This already happened on 2023-03-16, but I held off on updating the
repo's version a bit long: I wanted to wait a little to batch it up
with the next update, but the next update only arrived today, so...
I noticed that in 2.3, the game would launch with default controller
binds upon first launch (e.g. no pre-existing unlock.vvv or
settings.vvv), but in 2.4, this wasn't the case, and the default binds
would only be set the next time the game was launched. This would result
in you being essentially unable to use the controller on first launch
save for the analogue stick and D-pad to move between menu selections
and move the player.
Bisecting pointed to commit 3ef5248db9 as
the cause of the regression. It turns out returning early upon error or
a file not existing didn't set the default controller binds, because
they were done at the end of Game::deserializesettings(). But the binds
would be set on the next launch because if the file didn't exist, a new
file would be written, not with the default binds, but then the next
launch would read the file, see there were no binds, and then set the
default binds accordingly.
To fix this, I made it so that the default controller binds are set when
Game is initialized. That way, it covers all cases where the game can't
read a settings file.
Ever since 2.3, if you bind a controller button to the "menu" action
(i.e. back/escape) you won't be able to change that button to any other
action, because pressing it anywhere in the binding menu will exit to
the previous menu, without applying the binding.
I know action sets will overhaul everything, but 2.4 may (probably
"should") come out before we have action sets. This part is very
broken, and the fix is very easy: just move the bind-assigning code to
above the menu-return-on-esc code, and add a return.
For some reason, the accessibility option that was meant to disable
flashes doesn't disable ALL flashes, only screen flashes and screen
shaking. This commit disables a lot more, most importantly randomness
in colors, the player flashing on death/respawn, and teleporters
flashing.
This updates the interpolation positions of the player when transforming
into and out of VVVVVV-Man.
Otherwise, it can be seen that the player "zips" quickly during these
transformations if the Secret Lab entrance cutscene is played with
screen effects off.
Unfortunately, 1de459d9b4 caused a
regression where musicclass::niceplay() didn't work, because fading out
the music would cause haltdasmusik() to be called, which would reset the
fade variables.
To fix this, haltdasmusik() will now only reset the fade variables if
it's not called from a fade, which is signaled with a function
parameter.
`platv` is a room property that controls platform speed, and it has
always worked (other than some weird storage issues due to a bug).
However, the editor has no way to edit it currently, so people had to
resort to editing the level file by hand, or with a third-party tool.
This commit simply adds an easy way to modify platform speed.
VED has a fill bucket subtool for tiles and backgrounds, which is
really useful when creating rooms. This commit adds a fill bucket as
well, with an adaptive tool highlight, unlike VED.
This fixes a regression from 2.3 where the very beginning of A New
Dimension isn't silent.
A New Dimension's level music is set to Predestined Fate, but there is a
script box the player touches right upon spawning that stops playing
music. Then after the player ascends upwards, they touch a script box
that plays Predestined Fate. But in 2.2 and before, the very beginning
is silent due to the script box that stops music.
However in 2.3, due to the changes made to playing music during a fade,
the initial level music trying to play Predestined Fate during a music
fadeout from the main menu resulted in Predestined Fate being stored in
the nice fade variables, which kicked in after halting the music since
halting didn't reset those variables.
This resets those variables whenever music is halted, and now the
beginning of A New Dimension is back to how it was in 2.2 and before.
This fixes a regression where the red channel 0 glitch didn't work,
because the surface was always whitened, because LoadSprites would
whiten the image before converting it to surface.
This regression happened because of #923.
Fixes#962.
Misa asked me if this should only work for non-transparent textboxes,
and it shouldn't - that was kind of an oversight.
To make it work for transparent textboxes as well, I made a little
restructuring to avoid duplicating the code - fill_buttons() is now
called textbox_line(), and it replaces the direct accessing of the
textbox lines in the printing loops. The code that checks the width
of the textbox does not need to be copied, since the text box is
naturally not drawn for transparent text boxes.
This makes it so that `CustomEntity`s, at least internally, do not use
global tile position. Instead, they will use room-x and room-y
coordinates, which will be separate from their x- and y- positions.
This makes it much easier to deal with `CustomEntity`s, because you
don't have to divide and modulo everywhere to use them.
Since editorclass::add_entity and editorclass::get_entity_at expect
global tile position in their arguments, I've added room-x and room-y
arguments to these functions too.
