This regressed sometime -- I can't find the exact commit -- but in any
case I've moved this handling directly into the KeyEncoder so we can
unit test it and prevent future regressions.
Fixes#865
Related to #861
In #861, we fixed a deadlock that could happen if the writer mailbox was
full from the reader thread by waking up the writer thread for
processing.
Unfortunately, the writer thread ALSO handles messages that require the
terminal lock (i.e. resizing, focus state, etc.). If the mailbox
contains these messages, it cannot make forward progress on the writes
(which do not require a lock). This makes it possible still under heavy
write scenarios to fully deadlock the read/write threads.
This commit modifies the behavior so that while we are attempting to
queue a writer message after it fails, we release the lock. This is a
very slow path since we are releasing/acquiring locks under heavy
contention. We can improve it in the future but for now its okay because
this is also a rare situation that only happens under the heaviest loads
that also produce heavy writes.
Fixes#741
This completely reimplements double-click-and-drag logic for selecting
by word. The previous implementation was horribly broken. See #741 for
all the details.
The implemented logic now is:
* A double-click initiates a select-by-word selection mechanism.
- A double-click may start on a word or whitespace
- If the initial double-click is on a word, that word is immediately selected.
- If the initial double-click is on whitespace, the whitespace is not selected.
* A "word" is determined by a non-boundary character meeting a boundary character.
- A boundary character is `NUL` ` ` (space) `\t` `'` `"`
- This list is somewhat arbitrary to make the terminal "feel" good.
- Cell SGR states (fg/bg, bold, italic, etc.) have no effect on boundary determination or selection logic.
* As the user drags _on the same line_:
- No selection change occurs until the cursor is over a new word. Whitespace change does nothing.
- When selection is over a new word, that entire word added to the selection.
* When the user drags _up_ one or more lines:
- If the cursor is over whitespace, all lines from the selection point up to but not including the cursor line are selected.
* This selection is done in accordance to the previous rules.
- If the cursor is over a word, the word becomes the beginning of the selection.
- The end of the selection in all cases is the first word at or before the initial double-click point.
* When the user drags _down_ one or more lines:
- The same logic as _up_ but swap the "beginning" and "end" of selection terminology.
* With this logic, the behavior of Ghostty has the following invariants:
- Whitespace is never selected unless it is between two selected words
- Selection implies at least one word is highlighted
- The initial double-click point marks the beginning or end of a selection, never the middle.
This fixes#786 by adding a check to the callback that unhides the menu
bar to only unhide the menu bar if focus was lost to another Ghostty
window.
That's the desired behaviour: when focus is lost to another app's
window, we want the non-native-fullscreen window to stay unchanged in
background, but when changing focus to another Ghostty window, we want
to unhide the menu bar.
Why did this problem even show up?
It started to show up with the introduction of the Xib-file based
approach, in 3018377.
Before that, in 27ddc90, for example, the app would receive the same
notifications, but the `.autoHideMenuBar` didn't have an effect on the
window. Only after adding the Xib-file did it start to have an effect.
So I figured there's two ways we could fix it:
1. Figure out why the `.autoHideMenuBar` now works with Xib-files and
suppress it, or
2. Encode in the code the behaviour that we actually want: we only want
to show the menu bar when focus shifts to another one of Ghostty's
windows, but we want to leave it untouched when focus is lost to
another app's window.
I went with (2) but I think (1) is also valid and happy to close PR if
that's what we want to do.
Normally, we queue all the writes we need from a single `read()` syscall
and only wake up the writer thread at the end of processing that batch
of data.
But under very heavy load or large batches of data, it is possible for the
terminal sequences to generate enough writes to the pty to fill the
writer thread queue while we're still processing the data from `read()`.
This modifies our queuer to attempt to queue, but if the queue is full
we wake up the writer thread immediately then queue again (which should
succeed in every case -- eventually).
This maybe is a robust way to get Monaspace fonts working.
Previously, we used leading as part of the calculation in cell height. I
don't remember why. It appears most popular monospace fonts (Fira Code,
Berkeley Mono, JetBrains Mono, Monaco are the few I tested) have a value
of 0 for leading, so this has no effect. But some fonts like Monaspace
have a non-zero (positive) value, resulting in overly large cell
heights.
The issue is that we simply add leading to the height, without modifying
ascent. Normally this is what you want (normal typesetting) but for
terminals, we're trying to set text centered vertically in equally
spaced grid cells. For this, we want to split the leading between the
top and bottom.
Instead of making a separate enum that must be translated from the
ClipboardRequest type, simply re-use ClipboardRequest to determine the
clipboard confirmation reason.
PR #850 introduced several color palette OSC setting and querying
sequencing. In doing so, an active palette was introduced to enable
resetting back to the default (configured) palette. The active palette
was not initialized with the configured palette, thus any configured
theme by the terminal was not set at launch.
This behavior can be confirmed thanks to PR #852, which resets the
active palette on configuration reload. To observe the behavior:
1. Set the configured palette to something other than the default
2. Open ghostty. Observe the color palette
3. Reload the configuration
4. Press enter for a new shell prompt. Note the palette has changed
This patch sets the configured palette as the active palette at
initialization.
When the config is changed, update the active palette in addition to the
default palette, but only those colors which have not been changed with
OSC 4.
Default colors are those set by the user in the config file, or an
actual default value if unset. The actual colors are modifiable and can
be changed using the OSC 4, 10, and 11 sequences.
