This has been translated from Japanese by Google

* Clipping check processing of the VDP1 was corrected.

* Windowing of the VDP2 was corrected.

* Color RAM address formation processing at the time of the bit map of the VDP2 was corrected.

* Sprite priority number formation processing of the VDP2 was corrected.

* Computing in rotary parameter coordinate of the VDP2 was corrected.

* Hook processing of the system program was corrected.

* The Warring States blade and the like moves there is no BIOS with.  * It tried to replace the ABC button and the XYZ button with 9 keys of keyboard side.

* It modified ahead job backup file retaining of the STV below  the Backup folder.

* It tried to retain the contents of the EEPROM in each every game.

* The bug where speed deviates the dual core CPU use time was corrected.

* Because environment of the Windows98 was gone,  the Windows98 is deleted from the corresponding OS.

>> Get it HERE.

Simple Bugs..

It amazes me sometimes how much users of MAME moan and complain about certain bugs, when if they simply put a fraction of the effort they do complaining into trying to fix the bug then they’d almost certainly be able to fix it.

 

Mysterious Stones has been broken since around MAME 0.77u2. There are a few bad tiles near the start of the first levels, and by the 2nd level it’s apparently unplayable.

 

I took a quick glance at the vidhrdw/mystston.c file in MAME and saw the following code.

 

 

WRITE8_HANDLER( mystston_videoram_w )

{

    if (videoram[offset] != data)

    {

        videoram[offset] = data;

        tilemap_mark_tile_dirty(fg_tilemap, offset & 0×3ff);

    }

}

 

WRITE8_HANDLER( mystston_videoram2_w )

{

    if (videoram[offset] != data)

    {

        mystston_videoram2[offset] = data;

        tilemap_mark_tile_dirty(bg_tilemap, offset & 0×1ff);

    }

}

 

 

the error stands out a mile, even if you’re not a programmer an error like that shouldn’t be too hard to fix. The mystston_videoram2_w handler is broken. As an optimization the code is meant to check if the contents of videoram is the same as the value being written to it, and then only write it if it isn’t the same. The mystston_videoram2_w handler is checking the wrong video ram. The bug was obviously introduced as the result of a simple copy & paste error.

 

fixing the bug simply involved changing

 

if (videoram[offset] != data)

 

 

to

 

if (mystston_videoram2[offset] != data)

 

 

in WRITE8_HANDLER( mystston_videoram2_w )

 

A _very_ simple one line fix. The source is available so people can make fixes like this.

>> Before and After pictures HERE.

22nd April, 2006 - GameEx 5.96

 

* News module now includes new GameEx RSS feed.

* Applications set to launch at startup are now run before initialising DirectX again.

>> Get it HERE.

2006-04-21

 

* GPU:  Fixed some texture transparency problems 

* GPU:  Fixed some bugs when rendering 16-color rotoscale OBJs 

* GPU:  Fixed some bugs when rendering extended palette rotoscale BGs 

* MMU:  Added experimental SPI touchscreen handling 

* MMU:  Fixed some bugs in the handling of long filenames in the FAT 

* GUI:  Improved timer event handling a bit 

* GUI:  Fixed .zip loading 

* GUI:  Added support for  .7z loading 

 

The SPI touchscreen handling is not perfect yet and only works somewhat, so the old faked method is still used as a fallback for homebrew ROMs using libnds.

>> Get it HERE.