Of course, due to compatibility reasons, the XML files will still have
to use global tile position. For the same reason, warp token
destinations are still using global tile position too.
As reported by Lilithtreasure on the VVVVVV Discord server, it is
possible to get gray moving platforms and enemies in the main game.
This happens if you play the main game after loading a custom level with
a room that is gray at the same coordinates. E.g. if you play a custom
level with a gray room at (12, 4), then exit and go to Gantry and Dolly
in the main game which is also at (12, 4), then the platforms there
would be gray too.
This is because there is a missing map.custommode check.
The missing piece from sound effects was handling what to do when the
buffer ran out. Which seems to happen often when decoding from OGG,
unlike WAV. This handling involves callbacks to functions named
refillReserve and swapBuffers.
Without this code,some sound effects would be cut off early, as
documented in #953. This might also explain the division by 20 - which
I've copied too, just in case.
Now OGG sound effect tracks should be identical to music tracks (except
I've stripped the looping code out).
Fixes#953.
We were using this function to check if the format of the existing voice
is different from the format of the voice we intended to play. However,
whereas the format we use is the FAudioWaveFormatEx struct, the only
details we get from FAudioVoice_GetVoiceDetails is the
FAudioVoiceDetails struct, which has much less details and is missing
important things like whether the format is 8-bit or 16-bit or something
else.
And of course, if we don't check that the number of bits matches, then
it still leads to playback issues as described in #953.
There are no other functions in FAudio that operate on
FAudioWaveFormatEx structs. So instead, we'll just have to do it
ourselves, and store the format of the existing voice alongside the
format of the intended voice. And then use SDL_memcmp to make sure the
formats are the same before playing, and if not, then destroy and
re-create the voice.
Fixes#953.
There's a few places where textboxes are constructed through code, but
they pass in the color's RGB values in manually. This commit
unhardcodes most of them them, replacing them with a color lookup.
The ones that weren't changed are special cases, like `175, 174, 174`.
I thought 2.2 already had separate map and interact gamepad bindings,
and they simply got neglected and broken with 2.3's split Enter/E key
option. But actually, the new split Enter/E option also applied to
gamepad buttons, and a separate interact binding was added, without
really indicating anything if Enter and E are not split. And I guess
using the same button for map and interact by default also makes sense
for simplicity...
This commit makes sure the button glyph displayed in-game is at least
the correct button. The gamepad bindings menu is also slightly modified
to darken the interact option - the button glyphs code now
automatically causes them to show equal buttons anyway, so it wasn't
too big of a change to also darken the line and disable the binding
option. To me this says: "the interact key is fixed to be the same as
enter right now, but there is a way to change it."
It's still not ideal of course, and I know a similar change to the
gamepad menu to hide the interact option was rejected a year ago
because action sets would already fix it, but it's a year later now,
and showing misleading button glyphs should be fixed in 2.4, whether it
will already have action sets or not (And at this point I think the
plan already is to keep the existing input system for 2.4)
And it's a 3 line diff to darken and disable the option, compared to
fully hiding the option.
The language screen has a "Press Space, Z, or V to select" hint, which
I forgot to update for supporting button glyphs in #943, so this commit
does.
<action_hint>Press Space, Z, or V to select</action_hint>
<gamepad_hint>Press {button} to select</gamepad_hint>
> CMake is re-running because D:/a/VVVVVV/VVVVVV/desktop_version
> /build_default/CMakeFiles/generate.stamp is out-of-date.
> the file 'D:/a/VVVVVV/VVVVVV/desktop_version/CMakeLists.txt'
> is newer than 'D:/a/VVVVVV/VVVVVV/desktop_version/build_default
> /CMakeFiles/generate.stamp.depend'
> result='-1'
"newer" probably means "the last edit date is newer because it was just
cloned", so let's see if including it in the cache means the last edit
date will be kept the same too... We can freely do this, as the cache
key includes the hash of the file itself.
Since the initial `cmake -G "Visual Studio 17 2022" ..` seems to be a
real bottleneck on the Windows CI (sometimes taking multiple minutes
for seemingly no reason), cache the build folder based on the hash
of CMakeLists.txt. It should be possible to reuse the build folder if
CMakeLists hasn't changed, and if it does change (because someone adds
a source file or library or something), we'll just do that initial
cmake with -G once, and cache that version as well.
https://docs.github.com/en/actions/using-workflows/caching-dependencies-to-speed-up-workflows
For some reason, this test takes a dozen seconds or more. Maybe because
it needs to boot up powershell or something like that? No idea, but
it's just sitting there. The cache action has a return value that
indicates whether there was a cache hit, so there's no need to use
Test-Path, saving CI time.
