| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Tested-by: Yuri Vostrenkov
|
|
|
|
The `scaled_rotation' gimbal lock "fix" didn't take
into account sign changes, making center close to +-180
useless. remove the "fix".
This leaves with a gimbal lock problem at +-90 yaw. Fix
in another commit.
|
|
Tool: codespell
|
|
|
|
C++17 isn't fully supported.
|
|
|
|
we don't need to stop again, or else it crashes app
|
|
|
|
Using the close button must exit the program regardless
of whether the tray is enabled. Restore that behavior.
Reported-by: @albertolg
Issue: #750
|
|
|
|
|
|
Issue: #748
|
|
The pipeline code is hard to understand for programmers due to the
practice of putting all logic into a single function.
Split it up into logical components.
This needs more testing and bug reports. It's a big change with possibly
many regressions all over the place.
Issue: #688
|
|
|
|
|
|
|
|
It can cause problems for `freetrackclient.dll'
consumers. This fix is better than modulo rand().
|
|
Issue: #748
|
|
|
|
Issue: #748
|
|
|
|
The file must be in tracker-pt/ since it translates .ui
file located there.
Issue: #748
|
|
Remove useless abstract member functions, simplify
some.
Issue: #718
|
|
Some options weren't working, like -silent
Issue: #748
|
|
|
|
Remove cruft, adjust usages.
|
|
- move static variables into block scope if used in
only one function
- simplify control flow
|
|
- reproducible builds; -Wl,--no-insert-timestamp
- add HIGHENTROPYVA for 64-bit builds
Reducing code size is a pointless exercise, still:
- remove .eh_frame that only GNU uses
- reduce 32-bit code size; -march=pentium4
- add few feel-good CFLAGS. maybe reduces code size,
maybe doesn't do anything
|
|
Qt dpiawareness=1 doesn't look like it's messing up
after resolution changes. Use it for now.
|
|
DataID can be idempotent. It's "static unsigned
frameno" in npclient that's used in place of DataID.
|
|
- use `static constexpr inline' to avoid requiring
explicit declarations in object code
- use `const Foo* const' to maybe put into readonly
binary segment (at least for ELF DSOs)
- `constexpr' in function scope has storage, avoid
`static'
- don't use `constexpr' where there's no advantage,
like arrays
We'd like to avoid overhead of atomic initialization
for each function call. No idea how `static constexpr'
requiring storage in the standard plays with atomic
initialization requirement. Hearsay points that
`constexpr' without `static' in block scope behaves
more to our liking. It's all hazy though.
I'm not 100% sure if `static inline constexpr' has any
storage. Hopefully none, like a #define, and stuff
bigger than registers gets coalesced within the same
module, with small stuff being immediates.
|
|
Looks strange ingame.
|
|
This can be reproduced with hidpi, playing Falcon BMS
in a resolution lower than the desktop resolution. The
windows got resized to way more than as if they were
maximized.
This is a workaround, from the Qt API perspective it
doesn't make sense.
|
|
Causes a warning in Visual Studio. Rightly so.
|
|
Visual Studio runs "lrelease" targets without their
"lupdate" dependencies. Add a crude workaround.
Spotted by: @cpuwolf
Issue: #748
|
|
|
|
|
|
It's too complex to be understood when written in
CMake. Replace with explicit calls to set_property().
This hasn't been tested on OSX.
|
|
|
|
Also downcase win32 api header name. This fixes
building for Windows from Linux.
|
|
cf. #748
cf. #746
|
|
tracker/wii: add lang chinese && BT optimization
|
|
|
|
|
|
|
|
|
|
cf. #746
|
|
Now i18n regen won't make a mess.
|
