Shader animation events are accumulated per OS window and only consumed
when the window is next rendered. A focus-out followed by a focus-in with
no rendered frame in between, as happens when the window is hidden and
shown again, or a brief focus steal within one repaint interval, left both
bits pending. The state machine then treated them as simultaneous, so the
focus-out hard stop prevented os-window-focus-in groups from starting and
groups started by focus-out were turned on while focused.
Now a focus change replaces any pending opposite focus change, so only the
final focus state is reported. Genuinely simultaneous events, such as a
click that moves focus away, still give focus-out priority.
Closes#10871
The VS15 branch of draw_combining_char unconditionally moved the cursor
back by the width removed from the char. When the char was on the
previous line (cursor at x=0 after a soft wrap) this wrapped cursor->x to
UINT_MAX leading to out of bounds reads and writes. Only move the cursor
when it is on the same line as the char. Also preserve the soft wrap
marker in the cells cleared by halve_multicell_width.
File names, archive member names, symlink targets, error messages and
ebook metadata were written to the terminal unsanitized, allowing
crafted names to inject escape codes (e.g. OSC 52 to overwrite the
clipboard) merely by highlighting the entry.
A program running in the terminal could use edit-in-kitty to set --logo=/dev/zero
which would cause the image converter to read an infinite stream, blocking the
kitty main loop and consuming unbounded memory.
Docks can now be resized in whole cells by dragging their inner border,
with the resize_window action and with kitten @ resize-window, where
wider or taller always means bigger whatever the edge of the dock. A
dock keeps at least one cell and grows only into the content it is
docked to, so the docks inside it are never shrunk. OS window docks do
not shrink the tab docks of any tab, and when dragged, the other tabs
are re-laid out once, at the end of the drag. Percentage sizes stay
percentages. The new size is stored in the DockSpec, so it is reported
by kitten @ ls and saved in sessions. Resetting window sizes does not
change docks, as they are not part of the layout.
Previously the inner border of a dock showed a resize pointer but
dragging it did nothing, or in the Tall layout moved a layout divider.
Fixes#10857
Do not track decoration touches that start an interactive move or resize,
the compositor takes the contact over and need not send its up event, which
left the contact active and caused the next decoration touch to be ignored.
Drop a stale tracked contact whenever a new down arrives with the same id,
wherever it lands. A titlebar button held down by one finger no longer
blocks moving or resizing windows with another.
Restore the old mouse behavior of hit testing titlebar buttons only by x, so
that a button is still activated when released after dragging the pointer
below the titlebar.
Update the input serials only for decoration touches that are handled,
including their up events. Share the titlebar commit code between the
pointer and touch paths, remove checks that can no longer fail, fix
indentation in decoration_surface() and add a changelog entry.
OSC 9;9;path is now treated as a report of the current working directory,
like OSC 7, instead of popping up a desktop notification. Other ConEmu
numeric OSC 9 sub-commands (1-12) are ignored rather than shown as
notifications.
Scrolling down (CSI T) could cut off the lower rows of a scale > 1 text
sizing character at the bottom of the screen. Adding a combining char to
it then wrote past the end of the line buffer in add_combining_char and
rewrapping on resize read past the end of the line buffer in
copy_multiline_extra_lines.
Multiline chars that are cut off at the edge of the screen or of the
history buffer, by scrolling or by inserting lines, are partially
scrolled off and are rendered as such, so they are kept. Code walking the rows of a multiline char now only visits
the adjacent rows that actually continue the char, rather than trusting
the y and scale of a cell, so it stays in bounds and does not modify
unrelated cells for partial or torn chars.
Chars that would be torn, i.e. left with rows from unrelated lines
adjacent to their remaining part, are removed. This happens when
scrolling with margins, when a partial char is moved away from the edge
of the screen by scrolling or deleting lines, when inserting lines pushes
the lower rows of a char straddling the bottom margin out of the scroll
region, and when a partial char
cut off at the bottom is no longer at the bottom after a resize. Rewrap
drops partial chars, as it already did for chars with missing top rows,
placing cursors that were on them where the char would have been.
The checks for torn chars are skipped unless a multiline char has been
drawn in the screen, and chars cut off at the bottom of the screen,
which can only be created by scrolling down or inserting lines, are
tracked so that
linefeeds do not need to scan for them.
A hard reset now clears the alternate screen as well, so that the flag
tracking whether multiline chars have been used can be reset.
Also fix the lower rows of multiline chars being lost on rewrap when the
lines after the char were continuations that fit on its first row.
Create the temporary directory used for staging drops lazily, on the first
drop, instead of at startup, so the kitten works in directories that are not
writable when nothing is dropped into them.
Fixes#10855
An OS window dock is shown in every tab of its OS window, below the tab
bar and outside the docks of the tab, and lives as long as the OS window.