EDIT: Maybe the SDL cache hasn't ever been working. It seems to cache
30 bytes (huh?) and the folder it caches is C:\SDL, not C:\SDL2-2.24.0
as the folder would look like. So it was probably downloading SDL every
time anyway?
Also, I upgraded the cache@v2 action to cache@v3. I added a cache@v3
myself and noticed this was still using cache@v2, so I figured, why not
future-proof it? And that's when I notice this warning: "Node.js 12
actions are deprecated. Please update the following actions to use
Node.js 16: actions/cache@v2." So it's pretty good timing.
This was easier than I expected - just add an optional buttons="1"
attribute to cutscenes.xml. It's treated like the speaker attribute -
it's only there as context for the translator, and for the cutscene
test.
Due to rebasing messiness and diff noise, it's probably best if pull
requests either don't sync all the language files at all (and only
modify the English ones) OR only do it as a final commit. It's still
something we need to figure out, lol.
If a controller button is pressed, a controller is connected (even at
startup!) or an axis is moved, the game will switch to displaying
controller glyphs. If a keyboard key is pressed or the last controller
is removed, the game will switch to displaying keyboard keys.
This adds mappings from SDL's Xbox-based SDL_GameControllerButton
constants, to glyphs for the following layouts:
- LAYOUT_NINTENDO_SWITCH_PRO,
- LAYOUT_NINTENDO_SWITCH_JOYCON_L,
- LAYOUT_NINTENDO_SWITCH_JOYCON_R,
- LAYOUT_DECK,
- LAYOUT_PLAYSTATION,
- LAYOUT_XBOX,
- LAYOUT_GENERIC,
There may still be errors in these, but they should be mostly correct.
I'm leaving it up to Ethan to make it show the correct button glyphs
for the correct controllers being connected (and possibly to fix these
mappings where needed).
Violet's dialogue now looks like this:
squeak(purple)
text(purple,0,0,2)
Remember that you can press {b_map}
to check where you are on the map!
position(purple,above)
textbuttons()
speak_active
The new textbuttons() command sets the next textbox to replace {b_map}
with the map button, and {b_int} with the interact button. The
remaining keys would be added as soon as they need to be added to
ActionSets.h as well.
This adds a function that converts an action (such as interacting
in-game) to the corresponding button text ("ENTER", "E") or button
glyph (PlayStation triangle, Steam Deck Y, etc). This function
currently only gives the existing ENTERs or Es, because I don't know
how best to detect controller usage, or whether the game is running on
a Steam Deck, or what buttons need to be displayed there. Still, it
should now be really easy to adapt the rendering of keyboard keys to
consoles, controllers, or rebound keys.
To identify the actions that currently need to be displayed, this
commit also adds the initial enums for action sets as described by
Ethan in a comment in #834 (Jan 18, 2022).
These visualize the horizontal gravity line kludge for rooms beside
eachother. When you enter another room, gravity lines which look like
they're connected between the rooms try to have the same activated
state.
Basically, if you're in room (1,4) and you go into (2,4), if a
gravity line in (1,4) is activated (gray, on cooldown) and it's
touching the gravity line in (2,4), that gravity line will also be
activated.
This uses DDA (https://w.wiki/6RSQ) to draw a line between the previous
frame's mouse position, and the current frame's mouse position. This
means that there will no longer be gaps in lines of tiles if you move
your mouse fast enough (which is actually rather slow, so it gets
annoying quickly).
The editor's timestep is no longer hardcoded to 24, as I assume that
was only done so there would be less gaps in lines of tiles drawn.
With interpolation, that is no longer an issue, so I've removed the
editor's special case for the timestep.
Scripts used a weird "hook" system, where script names were extracted
into their own list. This was completely unneeded, so it has been
replaced with using the script.customscripts vector directly.
The script editor has been cleaned up, so the cursor's Y position is
relative to the entire script, rather than what's just displaying on
the screen currently. This simplifies a lot of code, and I don't know
why it was done the other way in the first place.
The script selector and script editor cursors have been sped up, since
both lists can be massive, and waiting 6 frames per line is extremely
slow and boring. This is still slow and boring, but we don't have
proper input repetition yet.
