## Link to GitHub Issue, if one exists
n/a
## Description of change
Complain if nvcuda.dll, nvcuvid.dll, or nvEncodeAPI64.dll are found in
modules or in root directory. This would complain for spice-provided
stubs as well.
## Testing
## Link to GitHub Issue, if one exists
Fixes#190
## Description of change
When `-cmdoverride` is specified, any command line arguments are taken
first before merging with what is in the config file.
## Testing
Still testing...
## Link to GitHub Issue, if one exists
#41 (partially)
## Description of change
This PR adds support for outputting IIDX tape LED data via spiceapi
(`iidx_tapeled_get()`). It returns the data as a `dict`, with the key
being the LED name and the value being the list of RGB values. You can
optionally pass in a variadic list of names to filter the returned data.
The new API function is only implemented in Python since I'm not able to
test the wrappers in other languages.
An example script using the new API is available in this Beef Board PR:
https://github.com/HWXLR8/beef-board/pull/133.
Demo use case: https://streamable.com/pq5e1e
## Testing
Tested in IIDX 32 with custom Beef Board firmware along with a Python
script to read and send the tape LED data over USB.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Rotate the printer output so that it's the correct orientation
(portrait, not landscape).
## Testing
Tested via test image, and also in-game, for LP and SDVX.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Add a stub for `cpusbxpkm.dll`. The stub doesn't work, it's only enough
to allow the game to load; but that is irrelevant since spice hooks the
printer at runtime.
## Testing
Observed the game boot, and was able to print out the test image from
test menu. Will do more testing on a real server.
## Link to GitHub Issue, if one exists
Fixes#322
## Description of change
Ensure game launches in full screen when windowed mode setting is off.
Before this PR, the game always launched in windowed mode.
## Testing
Seems to run fine in fullscreen with the monitor in portrait mode.
It does seem to run at even lower framerate (around 24fps, vs. 30fps in
windowed) but it might be expected.
Some weirdness with the mouse cursor when the overlay is active (two
cursors).
## Link to GitHub Issue, if one exists
Fixes#320
## Description of change
Replacing the existing logic that searches for game window to hook touch
on.
Eventually, we should fix all the other games that hook touch by
matching on window title with a call to `FindProcessWindowBeginsWith`.
## Testing
Tested with notepad open that was also titled `LovePlusAC`. Game hooked
correctly when the game was in focus during boot, and also when notepad
was in focus.
## Link to GitHub Issue, if one exists
#316
(this does not fix the issue!)
## Description of change
Add an option to omit `-noCamera` option. Doing so allows the game to
boot and search for a compatible camera using KS API, but will likely
fail since we don't know what cameras are compatible. You end up with a
camera error during boot, which can be ignored by flipping a switch in
the test menu to disable the camera.
No hooks are currently done to make the camera work!
## Testing
Tested to see that turning this on triggers the camera error. No known
camera at the moment.
## Link to GitHub Issue, if one exists
Fixes#139
## Description of change
Add an option for detecting window focus when accepting input.
The new default will be to have the following:
* naive bindings will check for window focus
* rawinput binds are not affected, will continue to be accepted
regardless of focus
Other options (always / never choices) can be picked by the user to
tweak the behavior if needed.
Clean up usage of `0xff` constant to represent invalid naive vkey.
## Testing
Still in progress, especially in regards to performance.
On a i7-9700 system, 10000 calls to `superexit::has_focus()` takes
250-350 microseconds, so this should not be a big deal at all.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Dump monitor resolution and refresh rate to the log on launch.
## Testing
Tested with two monitors.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Allow the server to override client's quality options for screen
mirroring.
On Windows / Android versions (flutter-based) clients, this is already
in the options of those apps. However, the iOS version lacks the quality
options.
## Testing
Tested with iOS and Windows clients.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Mark the option as "sensitive" which censors the option value (adapter's
IP address)
Remove the log entry from spice
This isn't perfect because most of the games like to log the IP address
and subnet anyway, but at least spice itself will stop doing it.