It is a single window running a single program, which is kept in the
active tab: it is moved to the new active tab on every tab switch, by a
new C function that moves the window between tabs of an OS window
without recreating its GPU resources. Every tab reserves the space for
the OS window docks, so switching tabs does not resize any windows.
The docks are not closed with tabs and do not keep a tab open, stay in
their OS window when the tab containing them is moved to another OS
window, and are closed when the last other window in the OS window is
closed. They are listed as part of the active tab, and the dock entry of
every docked group in the output of kitten @ ls now has a scope of
window, tab or os-window. They are saved in sessions, once per OS window.
Fixes#10854
Every on_resize and on_focus_change took its own 0s timer, so creating many
windows from a single callback filled the fixed size glfw timer table and
add_timer() returned 0, silently dropping the watcher calls. Now watcher
calls are queued and delivered in order behind a single pending timer, and
only when the window actually has a watcher for the event. Calls made during
delivery go to the next batch, preserving the recursion guard the timer
exists for.
Also fix add_timer() leaking a reference to the callback when the timer
could not be added.
Fixes#10853
A tab moved to another OS Window keeps its session, and new tabs and OS
Windows created for moved tabs and windows belong to the session of what
was moved. A window keeps its own session, which may be none, wherever it
is moved, so --add-to-session=! is respected; a window of no session
belongs to the session of the tab it is in.
Add Window.effective_session_name for the "window session, else its tab's
session" rule that was duplicated in five places, and use it in launch as
well, so --add-to-session=. and --cwd=current agree with
new_window_with_cwd. OS Windows created by launch --type=os-window or for
moved tabs and windows now also record their session.
Closes#10851
Report an error before writing if a filtered session serializes to
nothing. This keeps an existing session file intact and avoids
creating an empty one at a new path.
Let preview renderers supply extra cache key data instead of checking for
the cmd_renderer type. Include the command line in the key for custom
previews and the video_preview settings in the key for video thumbnails,
so that changing them no longer shows stale previews. GetPath and AddPath
now take this extra key data, replacing GetPathWithAlias.
Track the entry each path uses per extra key, so that previews of one file
with different settings do not evict each other. Do not remove an entry
while another path, such as another symlink to the same file, still uses
it. Cache hits are now recorded so that such shared entries are known.
Failing to remove an orphaned entry on a cache hit no longer turns the hit
into an error.
Fix ensure_entries() discarding unwritten changes when the entries file
does not exist yet, which lost the first path recorded in a new cache.
Also reset the path map when clearing the cache.
Custom preview commands receive the selected path. Include that path
in their cache key so that symlinks to the same file can have different
previews. Keep the resolved target in the key so retargeting a link
still invalidates its entry.
While rendering is paused, output can change the live scroll position,
pixel scroll offset, scrollback size and active screen. Image layers,
cursor visibility and the scrollbar used those live values while the
text was drawn from the frozen snapshot, so they drifted from it.
Capture the scrollback size and active screen at pause time and add
rendered_scrolled_by, rendered_history_count and rendered_on_main_screen
macros alongside rendered_pixel_scroll_offset_y. Move image layer
updates and scrollbar visibility and geometry into screen.c so the
renderer and the tests share the same code.
Closes#10847
When the first lines of the live prompt were in the scrollback, an unsigned
wrap made erase_last_command delete most of the screen and the newest
scrollback lines, including the live prompt. Delete the erased lines from
the middle of the scrollback instead, with a new historybuf_delete_lines()
that removes multicell chars cut by the deletion, and also remove such
fragments from the screen. Clear selections and the wrap flag of the line
above the erased lines on every path.
Also keep the last visited prompt on the same content when lines are
inserted or deleted (CSI L/M), forget it on a hard reset, and only fill from
the scrollback when unscrolling (CSI + T) if the scroll region starts at the
top of the screen, instead of overwriting the top line of the screen.
Store the last visited prompt as one row relative to the top of the live
screen (negative rows are in the scrollback) plus the viewport row it was
visited at, and derive the scroll offset when it is used. Previously it was
stored as a (viewport y, scrolled_by) pair that every code path moving lines
had to keep consistent by hand, which broke after erasing a command left
less scrollback than the saved offset, and when scrolling regions with
margins or reverse indexing.
Choose the tab to focus using only the windows that were actually saved,
moving the choice into a pure function so it can be tested without
constructing a TabManager. Test the first OS window logic with simple fake
tab managers instead of patching TabManager internals. Merge the changelog
entries and mention restore after logout.
TabManager.serialize_state_as_session wrote focus_tab as the index of the
active tab among all the tabs of the OS window, but loading a session
counts only the tabs it creates. With --match the tabs of other sessions
are not saved, so a session with such tabs before its active tab came back
on the wrong tab, or on its last tab once the index was clamped.
Write the index among the saved tabs. When the active tab is not saved, as
when it belongs to another session, write the saved tab with the most
recently focused saved window. That is the tab kitty focuses when switching
to the session while it is open, as window_for_session_name picks the
session's most recently focused window, so the session comes back as it was
last seen. Leaving focus_tab out instead would focus whichever saved tab
has the last focus command, usually the last one, which is arbitrary.
