Make the foreground_color and background_color fields in the Terminal
struct optional values so that we can determine if a foreground or
background color was explicitly set with an OSC 10 or OSC 11 sequence.
This makes the logic a bit simpler to reason about (i.e.
`foreground_color` is now always "the color set by an OSC 10 sequence"
while `default_foreground_color` is always "the color set by the config
file") and also fixes an issue where an OSC 10 or OSC 11 query would not
report the correct color after a config update changed the foreground or
background color.
The `cursor_color` field was already optional, with the same semantics
(it is only non-null when explicitly set with an OSC 12) so this brings
all three of these fields into alignment.
Fixes#2791
This goes back to the previous implementation of clone that directly
cloned values rather than using replay steps, because replay steps don't
preserve conditionals on the iterator.
This works.
Another fix will be when we support some sort of conditional syntax and
the replay iterator can yield that information. We have a unit test here
so we can verify that then.
The prior light/dark mode awareness work works on surface-level APIs. As
a result, configurations used at the app-level (such as split divider
colors, inactive split opacity, etc.) are not aware of the current theme
configurations and default to the "light" theme.
This commit adds APIs to specify app-level color scheme changes. This
changes the configuration for the app and sets the default conditional
state to use that new theme. This latter point makes it so that future
surfaces use the correct theme on load rather than requiring some apprt
event loop ticks. Some users have already reported a short "flicker" to
load the correct theme, so this should help alleviate that.
Related #2755
From the mode 2031 spec[1]:
> Send CSI ? 2031 h to the terminal to enable unsolicited DSR (device status
> report) messages for color palette updates and CSI ? 2031 l respectively to
> disable it again.
>
> The sent out DSR looks equivalent to the already above mentioned. This
> notification is not just sent when dark/light mode has been changed by the
> operating system / desktop, but also if the user explicitly changed color
> scheme, e.g. by configuration.
My reading of this paired with the original discussion is that this is
meant to be sent out for anything that could possibly change terminal
colors.
Previous to this commit, we only sent out the DSR when the actual system
light/dark mode changed. This commit changes it to send out the DSR on
any operation that _may_ change the terminal colors.
[1]: https://contour-terminal.org/vt-extensions/color-palette-update-notifications/#example-source-code
I genuinely don't know what they do, but Xcode recommended it and
cursory docs reading seems to indicate its safe. I don't think it'll
result in any noticeable changes.
Previously, we would access the `ghostty.config` object from anywhere.
The issue with this is that memory lifetime access to the underlying
`ghostty_config_t` was messy. It was easy when the apprt owned every
reference but since automatic theme changes were implemented, this isn't
always true anymore.
To fix this, we move to the same pattern we use internally in the core
of ghostty: whenever the config changes, we handle an event, derive our
desired values out of the config (copy them), and then let the caller
free the config if they want to. This way, we can be sure that any
information we need from the config is always owned by us.
Fixes a crash found in Discord.
Cloning the keybinding set previously shallow copied the actions, but
actions may contain pointers. These pointer values must be deep copied
to avoid dangling references when the underlying memory is freed.
Fixes#2747
I admit I don't fully understand this. But somehow, doing `var x: ?T =
undefined` in release fast mode makes `x` act as if its unset. I am
guessing since undefined does nothing to the memory, the memory layout
is such that it looks null for zeroed stack memory. This is a guess.
To fix this, I now initialize the type `T` and set it onto the optional
later. This commit also fixes an issue where calling `parseCLI` multiple
times on an optional would not modify the previous value if set.
