Items tagged: ⭐️ (60)

DIR: don't display invalid data with devices

When doing `DIR NUL` findfirst will actually return success
with an attribute indicating a device exists with that name.
MS-DOS shows 'File not found' in this case, but FreeCOM
(and also Comcom64) can display invalid data, so let's
correct that. Since the required constant FA_DEVICE isn't
always present in `dos.h`, let's define it when necessary.

[https://github.com/FDOS/freecom/issues/181]

Before (C: is MFS, D: is FAT16):
~~~
C:\>dir nul
 Volume in drive C is IR DXXXXS C
 Directory of C:\

NUL                      0  10-30-25  6:55p
         1 file(s)              0 bytes
         0 dir(s)     49,839 Mega bytes free
C:\>d:

D:\>dir nul
 Volume in drive D has no label
 Volume Serial Number is 4A09-DBB4

 Directory of D:\

NUL                      0  10-30-25  6:55p
         1 file(s)              0 bytes
         0 dir(s)      42,698,752 bytes free
~~~

With patch (C: is MFS, D: is FAT16):
~~~
C:\>dir nul
 Volume in drive C is IR DXXXXS C
 Directory of C:\

File not found.

C:\>d:

D:\>dir nul
 Volume in drive D has no label
 Volume Serial Number is 4A09-DBB4
 Directory of D:\

File not found.
~~~

cpuspeed: add NDIS v4 speed patch suitable for WFW 3.11

WFW 3.11 uses a similar timing loop as Win9x in NDIS.386 and so will hang
during startup on a blank screen on a machine with a fast CPU.

Fortunately due to its shared history it is possible to take the existing
NDIS v1/v3 source and adapt it for WFW 3.11. Add the updated source as a
new NDIS v4 source, include it PATCH_CPU_SPEED_NDIS_ALL, and update files.c
to patch NDIS.386 if detected.

Resolves: https://github.com/JHRobotics/patcher9x/issues/23

all: update files/documentation to include references to WFW 3.11

Now that patcher9x can also patch NDIS.386 in WFW 3.11, update any corresponding
files/documentation accordingly.

added spectrogram shader

to make this possible the previous frame is kept as a texture
the option of using the previous bar values in the shader is also
added

this commit also added the instalation of the line style shader

Emulate the "line style" spectrum analyzer from Winamp 2.

JS: Performance fix for chrome

Holy performance killer batman! Chrome was updating nodelists on
every change instead of batching them like firefox resulting in
exponentially worse performance. (Like "This fix made it go from
40 minutes to 7 seconds" exponentially)

CSS: Set contain on kint dumps to help client side performance

The contain and content-visibility properties are very new and
can actually break rendering entirely in different circumstances
for different browsers.

For now we'll stick to good old display:none and just hope contain
implementations get good enough to give us some perf down the line.