This commit moves everything left out of the previous commit to the
state system. This means a bunch of new functions were added as well,
to avoid the code in each function becoming too huge. A lot of cleanup
was done as well, simplifying logic, merging duplicated code, etc.
This commit does NOT touch "script hooks", script editor logic and
autotiling, as those seem to be their own separate beasts.
While warp lines were being drawn, they also got resized to
automatically fit between collision. In renderfixed, gravity lines are
resized the same way. Doing logic while drawing is very poor practice,
so resizing of these has been moved into logic, and merged together.
Aside from some more cleanup, this commit also removes the very poorly
done right click emulation, when you hold CTRL and click. It never
worked well in the past, and even requires a right click to use, so
there's not really any point to keeping it around.
The drawer could definitely be improved further, however I cleaned up a
little bit of the code duplication. I'll have to take a closer look
some other time, but I'm pretty sure that the duplicated code at the
bottom can be removed with a few tweaks, but I'll do that carefully
in a different commit.
Tools were a mess, spread all over the code with hundreds of `else-if`
statements. Instead of magic numbers denoting tools, an enum has been
created, and logic has been extracted into simple switch/cases, shared
logic being deduplicated.
The base of a state system for the editor has been created as well,
laying a good path for further organization improvements. Because of
this, the entire editor no longer gets drawn underneath the menus,
except for a few pieces which I haven't extracted yet. Either way,
this should be good for performance, if that was a concern.
I have this annoying issue where the game will open on the wrong monitor
in fullscreen mode, because that monitor is considered to be display 0,
whereas the primary monitor I want is display 1.
To mitigate this somewhat, the game now stores the display index that it
was closed on, and will save it to settings. Then the next time the game
opens, it will open on that display index. This should work pretty well
as long as you aren't changing your monitor setup constantly.
Of course, none of this applies if your window manager is busted. For
example, on GNOME Wayland, which is what I use, in windowed mode the
game will always open on the monitor the cursor is on, and it won't even
be centered in the monitor. But it works fine if I use XWayland via
SDL_VIDEODRIVER=x11.
Previously, the game would not store the size of the window itself, and
would always call SDL_GetRendererOutputSize() (via
Screen::GetWindowSize()) to figure out the size of the window. The only
problem is, this would return the size of the whole monitor if the game
was in fullscreen mode. And the only place where the original windowed
mode size was stored would be in SDL itself, but that wouldn't persist
after the game was closed.
So, if you exited the game while in fullscreen mode, then your window
size would get set to the size of your monitor (1920 by 1080 in my
case). Then when you opened the game and toggled fullscreen off, it
would go back to the default window size, which is 640 by 480.
This is made worse, however, if you were in forced fullscreen mode when
you previously exited the game in windowed mode. In that case, the game
saves the size of 1920 by 1080, but doesn't save that you were in
fullscreen mode, so opening the game not in forced fullscreen mode would
result in you having a 1920 by 1080 window, but in windowed mode.
Meaning that not even fullscreening and unfullscreening would put the
game window back to normal size.
The solution, of course, is to just store the window size ourselves,
like any other screen setting, and only use GetWindowSize() if needed.
And just to make things clear, I've also renamed the GetWindowSize()
function to GetScreenSize(), because if it was named "window" it could
lead one to think that it would always return the size of the screen in
windowed mode, when in fact it returns the size of the screen whatever
mode it is in - fullscreen size if in fullscreen mode and window size if
in windowed mode.
And doing this also fixes the FIXME above Screen::isForcedFullscreen().
This fixes the following bug that only occurs on Wayland: If the game is
configured to be fullscreened and in stretch mode, on startup, it won't
be in stretch mode. It will appear to be in letterbox mode, but the game
still thinks it's in stretch mode.
This is because during the ResizeScreen() call on startup, for whatever
reason, the window size will be reported to be the default size (640 by
480) instead of the screen resolution of the monitor, as one would
expect from being in fullscreen. It seems like when the game queries the
window size, the window isn't actually in fullscreen at that time, even
though this is after fullscreen has been set to true.
To fix this, I decided to always update the logical size before
SDL_RenderPresent() is called. To make this neater, I put the scaling
code in its own function named UpdateScaling().
This bug has existed since 2.3 and does not occur on X11. I tested this
on GNOME Wayland, and for testing it on X11, I used Openbox in a Xephyr
session while running VVVVVV with SDL_VIDEODRIVER=x11.
