated with a very large
string causing memory corruption and making the logging fail (among other
weirdness). The function to show the log failure was also somewhat broken
as the long filename log file would be trimmed down to 20 characters but
the display window would be 76 characters wide.
* CheckZipCode6 has been enhanced to report if sectors were not seen or if
sectors were seen multiple times on a decoded track. Sometimes sectors
can be duplicated, resulting in some sectors being duplicated and some
missing. Also illegal sector values are now reported.
* D81 subdirectory headers were being created with improper padding
characters ($00) instead of $A0.
* The File Info (Ctrl-N) for MP3 files didn't take into account a ID3 V2
TAG header. It is now detected and skipped. Now, the proper MP3 info is
displayed for files with a V3 ID3 V2 header. The info box also shows if
either a V1.x and/or a V2.x tag exists.
* Corrected small error in ZIPDISK.DOC in my formats documents.
* Updated copyright notice to the year 2011
* Updated various locations that I am now compiling with Open Watcom v1.9
|
|
|
|
 | Search CSDb |
|
| Navigate | |
|
| Detailed Info | |
|
| Fun Stuff | |
· Goofs
· Hidden Parts
· Trivia
|
|
| Forum | |
|
| Support CSDb | |
|
 |  |
|
