Age | Commit message (Collapse) | Author |
|
Used by Phant2.
|
|
|
|
|
|
|
|
This happened to work previously because the size was not checked
for validity, and because the audio player calculates its own size
so this value was never actually used.
|
|
The invalid entries, which are on CD 6, appear to correspond to
audio that's on CD 4 (though not with the correct offset for CD
4's RESOURCE.AUD).
Skipping the invalid map entries on CD 6 should cause these audio
files to be loaded from the CD 4 audio bundle if they are
requested during chapter six since ScummVM combines resources from
all CDs and matches on their IDs.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Add Colonel's Bequest patch to fix an issue in room 35.
When you tell Lilly about Gertie in act 2, Lilly will then
walk out of the room. In case Laura (ego) is in her way, the game
will effectively block and will appear frozen.
This also happened, when using the original interpreter.
This also adds a new functionality to the script patcher.
PATCH_GETORIGINALUINT16 is able to get a full UINT16 from
original script data and insert it somewhere else.
PATCH_GETORIGINALBYTE and PATCH_GETORIGINALBYTEADJUST commands
were merged internally and use the same internal command now.
Some script patches were adjusted accordingly.
Thanks again to wjp for helping on the script issue.
|
|
When clicking anywhere while entering the middle museum party room
(which has a large golden Egyptian head inside of it), the game
will keep the script for entering active, which will then make it
impossible to leave this room.
This script bug also happened, when using the original interpreter.
The patch applies to at least English + German floppy as well as
English CD.
Very many thanks to wjp for helping figuring the actual cause out.
|
|
This fixes QfG4 character import, which specifies a size of 52 for a
buffer of size 40.
|
|
|
|
|
|
|
|
|
|
Also added in depth comments about the currently known variations
Returning the RNG seed is now implemented (although it seems no
games actually use this functionality, it seems to be just script
bugs).
Also remove Torin kRandom workarounds.
|
|
Since blockSize is asserted to be positive, buf can never be equal to _buf.
|
|
|
|
|
|
|
|
|
|
I did a full-score playthrough of the game with a warning() on
the workaround condition as m_kiewitz suggested.
The following room numbers were logged:
* room#1: intro logo
* room#21: time rip (intro and ending)
* room#531: time pod time travel
* room#290: astro chicken star background
* room#376: right after astro chicken game over (called only once)
|
|
More details in PR#889
|
|
|
|
This fixes the horse at the fountain drawing on top of the
fountain instead of behind it.
|
|
|
|
|
|
Fixes the ego disappearing behind the castle doors in MGDX.
|
|
|
|
Compression overhead takes up more space than uncompressed files,
plus the uncompressed files it creates are marginally quicker and
easier to debug with external tools.
|
|
|
|
|
|
|
|
|
|
Hoyle5 will spin on kGetTime between 15 and 300 ticks in multiple
game scripts in order to delay execution (for example, after
choosing opponents and clicking "okay"). This causes ScummVM to
be unresponsive and wastes CPU time.
This commit patches the spin subroutines to instead call a kernel
function (kWait) that waits without a spin loop. This kernel
function was removed in SCI2, and has been added back in ScummVM
specifically for Hoyle5, so this patch will not work with the
original interpreter.
|
|
|
|
|
|
Fixes Trac#9648.
|
|
In PQ4CD, when leaving the shooting range through the front door,
some transition screen items will be generated that have invalid
dimensions. SSCI simply clips these rectangles to zero.
|
|
|
|
PQ4CD and several other games contain a hack in two renderer
methods to avoid rendering invalid screen items with zero or
negative-dimension target rects. This prevents PQ4CD from
crashing during the fifth phase of target practice.
|
|
|
|
|
|
|