Fixes#2908
When loading `config-file`, we need to ensure that all loaded
configuration is loaded _prior_ to any `-e` values from the CLI.
To do this, I inserted a new `-e` special tag type in our replay steps.
This can be used to find when `-e` starts and ensure it remains at the
end of replay steps when the replay steps are being modified.
This commit also found a similar (but not exercised) issue where this
could happen with light/dark themeing.
❤️👻
This is the icon that we'll launch Ghostty 1.0 with. It was designed by
Michael Flareup at PixelResort. It retains the style of the original
Ghostty icon by Alasdair Monk, but brings in the new Ghost character and
adds details that make it more Apple-like.
The new Ghost character is an important evolution from the original
since it separates us from looking too much like PacMan. The new Ghost
is more unique and recognizable to Ghostty (or, hopefully will be!).
The icon itself has more details: the aluminum around the edge has
texture for the large enough sizes, there are visible scanlines, the
glow of a screen emanates from the ghost.
The icon itself is stylistic more Apple-like than other platforms. I
think Apple icons tend to look very good in more environments than the
reverse and I'm a big fan of the Apple aesthetic so I wanted to bring
that to Ghostty for all platforms.
This is probably going to cause some issues, this is the soonest we've
ever tried this. CI seems quite happy about the change though so moving
forward quickly is good with me.
I'm unsure if this is an environmental issue just for me or if this is
more widespread or what other downsides this may have. I'm more than
willing to revert this if it ends up causing different issues.
I found that with NixOS 24.11 and GTK 4.14 on my system, the default
Wayland GDK backend fails to initialize EGL. With GTK 4.16 everything
is fine.
If I force X11 then everything also works fine. This commit forces X11
for GTK 4.14 specifically (4.16+ is allowed to use Wayland).
Fixes#2877
As the comment in the diff states, we rely on `mmap` to zero our memory.
When we reset we are reusing previously allocated memory so we won't hit
an `mmap`. We need to zero the memory ourselves.
This is pretty slow if there is a lot of memory but in every case except
allocation failures, we expect there to be only a few pages allocated.
Fixes#2857
Some terminal modes always reset, but there are others that should be
conditional based on how the terminal's default state is configured.
Primarily from #2857 is the grapheme clustering mode (mode 2027) which
was always resetting to false but should be conditional based on the
the `grapheme-width-method` configuration.
Fixes#2848
The proper way to convert a unicode scalar in Swift is to use the
`String` initializer that takes a `UnicodeScalar` as an argument. We
were converting a number to a string before, which is incorrect.
Fixes#2781
This commit contains two separate changes but very related:
1. We update the color scheme state of the app on app start. This is
necessary so that the configuration properly reflects the conditional
state of the theme at the app level (i.e. the window headerbar).
2. We take ownership of the new config when it changes, matching macOS.
This ensures that things like our GTK headerbar update when the theme
changes but more generally whenever any config changes.
And some housekeeping:
- We remove runtime CSS setup from init. We can do it on the first tick
of `run` instead. This will probably save some CPU cycles especially
when we're just notifying a single instance to create a new window.
- I moved dbus event setup to `run` as well. We don't need to know these
events unless we're actually running the app. Similar to the above,
should save some CPU time on single instance runs.
This sets the macOS QoS class of the renderer thread. Apple
recommends[1] that all threads should have a QoS class set, and there
are many benefits[2] to that, mainly around power management moreso than
performance I'd expect.
In this commit, I start by setting the QoS class of the renderer thread.
By default, the renderer thread is set to user interactive, because it
is a UI thread after all. But under some conditions we downgrade:
- If the surface is not visible at all (i.e. another window is fully
covering it or its minimized), we set the QoS class to utility. This
is lower than the default, previous QoS and should help macOS
unschedule the workload or move it to a different core.
- If the surface is visible but not focused, we set the QoS class to
user initiated. This is lower than user interactive but higher than
default. The renderer should remain responsive but not consume as
much time as it would if it was user interactive.
I'm unable to see any noticable difference in anything from these
changes. Unfortunately it doesn't seem like Apple provides good tools to
play around with this.
We should continue to apply QoS classes to our other threads on macOS.
[1]: https://developer.apple.com/documentation/apple-silicon/tuning-your-code-s-performance-for-apple-silicon?preferredLanguage=occl
[2]: https://blog.xoria.org/macos-tips-threading/
Fixes#2817
The test is pretty explanatory. I also renamed `end` to `count` since I
think this poor naming was the reason for the bug. In `eraseChars`, the
`count` (nee `end`) is the number of cells to erase, not the index of
the last cell to erase.
