Related to #1802
CSIu format sends a unicode codepoint but we were just sending the first
byte of the utf8 sequence, which is very much not right. This fixes that
by parsing the utf-8.
It isn't defined what to do if the utf-8 sequence is invalid or has
multiple codepoints so we just skip CSIu encoding in that case.
Reproduction: HU layout Ctrl+ő
On release, we were previously not sending an alt text (we were sending
it properly on press). Kitty sends it also on release and the spec makes
it clear we should send it on release. This was just due to some faulty
logic; added a test and fixed that.
Fixes#1802
This allows `keybind` configurations to map to any Unicode codepoint. This enables keybindings for which we don't have a registered keycode or for custom keyboard firmwares that may produce arbitrary text (but the Ghostty support is limited to a single codepoint).
The `keybind` syntax is unchanged. If a bound character doesn't map to a known logical key that Ghostty knows about, we map it to a Unicode codepoint. The unicode codepoint is compared against the _unshifted codepoint_ from the apprt key event.
Note that this binding is to a single _codepoint_. We don't support arbitrary sequences of characters or multi-code point graphemes for keybindings due to the complexity in memory management that would introduce.
This also provides a good fallback for scenarios where it might make sense to educate Ghostty about a key code or fix a bug in our keyboard input system, but the unicode data is correct. In that scenario, unicode key binds should allow key binds to still work while we investigate the input issues.
Example:
```
shift+ö=text:hello
```
This now works as expected on a US hardware keyboard with the Hungarian keyboard layout.
Allows for high dpi displays to get odd numbered pixel sizes, for
example, 13.5pt @ 2px/pt for 27px font. This implementation performs
all the sizing calculations with f32, rounding to the nearest pixel
size when it comes to rendering. In the future this can be enhanced
by adding fractional scaling to support fractional pixel sizes.
In korean input method on macos, it should simply return empty string.
Issue was created after https://github.com/mitchellh/ghostty/pull/1659.
```
gksr<BS> # 한ㄱ<BS>
```
'ㄱ' at the end should be removed with single <BS>, but for now
it requires two <bs> to remove 'ㄱ'.
Fixes#1618
Font sizes in configuration were always a u8, but the keybinding and
internal state was a u16 so it allowed for an ever-growing font size. At
a certain point, there is an integer overflow which causes it to wrap
around. This is all silly, 255 should be large enough for anyone[1]
[1]: Ready to be super wrong about this
Ghostty does not report associated text on macOS when
macos-option-as-alt is enabled for _any_ key press, whether or not the
Alt modifier is actually present. The "option as alt" decision should
only be made when the alt modifier is present.
This follows a specific behavior Kitty has, but no other terminal seems
to have, but it is a reasonable behavior that is in use by real programs
as found by our beta testing community.
We diverge from Kitty though in that we only apply this to ASCII
letters. Kitty does not CSIu encode special characters like `@` or `$`.
For these characters, Ghostty still encodes it as specified by fixterms.
Related to #1507
The comments in this change have most of the detail. The primary gist:
- caps-lock is handled correctly so #1501 is still fixed
- only characters pressed with the shift key are affected
Fixes#1505
I verified this behavior with every other terminal and I've added test
cases for it. We previously had a test case to assert the opposite,
which is incorrect.
Fixes#1416
At a high level, the issue is that when mouse capture is enabled (i.e. in
neovim), "shift" escapes the capture. So "cmd+shift" is equal to "cmd"
which doesn't get sent to the TUI program and so on. For link
highlighting which now requires "cmd" (super) is held, we were sending
"cmd+shift" to the renderer so we weren't checking for links.
So the core of this commit is respecting this scenario and stripping the
shift modifier.
This commit also found that when the mouse wasn't over a link, we were
always checking and highlighting links on line one of the visible
screen. This bug is fixed and should also result in a very slight
performance improvement on rendering in all cases.
Previously, we encoded `ctrl+_` in the CSIu format[1]. This breaks most
notably emacs which expects the legacy ambiguous encoding.
This commit utilizes the generator from Kitty to generate our control
key mappings. We also switch from keycode mapping to key contents
mapping which appears to be the correct behavior also compared to other
terminals.
In the course of doing this, I also found one bug with our fixterms
implementation. Fixterms states: "The Shift key should not be considered
as a modifier for Unicode characters, because it is most likely used to
obtain the character in the first place (e.g. the shift key is often
required to obtain the ! symbol)." We were not applying that logic and
now do.
[1]: https://www.leonerd.org.uk/hacks/fixterms/
This gets `zig build -Dtarget=aarch64-ios` working. By "working" I mean
it produces an object file without compiler errors. However, the object
file certainly isn't useful since it uses a number of features that will
not work in the iOS sandbox.
This is just an experiment more than anything to see how hard it would be to
get libghostty working within iOS to render a terminal. Note iOS doesn't
support ptys so this wouldn't be a true on-device terminal. The
challenge right now is to just get a terminal rendering (not usable).
Fixes#1099
We previously applied application keypad mode logic (`ESC=` or mode 66)
whenever it was active. However, from looking at the behavior of other
terminals (xterm and foot) it appears this isn't correct.
For xterm, application keypad mode only applies unconditionally if the
keyboard mode is VT220 (`-kt vt220`). For modern terminals, application
keypad mode is only applied if mode 1035 is disabled.
Mode 1035 is the "ignore numpad state with keypad mode" mode. It
defaults to true on terminal startup. If this is true, keypads are
always encoded in numerical mode. If this is false, the numlock state
will be respected.
Associated text should only be reported when it was generated from the
key event. To my knowledge, there are no release events which produce
text. This now matches kitty reporting for release events.
> When the key event is related to an actual modifier key, the corresponding
> modifier's bit must be set to the modifier state including the effect for the
> current event. For example, when pressing the :kbd:`LEFT_CONTROL` key, the
> ``ctrl`` bit must be set and when releasing it, it must be reset. When both
> left and right control keys are pressed and one is released, the release event
> must have the ``ctrl`` bit set. See :iss:`6913` for discussion of this design.
Fixes#1052
This implements a `close_all_windows` binding in the core and implements
it for macOS specifically. This will ask for close confirmation if any
surface in any of the windows requires confirmation.
This is bound by default to option+shift+command+w to match Safari. The
binding is generall option+command+w but users may expect this to also
mean "Close All Other Tabs" which is the changed behavior if any tabs
are present in a standard macOS application. So I chose to follow Safari
instead.
This doesn't implement this feature for GTK, that's left as an exercise
for a contributor.
This allows bindings with `tab` to work properly on Linux. The issue is
that in the key translation, we weren't mapping this and thought it was
invalid IME input so we were ignoring it.
