This is more correct: a pagelist is a linked list of nodes, not pages.
The nodes themselves contain pages but we were previously calling the
nodes "pages" which was confusing, especially as I plan some future
changes to the way pages are stored.
Fixes#2601
It is more expected behavior that `-e` affects only the first window. By
introducing a dedicated configuration we avoid making `-e` too magical:
its simply syntax sugar for setting the "initial-command" configuration.
Fixes#2537
This matches Terminal.app. iTerm2 requires cmd+option (our old
behavior). Kitty doesn't seem to support rectangle select or I couldn't
figure out how to make it work. WezTerm matches Terminal.app too.
Outside of terminal emulators, this is also the rectangular select
binding for neovim.
Mode 2048 and CSI 14 t are size report control sequences which contain
the text area size in pixels. The text area is defined to be the extents
of the grid (rows and columns). Ghostty calculates the available size
for the text area by setting the available padding, and then filling as
much of the remaining space as possible. However, if there are remainder
pixels these are still reported as part of the text area size.
Pass the cell_size geometry through so that we can always report the
correct value: columns * cell width and rows * cell height.
This continues our work to improve the amount of explicit error sets
we use in the codebase. Explicit error sets make it easier to understand
possible failure scenarios, allow us to use exhaustive matching, create
compiler errors if errors are unexpectedly added or removed, etc.
The goal eventually is 100% coverage but we're not even close yet.
This just moves us a little closer.
This simplifies the math for calculating scroll vectors based on mouse
scroll events. This was done to fix inverted horizontal scrolling on
macOS with natural scrolling enabled. Many assertions were added for
assumptions and our preconditions are clearly documented.
The preconditions are:
* Apprt scroll offsets are negative down/left, positive up/right
* Terminal vertical scroll is postive down, negative up (opposite
since scroll for a terminal means how many rows to move down).
* `Surface.scrollCallback` is always call with an apprt offset.
* Apprt is responsible for implementing natural scrolling. Surface
always assumes negative is down/left.
This fixes a longstanding piece of tech debt which caused all sorts of
bugs. Instead of trying to jam cursorPosCallback into a modsChanged
event we bring out the link refreshing into a dedicated method that is
shared.
Ghostty was previously treating shift as a way to always stop mouse
reporting. That's true for mouse button events, but not for mouse
movement events. For mouse movement events, shift should be treated as
a modifier until a button (any mouse button) is pressed. Once it is
pressed, we pause mouse reporting until all buttons are released.
Found by @ldemailly. This matches the behavior of Kitty, Alacritty,
WezTerm, and xterm.
Fixes#2364
This adds the bold, italic, and bold italic variants of JB Mono so it is
built-in. This also fixes up the naming convention for the embedded font
files across tests and removes redundant embedded font files.
Fixes#2345
The new docs for `copy-on-select`:
Whether to automatically copy selected text to the clipboard. `true`
will prefer to copy to the selection clipboard if supported by the
OS, otherwise it will copy to the system clipboard.
The value `clipboard` will always copy text to the system clipboard
(for supported systems) as well as the system clipboard. This is sometimes
a preferred behavior on Linux.
Middle-click paste will always use the selection clipboard on Linux
and the system clipboard on macOS. Middle-click paste is always enabled
even if this is `false`.
The default value is true on Linux and false on macOS. macOS copy on
select behavior is not typical for applications so it is disabled by
default. On Linux, this is a standard behavior so it is enabled by
default.
Previously, once we had one click registered, shift+click would always
go into selection extend mode. This is not the behavior we want, since
we want shift+double/triple click to work in alternate screens.
This commit changes the behavior so that we only extend the selection
after the multi-click interval has passed.
I see a lot of opportunity to improve this whole callback much more but
I don't want to risk introducing new bugs since this is a hard to test
area, so I'm going to leave it for now.
First, this commit modifies libghostty to use a single unified action
dispatch system based on a tagged union versus the one-off callback
system that was previously in place. This change simplifies the code on
both the core and consumer sides of the library. Importantly, as we
introduce new actions, we can now maintain ABI compatibility so long as
our union size does not change (something I don't promise yet).
Second, this moves a lot more of the functions call on a surface into
the action system. This affects all apprts and continues the previous
work of introducing a more unified API for optional surface features.
Fixes#1547
The core change to make this work is to make the cursor position
callback support taking updated modifiers. On both macOS and GTK, cursor
position events also provide the pressed modifiers so we can pass those
in.
Currently, clicking on a desktop notification will bring Ghostty
to the foreground, but it won't necessarily bring the right window
to the top and it won't switch tabs or change the focus on splits.
With this patch, clicking on a desktop notification will raise the
correct window, change to the correct tab, and focus on the correct
split that send the original desktop notification.
The underlying API call returns a sentinel slice so selectionString
should do the same or there are problems later trying to free the
allocated memory.
When the focus reporting mode (1004) is enabled, send the current focus
state. This allows applications to track their own focus state without
first having to wait for a focus event (or query
it by sending a DECSET followed by a DECRST).
Ghostty's focus state is stored only in the renderer, where the termio
thread cannot access it. We duplicate the focus state tracking in the
Terminal struct with the addition of a new (1-bit) flag. We duplicate
the state because the renderer uses the focus state for its own purposes
(in particular, the Metal renderer uses the focus state to manage
its DisplayLink), and synchronizing access to the shared terminal state
is more cumbersome than simply tracking the focus state in the renderer
in addition to the terminal.