## Testing
Tested 64 bit LDJ with parameter set.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Always log the DLL identifier string in patch manager, even if there are
no patches.
## Testing
Tested 32 / 64 bits.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Log a message if d3d9.dll (and friends) are found in the game root or
the modules directory.
This is to help with troubleshooting in cases where the user is unaware
that d3d9 is being used to inject other DLLs (e.g., ifs_layeredfs), or
when DX9 reimplementation is causing issues / crashes (e.g., dxvk).
## Testing
Observed the new messages being logged in both modules / root directory.
## Link to GitHub Issue, if one exists
Fixes#307
## Description of change
Add an option for making TDJ/UFC subscreen window borderless in windowed
mode
Add another option for keeping the subscreen "always on top".
Add size / position arguments to SDVX windowed subscreen, exactly the
same as what is allowed for IIDX.
## Testing
Seems to work....
## Link to GitHub Issue, if one exists
#304
## Description of change
Complain loudly about DLL conflicts. Show warnings for each DLL that
exists in both modules directory and the root.
This is not based on LoadLibrary hooks, just enumerating DLLs via
filesystem access.
Deliberately not using `log_fatal` since these are not necessarily going
to crash the game. Some people just love to have messy installs and if
the game booted fine before, it should continue to boot.
## Testing
Tested with/without `-modules`, with/without duplicate DLLs.
## Link to GitHub Issue, if one exists
n/a
## Description of change
The check to warn the user about potential MAX_PATH violation should be
stricter. Saw a user with 150 characters run into a crash during boot
for SDVX3.
## Testing
Tested paths under 130 / over 130, in both 32/64 bits.
## Link to GitHub Issue, if one exists
n/a
## Description of change
For JSON config files of the following features:
* patch manager
* screen resize
* IIDX camera hook
* card manager
when saving a new file, store them in %appdata%\spice2x instead of
%appdata%.
On load:
1. If the JSON file exists in %appdata%\spice2x, use that (new path)
1. If the JSON file exists in %appdata%, continue to use that (legacy
path)
It's common for people to have mixed versions of spicetools/spice2x so
we'll continue to read from the %appdata% root if the files are there,
but with a preference for the new path. We will not forcibly move files.
spicetools.xml will continue to live in the %appdata% root. Moving this
will confuse a lot of people, so I'm avoiding this.
Also, this fixes `-patchcfgpath` and `-resizecfgpath` to create
directories as needed (previously the parent directory must have existed
first)
## Testing
Tested -
* existing config files are continued to be read from %appdata%
* new files get created in %appdata%\spice2x\... (both in spicecfg and
in overlay)
* can provide custom path for `-patchcfgpath` `-resizecfgpath` and
observe directories + file created in custom path, try absolute or local
relative paths
## Link to GitHub Issue, if one exists
Fixes#280
## Description of change
Introduce a variant of 32bit spice.exe that is large address aware.
## Compiling
👍
## Testing
Minimal testing has been done to see if the LAA variant can boot
existing 32 bit games.
## Link to GitHub Issue, if one exists
Fixes#277
## Description of change
To avoid confusion, temporarily remove `-sdvxlandscape` because of the
issue described in the bug above.
`forceresswap` is still kept, so there is still a way to do exactly what
`-sdvxlandscape` used to do, if someone really wants it.
## Testing
Checked that `-sdvxlandscape` is ignored, and that `-sdvxlandscape`
continues to be honored.
## Description of change
Adds `ccache` support for local compiling and Github Actions
environments. This significantly speeds up incremental builds, but cold
starts will take a little longer to compile.
The build script now has an `-i` flag to ignore the build cache if
needed. You can also manually delete the `.ccache` folder to force a
fresh build too.
For the GA workflow, it uses the
[`ccache-action`](https://github.com/hendrikmuhs/ccache-action) action
to manage and store the cached build objects. This cache should be
accessible to all branches.
I also removed the `--no-cache` flag when building the
`spicetools/spice` image since Docker will automatically invalidate
`COPY` layers if it detects the copied files have changed.
