Use clone3 / CLONE_INTO_CGROUP to have the Linux kernel create the process in the
correct cgroup rather than move the process into the cgroup after it is created.
For now, bash integration must be explicitly enabled (by setting
`shell-integration = bash`). Our automatic shell integration requires
bash version 4 or later, and systems like macOS continue to ship bash
version 3 by default. This approach avoids the cost of performing a
runtime version check.
Fish automatic integration taken as an example.
Just like fish, Elvish checks `XDG_DATA_DIRS` for its modules.
Thus, Fish integration in zig is reused, and integration in
Elvish now removes `GHOSTTY_FISH_XDG_DIR` environment variable
on launch.
When the -c option is present, then commands are read from the first
non-option argument command string. Our simple implementation assumes
that if we see at least the '-c' option, a command string was given, and
the shell is always considered to be non-interactive - even if the '-i'
(interactive) option is also given.
bash reads HISTFILE at startup to locate its history file, but this is
apparently too early for it to be able to expand home-relative paths. We
now manually expand the full path and add that to the environment.
This change adds automatic bash shell detection and integration.
Unlike our other shell integrations, bash doesn't provide a built-in
mechanism for injecting our ghostty.bash script into the new shell
environment.
Instead, we start bash in POSIX mode and use the ENV environment
variable to load our integration script, and the rest of the bash
startup sequence becomes the responsibility of our script to emulate
(along with disabling POSIX mode).
When a shell is forced, we would supply its /-prefixed executable name
to mimic a path location. The rest of the integration detection logic
assumes just a base executable name. Fix the forced names accordingly.
Also add a unit test for this "force shell" behavior.
We don't actually use libc wcswidth to determine width and even if we
did some terminals use wcwidth (which behaves differently), some use
the Python wcswidth library (which behaves differently), etc. The
reality is there is no real consistency on "legacy" behavior so by
naming it "legacy" we show that we're doing our best but also gives us
wiggle room to change our behavior in the future.
Functionally nothing changes with this commit.
This adds a new option to the shell integration feature set, `no-title`.
If this option is set, the shell integration will not automatically
update the window title.
Using a saturating addition here just to avoid overflow, since setCursorPos handles proper clamping to the screen size so we don't need to duplicate that logic.
Fixes#1514
We previously required all preedit inputs to fit into the small copied
message size. That's true for 99% of all inputs, but if a long pre-edit
input comes in, this may not be true. We should try the small array
fast-path but fall back to heap allocation if we must.
Fixes#1403
This changes the behavior of `grapheme-width-method = unicode` to change
the default state of mode 2027 to true. Prior to this, setting this
config would force grapheme clustering regardless of mode 2027. Now,
this only sets the default and running TUI programs can disable it if
they want.
Fixes#1401
SCOSC is ambiguous with regards to DECSLRM. This commit copies the logic
of xterm: if left/right mode is enabled, then CSI S is always DECSLRM.
But if left/right mode is disabled then CSI S empty always uses SCOSC.
SCORC always works.
A lot of the state that we put on Exec is just there to copy to
StreamHandler, but we already have it in DerivedConfig. I think this
whole copy copy copy is just legacy cruft since termio.Exec is one of
the older parts of the source code.
This rearchitects the Exec struct to act more like Surface and Renderer
where it stores its derived config. This lets us avoid a few extra
allocations and removes a LOT of struct member noise from termio.Exec.
For pointer lifetimes, the memory allocated is now owned by
DerivedConfig. When changeConfig is called, its the only time BOTH are
still alive, so we can safely swap pointers and deinit without having to
duplicate across threads. This is the same as renderer/surface.