Software Explosion 2/Software Explosion 2 - Virtual Computer Pets (1997).iso/winspec/docs/history.txt

Text File | 1996-12-29 | 13KB | 486 lines

30th December

    . R register is now updated in the ldir() and lddr() loop code.

    . ld_a_r and ld_r_a() corrected to match exactly the Z80
    processor. Now uspirits.tap runs.

    . Added code to return 0 when reading port 0xFFFD, just to run
    uspirits.tap.

    . neg() error corrected. Now Super Hang-on runs correctlely.

    . v1.21 released.

24th December

    . Corrected bit_n_phl, thanks to a Blood's hint about 'Doods and the
things' crashing the emulated Spectrum after clicking the start option.

3rd October

    . Executable compiled for 386 processor.

    . v1.20 released.

2nd October

    . Added ZX's .SNP snapshot format.

30th September

    . Several non-critical Window-related bugs corrected.

19th September

    . Created Windows help file.

18th September

        . Corrected .BMP saving code in WSpecEm.c, with the help of my
friend Jose Caetano Silva [[e-mail]].


14th September

        . Changed Speed Popup to Options menu.


13th September

    . Created Windows'95 compliant setup with InstallShield SE.


3rd September

    . Reorganized system menu.

        . Added keyboard template and joystick control from Axel Quentin,
who sent a modified WSpecEm version.      Also, corrected it to allow the
floating keyboard template to generate keypresses with a mouse.


1st September

        Returned to Portugal. Finally I am no longer forbidden to include
other's people code and have a proper compiler! [I could not compile some
code sent to me in the university campus];

    . SEM snapshot format appended to the emulator recognized formats.
As found out, it appears it doesn't save the interrupt flags, as apparently
in SpecEmu, each time the menu has control, they are disabled. Someone,
comments?

    . OUTI and OUTD (inout.c) did nothing. Pointed out by Gilad Raz;

    . 20ms constant timing var in wpsecem.c was short instead of long,
so automic speed control was there from the beginning, but didn't work.
Pointed again by Gilad Raz;

    . Changed the timing check to just before the execute() call. This way
it will be more accurate.


1st August

    . Fixed special case of C flag handling in add_hl_ss and adc_a_r
adc_a_pss. Now CIRCLE basic instruction and any games that call the ROM
circle subroutines works perfectly. Thanks, Blood.

    . Released v1.13.


20th June

    . From now on, the next file loaded after a .ach file, will reload
the last active ROM file.

    . Released v1.12.

    . Handed the project at the University of West of England, at
Bristol, UK.


13rd June

    . Fixed the overflow logic calculation (flag P), at MATH16BI.c, for
instructions adc_hl_ss() and sbc_hl_ss() family. Now apolo11.z80 works.
It was some glitch, because the 8 bit-overflow is well done.
Thanks again, Blood.


6th June

    . Corrected TAPE2TAP, it appeared to have some problems with
16 bit-SoundCards and upper.


2nd June

    . Implemented 4x4 size;

    . Included on the menu the equivalent keys, as usually Windows
programs do. Thanks for the hint, Cotrina;

    . Released v1.11.


30th May

    . Removed a bug where if you paused the emulator, loaded a snap
and unpaused it again, you would be back to the name of the previous
snapshot. Report: Blood [[e-mail]].


29th May

    . Updated document with development notes, relating to questions
a few people did by email, and some new insigths.


28th May

    . Appended a document with tips for reading Spectrum tapes after
receiving a mail from Italy asking me related questions.


6th May

    . Appended (...) to Speed and Poke on the menu to conform to
Windows menu notation;

    . Modified the position of the video flashing instruction, to
avoid an error where sometimes the last byte wrote to the pixel area
wouldn't show up.