### Compile times
Compiled with a 7800X3D, 32GB 6000MT/s RAM.
| `main` | `ccache` (fresh) | `ccache` (cached) | `ccache` (ignore
cache) |
|--------|--------|--------|--------|
| 2m46.047s | 3m8.886s | 0m32.638s | 2m45.148s |
## Testing
Tested incremental builds by using another branch that's rebased on top
of this.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Previously, the artifact produced by the CI action was `spice2x.zip`.
With this change, it will be `spice2x-ci-1234567.zip` where 1234567 is
the first 7 chars of the commit hash.
Note that for PRs, github produces a merge commit before queueing the
action, so the commit hash won't match up with the HEAD of a fork. The
goal is to just generate unique-ish filenames so this is not a huge
deal.
## Compiling
## Testing
Action has been run in the fork.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Change the usage pattern for Card Manager UI.
### Previous behavior
Previously, it was a list of cards that you select, and click a button
to insert as P1 or P2.
This is fine, but users got confused when they pressed `Insert Card`
overlay key and got a different card inserted, or when auto-insert
didn't work as expected.
Additionally, if you play a game without continue (IIDX/DDR for
example), then it's cumbersome to bring up the overlay and click on
insert card every time.
### New behavior
Insert Card overlay now lets you pick a card and *slot* it into P1 or P2
side.
From that point on, the slotted card number is used as the override,
replacing what was previously passed to `-card0` or `-card1`. This means
that `Insert Card` button and auto-insert will pick up the new card and
use that instead.
## Compiling
👍
## Testing
Tested 1p-only game (popn) and 2p games (DDR, IIDX).
## Link to GitHub Issue, if one exists
Fixes#295
## Description of change
Add missing `Start` and `10KEY` lights output, as see in the test menu.
## Compiling
👍
## Testing
Seems fine in museca test mode.
## Link to GitHub Issue, if one exists
Fixes#292
## Description of change
This PR adds CI with Github Actions. Runs on every push and pull
request. If a new commit is pushed, any previous runs are cancelled for
that branch.
In total, these optimisations (on a 7800X3D with 32GB 6000MT/s RAM and
an NVMe Gen4 SSD):
* Reduces `spicetools/deps` image size by around 200 MB, and build time
by almost 50%
* Reduces compilation time by around 9%
Changes:
* Pull archlinux image from GHCR so that we don't run into pull limits
from docker.io
* Use `base-devel` archlinux image
* Reduces image build time from 141 seconds to 109 seconds on my machine
* Clean up unneeded packages
* Small reduction in `spicetools/deps` image size (4.35 GB vs 4.14 GB)
* Use precompiled `yay` binary
* Reduces image build time from 163 seconds to 141 seconds on my machine
* Use `nproc` to get the number of cores instead of parsing
`/proc/cpuinfo` with `awk`
* Use `ninja` to compile
* Speeds up compile times a little, around 13 seconds on my machine
* Don't run `docker run` in interactive TTY in `build_docker.sh`
* Github Actions disallows interactive TTYs
* Still allowed for the batch script
* `time` the compilation process
* Enable BuildKit in the bash script
* Required for build contexts
* Don't use `USER` instruction
* Breaks Github Actions, doesn't seem to affect build?
* More info:
https://docs.github.com/en/actions/sharing-automations/creating-actions/dockerfile-support-for-github-actions#user
* `USER` also seems to cause issues on a mounted ReFS image
* Add executable bits to bash scripts
Future ideas:
* Combine `external/Dockerfile` and `./Dockerfile` into one to
deduplicate images and save disk space
* Fix the weird `/src/src` directory structure in the Dockerfile and
build scripts
* Investigate CMake and ninja optimisations?
## Compiling
Check CI status on my fork.
## Testing
Tested with docker on Windows. Also play tested a bit to make sure the
Github Actions artifact works.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Update tooltip for `-monitor` option. Move to `Graphics (Full Screen)`
to reduce clutter.
## Compilin
👍
## Testing
n/a
## Link to GitHub Issue, if one exists
Fixes#120
## Description of change
Don't call `is_touch_available` in launcher before RawInput stack is
initialized. This causes us to incorrectly fall back to the Win7/Win8
touch handler.