Boss.serialize_state_as_session passed is_first to the first OS window it
serialized, which is any but the focused one when there are several. With
--match that OS window can have no matching windows and write nothing, and
the next one then started the file with new_os_window, so loading the
session opened it in a new OS window instead of the current one. Whether
it did depended on which OS window had focus when saving.
Pass is_first to the first OS window that writes anything.
Always reply from the calibre worker, even when reading metadata raises or
calibre emits stray output, so that the request does not block until it
times out.
Accept cached calibre metadata generated for a different path to the same
file, as the disk cache key resolves symlinks. Otherwise a symlink and its
target keep invalidating each other's cache entry.
The mouse-demo kitten popped and pushed the pointer shape on every mouse
event, and kitty applied each OSC 22 to the OS pointer immediately, so the
pointer briefly flipped to the default shape and back, restarting
animations. Only change the shape in the kitten when it actually changes,
and apply pointer shape stack changes to the OS pointer once per batch of
parsed child output.
Also have ApplyHoverStyles replace the current shape instead of pushing a
new one, so the shape stack does not grow as the mouse moves between
regions.
Fixes#10838
ColorProfile.as_dict() used the key cursor_text instead of the kitty.conf
name cursor_text_color, so get-colors listed the window override under a
different name than the configured value, which then shadowed it.
The bootstrap script extracted the tarball with umask 000, so tar
created any directories not present in the tarball, such as ~/.terminfo
and the ssh kitten data directory, with mode 0777. Since tar is run with
-p, the permissions of entries in the tarball are already restored
exactly, so the umask override is unnecessary. Remove it so that
directories not in the tarball respect the user's umask, matching the
behavior of bootstrap.py.
Fixes#10834Closes#10835
Replace the grab/release error handler pattern with a process wide
handler that is installed for the lifetime of the display. Errors are
captured by nestable traps identified by request serial numbers, so
traps only capture errors from requests made while they are open, and
errors from requests not covered by any trap are reported rather than
exiting the process.
Requests to windows owned by other clients (selection requestors, drag
and drop peers) are trapped individually, without waiting for a round
trip to the server, and without covering calls into the application.
Also fix reading uninitialized memory when handling a MULTIPLE
selection request whose requestor window has been destroyed, trap
errors when replying to drag sources, report failure to deliver drag
data, and fix detectEWMH() leaving the error handler installed on an
early return.
Follow up to the previous commit fixing hangs when previewing named pipes:
- Skip previewing anything that is not a regular file (or a symlink to
one) in preview_for, before any previewer opens it. This avoids
unblocking and then breaking writers waiting on a FIFO and covers the
non-text previewers (images, video, archives, calibre, commands) too.
- Apply the text preview size limit to the symlink target instead of the
symlink itself, and enforce it while reading so that files growing
after stat() are also handled.
- Pre-allocate the read buffer based on the file size, as os.ReadFile does.
- Read ignore files during scanning with the same non-blocking reader, so
a FIFO named .ignore or .gitignore cannot stall the scan.
- Unblock goroutines stuck on FIFOs when tests time out.
Open text preview files in nonblocking mode and check the type on
the opened descriptor before reading. Reuse the same read for syntax
highlighting because it can reopen a path after the file changes.
bash is compiled with the paths to its system wide profile and bashrc
(SYS_PROFILE and SYS_BASHRC) and there is no way to query bash for them.
Previously kitty guessed locations in /etc, which fails on distros that
keep these files elsewhere, such as hermetic /usr distros like Clear Linux
(/usr/share/defaults/etc) or Nix (/etc/bashrc).
Now find the paths by scanning the bash executable for them, both in kitty
and in kitten run-shell (so it works on remote hosts via the ssh kitten).
The result is cached per executable in kitty. If the scan fails, or the
locations are explicitly specified via the environment, the previous
behavior is used.
Fixes#10819
Use Capture.ByteRange() from v2 instead of a hand-rolled rune to byte
offset mapping. The old mapping required capture offsets to be monotonic
across matches, so a capture inside a lookaround overlapping a later match
in non-ASCII text caused all matches to be silently dropped.
Also propagate match errors instead of discarding them. Drops the v1
dependency, v2 was already pulled in by chroma.
Shake only the window in which a bell occurred, using
d.bell_window_geometry, either side to side or in random directions.
Useful for shaking the window when a shell command fails. Add docs with
shell snippets to ring the bell on command failure and demo entries.
Fire the event directly from screen_bell() instead of piggy-backing on
start_visual_bell_at in the render loop. Previously the event only fired
when visual_bell_duration was non-zero and bells in windows that were not
visible fired it late, when the window was next rendered. Now only bells
in visible windows (in the active tab and not hidden by an overlay or the
layout) fire the event.