25th May

    . Screen generation modified -- instead of generating always 50
frames per second, now it's user definable. 25 will do perfectly, it was
stupid and slow to do 50 -- anyway at europe we have 24 at TV and nobody
complains --- emulator is a lot more faster, now; did that after
sugestions from Cotrina and Marat Fazullin;

    . As now we have a faster emulation, built a configurable delay
option;

    . When a pause is made, invalidate to screen to force Windows to
paint it, so it will work with new frame time scheme.

    . Include v3.0 of spectrum faq;

    . v1.10 released.


22nd May

    . Tried to increase task priority, but that appears to be a feature
of Windows 3.0 no longer working under 3.11, see WinMain() at wspecem.c for
details;

    . v1.03 released.

    . Poke option was recognizing numbers as signed and didn't work with
numbers that exceed 32767. Thanks goes to Alvin [[e-mail]];

    v1.03.01 released.


21st May

    . Due to a error, the ghost keys weren't working...

    . Corrected issue 2 emulation, thanks to a warning from James
McKay [[e-mail]]. [Hauting Hedges didn't work];

    . Now 5 on the numeric keypad is fire too, besides 0, when sinclair
joystick is active;

    . group of instructions ld Iy,r and ld iY,r weren't present on array
of functions instruc_tablIY;

    . ld_iY_a() corrected. It loaded A on the HY register. American 3D
Pool is now working [am3d.z80]. It displayed all the characters garbled
on the screen. [The game was again a hint from James];

    . Date of release placed on the Windows about box;

    . T-states of bit_r family of instructions corrected and 1 T-state
of delay introduced when reading/writing the ULA. Thanks goes to Gerton
Lunter for the file benchmrk.z80 and the Spectrum info[[e-mail]].


20th May

    . Corrected more two errors: still had a problem dealing with
a non-existant ROM at program startup time, and got a infinite loop when
typing LOAD "" without Loading in the menu a .BLK or .TAP tape file;

    . Increased stack size;

    . Found another bug: when loading any kind of files, except
snapshots, F was modified;

    . Send now image dirty signals to windows after opening a  load,
saveas or about box;

    . Corrected .TAP and .BLK file handling. If the block wasn't fully
loaded (e.g. searching for blocks, problems ocurred). Error spoted thanks to
80 (header+body) tape blocks of AlchNews Issue 19;

    . tap_load() modified to follow more closely the real ROM routine.

    . v1.02 released


19th May

    . .SLT loading scheme slightly modified, to allow for future
extensions of the format, again following Damien sugestions;

        . After notices from Cotrina implemented ghost keys (fake
closures);

    . Now it unpatches the ROM before saving it: it was trashing it
in others versions in a way that it could not be used with other emulators
in case it was saved by mistake;

    . Corrected bug concerning .BLK tape files.

    . v1.01 released


18th May

    . PCX saving implemented.


16th May

    . Now emulation can load properly .SLT files and it's levels,
after complaints from Damien;

    . v1.0 Released.


15th May

    . Corrected it, didn't release memory when leaving;

    . Sent v0.01.05b to Damien and Cotrina;

    . Corrected a few mistakes from the documentation;

    . Included WSpeEm.ico designed by Damien Burke;

    . Modified scheme names presentation on the windows caption,
Model 3 to Issue 3 in the options menu and F5 to reset after Damien Burke
sugestions;

    . Forgot to modify version number and sent it to Cotrina;

    . Prevent saving a snapshot which extension is not recognized,
folowing a complaint from Damien Burke;

    . Released to the public.


14th May

    . Corrected another error, where the caption box would not be
modified for snaphots when icons were clicked or dragged and dropped into
the application;

    . Sent v0.01.04b to Damien and Cotrina.


13rd May

    . Corrected error where if rom file wasn't found, the emulator was
locked in a loop, relating to a dialog box [bug report: Damien Burke];

    . Implemented F2 as saveas and F3 as load keys, folowing a sugestion
from Damien Burke.