The desired behavior is to use rawinput touch by default for all games,
unless overridden by the user via `-wintouch`.
Add a bunch of logging so we can spot who's calling
`is_touch_available`.
As a side effect, a handful of games that previously used Win7/8 touch
will now use RawInput touch instead. If this causes any issues, they can
turn `-wintouch` on to force Win7/8 touch logic again.
## Compiling
👍
## Testing
Test:
* RB
* -wintouch
* TDJ, UFC subscreen and native touch
* overlay, especially in multi-mon
## Link to GitHub Issue, if one exists
n/a
## Description of change
No one uses this, remove them to reduce clutter. This is in preparation
to migrate from deprecated ImGui IO model to new one, and this is part
of that.
## Compiling
👍
## Testing
## Link to GitHub Issue, if one exists
n/a
## Description of change
Update ImGui to `v1.91.4`. We can't update to `v1.91.5` or newer just
yet because impl_spice depends on the old IO model of ImGui which got
removed in `v1.91.5`.
We also can't update the DX9 renderer right now, because at some point
overlay breaks in TDJ (and only in TDJ...)
Enable the new debug logging functionality in ImGui. Add a button in
overlay to toggle debug log window.
Fix some minor bugs spotted by the new ImGui debug log: mismatched API
calls, font loading issues.
## Compiling
👍
## Testing
In progress.
## Link to GitHub Issue, if one exists
n/a
## Description of change
Functionally, no change.
Move ImGui library files around so that they mirror how they are
structured in imgui releases. This will make it easier to migrate to a
future version of ImGui.
Add a handful of static asserts to check that we have the necessary
features enabled / disabled in imgui header.
## Compiling
👍
## Testing
Tested 32-bit overlay, 64-bit overlay, TDJ, and UFC overlays, and
spicecfg.
Interesting finds during ImGui testing (not part of this PR, but for
future reference when upgrading imgui version):
* Updating dx9 backend to the newest causes TDJ to stop rendering the
overlay completely (fine in UFC or any other game)
* Omitting `IMGUI_USE_BGRA_PACKED_COLOR` causes spicecfg software
renderer to use the wrong color (RGB format parsing error probably)
* At some point - v1.90? the `About` and `Licenses` buttons in spicecfg
stop responding after an upgrade, maybe our improper usage of MenuItem?
* Using the docking branch of ImGui is not required but we should just
stick to it for multi-viewport functionality which we may want to use
later.
## Link to GitHub Issue, if one exists
n/a
## Description of change
`ImGui::Columns` is a deprecated component with many issues. Migrate
Buttons/Analog/Overlay/Lights/Options... tabs to the new table API.
Clean up some UI elements around it to reduce clutter.
Note: there are still other usages of Columns API (MIDI debug output,
ACIO viewer, window manager) but no one hardly ever sees those.
Eventually they will need to be migrated if ImGui ever drops support for
them.
## Compiling
👍
## Testing
Testing all games in spicecfg and overlay.
## Link to GitHub Issue, if one exists
Fixes#27
## Description of change
In standalone configurator mode, when a mouse cursor moves outside the
window and comes back, it may get stuck in the wrong cursor type (e.g.,
get stuck in horizontal resize mode). This is a side effect of software
rendering we do that doesn't deal nicely with mouse events (i.e., not at
all).
## Compiling
Should be fine.
## Testing
Testing spicecfg, spice -cfg, spice64 -cfg, and in-game overlay.
## Link to GitHub Issue, if one exists
#0
## Description of change
When patcher URL returns an invalid response (as in, cannot be parsed as
JSON), show an error message to user.
Add a "copy text" button to copy the error message.
Fix WarnTooltip so that can handle the case where description is an
empty string for a patch - previously this was just showing as multiple
lines of empty space and then the warning.
## Compiling
It's fine except Docker build currently fails due to cmake version
issue.
## Testing
Tested valid / invalid datecodes and checked against responses.
bicarus-dev