With dpi-aware=auto (the default), scale fonts using DPI *only*
if *all* available monitors have a scaling factor of one.
The idea is this: if a user, with multiple monitors, have enabled
scaling on *at least* one monitor, he/she has most likely done so to
match the size of his/hers other monitors.
For example, if the user has one monitor with a scaling factor of one,
and another one with a scaling factor of two, he/she expects things to
be twice as large on the second monitor.
If we (foot) scale using DPI on the first monitor, and using the
scaling factor on the second monitor, foot will *not* look twice as
big on the second monitor (this was the old behavior of
dpi-aware=auto).
Part of #714
If the call to fdm_wayl() with EPOLLHUP was followed by calls to
wayl_roundtrip(), foot would hang.
The reason for this is that the EPOLLHUP handler in fdm_wayl() would
call wl_display_cancel_read().
wayl_roundtrip() also calls wl_display_cancel_read() (since we
normally have called wl_display_prepare_read()), before calling
wl_display_roundtrip().
When calling wl_display_cancel_read() two times in a row, without a
wl_display_prepare_read() in between, wl_display_roundtrip() hangs.
Fix by not calling wl_display_cancel_read() in fdm_wayl(). This
ensures our invariant holds: wl_display_prepare_read() is *always* in
effect outside of fdm_wayl().
Closes#651
We still use the primary font, but use a custom size, based on the
title bar’s height.
This fixes an issue where the window title could be way too small, or
way too big. And changed size when the terminal font size was changed.
Up until now, *all* buffers have been tracked in a single, global
buffer list. We've used 'cookies' to separate buffers from different
contexts (so that shm_get_buffer() doesn't try to re-use e.g. a
search-box buffer for the main grid).
This patch refactors this, and completely removes the global
list.
Instead of cookies, we now use 'chains'. A chain tracks both the
properties to apply to newly created buffers (scrollable, number of
pixman instances to instantiate etc), as well as the instantiated
buffers themselves.
This means there's strictly speaking not much use for shm_fini()
anymore, since its up to the chain owner to call shm_chain_free(),
which will also purge all buffers.
However, since purging a buffer may be deferred, if the buffer is
owned by the compositor at the time of the call to shm_purge() or
shm_chain_free(), we still keep a global 'deferred' list, on to which
deferred buffers are pushed. shm_fini() iterates this list and
destroys the buffers _even_ if they are still owned by the
compositor. This only happens at program termination, and not when
destroying a terminal instance. I.e. closing a window in a “foot
--server” does *not* trigger this.
Each terminal instatiates a number of chains, and these chains are
destroyed when the terminal instance is destroyed. Note that some
buffers may be put on the deferred list, as mentioned above.
All SHM pixmap cookies depend on the terminal instance’s memory
address. Thus, after a terminal instance has been destroyed, shm
pixmaps that belonged to it will never be purged automatically.
CSDs aren’t typically toggled on and off. Thus, when disabled,
immediately purge their corresponding pixmap buffers, to free up some
memory, and release file descriptors.
Unlike other surface types, the SHM cookie depends on the address of
each URL instance. This means if we enable, disable, and then enable
URL mode again (thus showing exactly the same URLs as the first time),
the URLs will have new addresses, and thus the old SHM pixmaps will
not get purged automatically.
So, manually purge them when destroying the URLs.
* Set win->configure.csd_mode, not win->csd_mode. Otherwise our
‘configure’ handler will swap out the CSD_YES we set, for
CSD_UNKNOWN(?), resulting in no CSDs at all.
* Don’t instantiate the CSDs in wayl_win_init(), let the ‘configure’
event handler do that. Just like it does when we have a decoration
manager emitting decoration configure events.
The configure event asks the client to change its decoration
mode. The configured state should not be applied immediately.
Clients must send an ack_configure in response to this event.
See xdg_surface.configure and xdg_surface.ack_configure for
details.
In particular, ”the configured state should *not* be applied
immediately”.
Instead, treat CSD/SSD changes like all other window dimension related
changes: store the to-be mode in win->configure, and apply it in the
surface configure event.