12nd May:

    . Now I am also emulating issue 3, since Cotrina found Abu Simbel
Profanation had problems with emulation. I found out that this game only
works on issue 3 speccys. I didn't emulate issue 3, because I tougth it was
insane for games to ignore the issue 2 market;

    . Include menu option issue 3/issue 2;

        . Modified timings acording to data provided by Ian Collier;

    . Modified EI routine to prevent interrupts at the end of EI only if
interrupts were disabled at the beginning;

    . Correct bug where if you double-clicked an icon, the ROM file
would not be opened;


11st May:

    . Corrected sizing of window. It did only take in account one frame
and not the two frames. Now the screen updates are indeed faster. It appears
the program was being penalized for not being in a multiple of the screen
resolution. [anyone who played with AVI files will understand this]. That's
way it was so slow...and that's because I lost so much of may time speeding
it up.


10th May:

    . Corrected Pause and iconify buggy behaviour after a Cotrina bug
report. The program was trying to paint the screen in the icon space, and
Windows kept updating the icon, this happenning ad eternum.


9th May:

    . A kind of paint cache implement to improve speed of emulation;

    . Poke interface implemented;

    . v0.01.03b released [to Cotrina].


8th May:

    . Flash implemented;

    . Tried to implement HCR-like resolution with horrendous and too slow
results;

    . v0.01.02b released [to [e-mail] (Francisco Cotrina)].


7th May:

    . Implemented .SCR format;

    . Handle DI/HALT situation;

    . TAP saving corrected (only MULTIFACE like, yet);


2nd May 96:

    . Implemented .TAP format sucefully;


30th April 96:

    . Started implementing .SLT format;


25th April 96:

    . Cleaned a bit the files. Started implementing .TAP format;


19th April 96:

    . No more WM_TIMER messages: painting logic now is done at the end of
the Spectrum virtual time. The screen appears now to not be so 'jerky';


18th April 96:

    . ED/FD level loader trap implemented;
    . v0.01.01b released [to [e-mail]
                (Damien Burke)].


16th April 96:

    . Cache implemented in SNALOAD.c --- snapshots load time greatly
reduced;

    . v0.01.00b released. [to [e-mail](Filipe Silva) and
                [e-mail](Paulo Augusto)];


13th April 96:

    . Improved application menus.


12th April 96:

    . Another error corrected: bits 6 and 7 of ULA port were always 0
instead of 1 [Rick Dangerous didn't run];

    . Designed debugger screen;

    . Investigated why Rick Dangerous doesn't work with a joystick, but
it is a problem with the snapshot -- maybe it was loaded in a emulator with
no joystick support.

11th April 96:

    . Found several bad defined instructions in the instruction table.
dec_iY, dec_Yi were both defined as dec_yi. [Navy Moves 2 crashed instead of
asking for the codes].

10th April 96:

    . Downloaded a few snapshots from Internet;
    . Added support for longfilenames under Windows'95.


9th April 96:

    . Sound added to emulation.


8th April 96:

    . Replaced parity function at MISC.c by a parity table.


7th April 96:

    . Finished modifications. The emulation appears to be notabily faster.


6th April 96:

    . Continued to modify emulation for a faster speed. Appended all the
undocumented opcodes that never were implemented to date (the ones with
the CB prefix).


5th April 96:

    . Modified IX/IY logic;

    . Modified handling of IX and IY prefixes --
    extensive changes in the code. This will fix the error
    spoted and will increase once more the emulation speed;

    . Modified ld_a_r() and ld_r_a() for a new handling of
    R register.


4th April 96:

    . modified R register handling;

    . corrected runaway condition in getbyte() with
           corrupted snapshots (would kill emulator);

    . Modified central loop of emulation (execute()) for a faster
    emulation;

    . Spoted an error in Z80 emulation - handling IX or IY
    prefixes, if next instruction is not a HL instruction,
    the prefix affected all instructions until if found
    a HL instruction or a ED prefix -- but it's surprising
    how only failed 4 of nearly 2000 spectrum programs tested.

    . Added .SIT and .BLK handling logic.