This fixes an issue where foot incorrectly resized the window when the
server switched between CSD/SSD at run-time.
ayl_roundtrip() has the following code:
wl_display_roundtrip(wayl->display);
while (wl_display_prepare_read(wayl->display) != 0)
wl_display_dispatch_pending(wayl->display);
wayl_flush(wayl);
If the first wl_display_roundtrip() fails, for example because the
Wayland socket has been closed, we may get stuck in the while-loop.
This happens if the read queue isn’t empty, in which case
wl_display_prepare_read() will return -1 and we’ll continue trying to
dispatch the pending events forever, never succeeding since the socket
is gone.
Closes#542
We only needed term->font_scale to be able to detect scaling factor
changes (term->font_scale != term->scale).
But, we already have the old scaling factor in all places where
term_font_dpi_changed() is called, so let’s pass the old scaling
factor as an argument instead.
Only enable XDG activation when compiling against wayland-protocols
1.21. Older versions don’t have this protocol.
When available, define HAVE_XDG_ACTIVATION.
Make all usages of xdg_activation_v1 and xdg_activation_token_v1
conditional.
Defensive programming; output_geometry() etc are typically only called
once for an output instance. But let’s ensure we’re not leaking memory
if it’s called more than once.
This ensures different seat’s don’t step on each others IME pre-edit
state.
It also removes most dependencies on having a valid term pointer for
many IME operations.
We’re still not all the way, since we support disabling IME with a
private mode, which is per terminal, not seat.
Thus, we still require the seat to have keyboard focus on one of our
windows.
Closes#324. But note that *rendering* of multiple seat’s IME pre-edit
strings is still broken.
In many places we have the following pattern:
tll_foreach(list, it)
free(it->item.thing);
tll_free(list);
Since all tll functions are macros, and thus inlined, and since
tll_free in itself expands to a tll_foreach(), the above pattern
expands to more native code than necessary.
This is somewhat smaller:
tll_foreach(list, it) {
free(it->item.thing);
tll_remove(list, it);
}
Up until now, the various key binding modes (“normal”, “search” and
“url”) have used their own struct definitions for their key bindings.
The only reason for this was to have a properly typed “action” (using
the appropriate “action” enum).
This caused lots of duplicated code. This patch refactors this to use
a single struct definition for the “unparsed” key bindings handled by
the configuration, and another single definition for “parsed” bindings
used while handling input.
This allows us to implement configuration parsing, keymap translation
and so on using one set of functions, regardless of key binding mode.
Instead of disabling content centering, delay the TIOCSWINSZ (a.k.a
delay sending the new dimensions to the client) by a small amount
while doing an interactive resize.
Non-interactive resizes are still immediate.
For now, force a resize when the user stops the interactive
resize. This ensures the client application receives the new
dimensions immediately.
It still works without the last, forced, resize, but there typically
be a small delay until the client application receives the final
dimensions.
Closes#301Closes#283
This is needed to apply content centering after an interactive resize,
as otherwise we’d just ignore the last configure event since the only
difference between that event and the next to last event is that the
is-resizing bit has been cleared. I.e. the window size is identical to
what we already have, and our resize code would thus ignore the event.
Store a list of currently pressed buttons, and which surface they
belong to (i.e. which surface that received the press).
Then, in motion events (with a button pressed, aka drag operations),
send the event to the “original” surface (that received the press).
Also send release events to the originating surface.
This means a surface receiving a press will always receive a
corresponding release. And no one will receive a release event without
a corresponding press event.
Motion events with a button pressed will *always* use the *first*
button that was pressed. I.e. if you press a button, start dragging,
and then press another button, we keep generating motion events for
the *first* button.
This is done by allocating cells for the pre-edit text when receiving
the text-input::done() call, and populating them by converting the
utf-8 formatted pre-edit text to wchars.
We also convert the pre-edit cursor position to cell positions (it can
cover multiple cells).
When rendering, we simply render the pre-edit cells on-top off the
regular grid. While doing so, we also mark the underlying, “real”,
cells as dirty, to ensure they are re-rendered when the pre-edit text
is modified or removed.
