The global config doesn’t necessarily reflect the correct
configuration to use - we should *always* use the current terminal
instance’s conf pointer.
* Move selection override modifier mask to the key_binding_set struct
* Always warn if XDG activation is unavailable, not just if
bell.urgent is set (we no longer have access to this information)
* Pass ‘bool presentation_timings’ as a parameter to wayl_init()
* Remove ‘presentation_timings’ member from the ‘terminal’ struct
Closes#932
Up until now, our Wayland seats have been tracking key bindings. This
makes sense, since the seat’s keymap determines how the key bindings
are resolved.
However, tying bindings to the seat/keymap alone isn’t enough, since
we also depend on the current configuration (i.e. user settings) when
resolving a key binding.
This means configurations that doesn’t match the wayland object’s
configuration, currently don’t resolve key bindings correctly. This
applies to footclients where the user has overridden key bindings on
the command line (e.g. --override key-bindings.foo=bar).
Thus, to correctly resolve key bindings, each set of key bindings must
be tied *both* to a seat/keymap, *and* a configuration.
This patch introduces a key-binding manager, with an API to
add/remove/lookup, and load/unload keymaps from sets of key bindings.
In the API, sets are tied to a seat and terminal instance, since this
makes the most sense (we need to instantiate, or incref a set whenever
a new terminal instance is created). Internally, the set is tied to a
seat and the terminal’s configuration.
Sets are *added* when a new seat is added, and when a new terminal
instance is created. Since there can only be one instance of each
seat, sets are always removed when a seat is removed.
Terminals on the other hand can re-use the same configuration (and
typically do). Thus, sets ref-count the configuration. In other words,
when instantiating a new terminal, we may not have to instantiate a
new set of key bindings, but can often be incref:ed instead.
Whenever the keymap changes on a seat, all key bindings sets
associated with that seat reloads (re-resolves) their key bindings.
Closes#931
When this option is used, the child process in the new terminal
instance will inherit its environment from the footclient process,
instead of the foot server’s.
Implemented by sending (yet another) dynamic string list as part of
the client -> server setup packet. When the new option is *not* used,
the setup packet is now 2 bytes larger than before.
On the server side, the slave process now uses execvpe() instead of
execvp(). There’s plumbing to propagate a new ‘envp’ argument from
term_init() all the way down to slave_exec(). If ‘envp’ is NULL, we
use ‘environ’ instead (thus matching the old behavior of execvp()).
Closes#1004
A foot --server instance would exit with code 0, even on failure, if
the number of currently open terminal instances were 0.
This is because ‘ret’ assumed failure, and then tried to set it to
‘success’ after the even loop had terminated, basted on the server’s
current state.
Fix by:
* set ‘ret’ to success just before entering the event loop
* set ‘ret’ to failure when we detect an FDM failure
* don’t try to second-guess success/failure after having exited the
event loop
Closes#943
Not doing so before calling sigaction(3) is "undefined" according to
POSIX[1]:
> Applications shall call either sigemptyset() or sigfillset() at least
> once for each object of type sigset_t prior to any other use of that
> object. If such an object is not initialized in this way, but is
> nonetheless supplied as an argument to any of pthread_sigmask(),
> sigaction(), sigaddset(), sigdelset(), sigismember(), sigpending(),
> sigprocmask(), sigsuspend(), sigtimedwait(), sigwait(), or
> sigwaitinfo(), the results are undefined.
The use of designated initializers means that sa_mask members were
still being initialized, but sigset_t is an opaque type and implicit
initialization doesn't necessarily produce the same results as using
sigemptyset(3) (although it typically does on most implementations).
[1]: https://pubs.opengroup.org/onlinepubs/9699919799/functions/sigaddset.html
Fcft no longer uses wchar_t, but plain uint32_t to represent
codepoints.
Since we do a fair amount of string operations in foot, it still makes
sense to use something that actually _is_ a string (or character),
rather than an array of uint32_t.
For this reason, we switch out all wchar_t usage in foot to
char32_t. We also verify, at compile-time, that char32_t used
UTF-32 (which is what fcft expects).
Unfortunately, there are no string functions for char32_t. To avoid
having to re-implement all wcs*() functions, we add a small wrapper
layer of c32*() functions.
These wrapper functions take char32_t arguments, but then simply call
the corresponding wcs*() function.
For this to work, wcs*() must _also_ be UTF-32 compatible. We can
check for the presence of the __STDC_ISO_10646__ macro. If set,
wchar_t is at least 4 bytes and its internal representation is UTF-32.
FreeBSD does *not* define this macro, because its internal wchar_t
representation depends on the current locale. It _does_ use UTF-32
_if_ the current locale is UTF-8.
Since foot enforces UTF-8, we simply need to check if __FreeBSD__ is
defined.
Other fcft API changes:
* fcft_glyph_rasterize() -> fcft_codepoint_rasterize()
* font.space_advance has been removed
* ‘tags’ have been removed from fcft_grapheme_rasterize()
* ‘fcft_log_init()’ removed
* ‘fcft_init()’ and ‘fcft_fini()’ must be explicitly called
This ensures processes spawned by us (e.g. the shell, new terminal
instances etc) don’t inherit a flawed signal mask, or having signals
unknowingly ignored.
Closes#854
Foot does not support running under non-UTF8 locales. If we detect a
non-UTF8 locale, we log this and exit with an error.
However, it appears a fairly common situation is this:
user (knowingly or unknowingly) only configures his/hers locale in
e.g. the shell RC files. These aren’t sourced when the compositor is
started via a display manager.
Thus, compositor key binds will fail to launch, and the user typically
has no way of seeing foot’s output.
So, the user proceeds to start another terminal, and from that one
tries launching foot. And it works... (because the shell in the other
terminal sourced the locale configuration).
User is left confused and often don’t know how to debug.
This patch is somewhat hackish; in addition to logging the locale
error, it also pushes a user notification. For this to be visible, we
need to actually start a terminal window.
So, we ignore the configured shell, and we ignore any custom command
line, and instead spawns “/bin/sh -c ‘’”. This allows us to get
something running (that hopefully doesn’t produce too much output we
can’t decode due to the non-UTF8 locale). But, it will exit
immediately. So we also set the --hold flag.
This works, and may be a good enough “solution”. However, it only
works for standalone foot instances, not footclients.
Our getopt_long() based argument parser would (correctly) break out of
the while loop immediately. However, we would then proceed and
decrement ‘argc’ by 1, resulting in it being -1.
This was then passed to the terminal constructor, which passed it on
to the client application fork+exec logic. That finally resulted in an
exec(3) failure:
err: slave.c:339: SHELL=/bin/zsh: failed to execute: No such file or directory
err: fdm.c:215: no such FD: 7
info: main.c:604: goodbye
We want to handle SIGPIPEs without crashing... One way to trigger this
was to use e.g.
pipe-visible=[cat /foo/bar] Control+Shift+q
That is, pipe output to something that did not consume it. This led
to a SIGPIPE when we tried to write the terminal contents to the pipe,
and crashed the whole foot instance.
There's no need to call setlocale() twice, since the first call
returns the same value as the second.
We also handle a NULL return value explicitly and show a distinct
error message for that case, since it typically indicates a
misconfigured locale and not just a non-UTF8 locale.
This is an application of the xdg activation protocol that will allow
compositors to associate new foot toplevels with the command that
launched them.
footclient receives an activation token from the launcher which the
compositor can use to track application startup. It passes the token
to the foot server, which then activates the new window with the token
to complete the startup sequence.
FOOT_DEFAULT_TERM is a string literal passed as a -D argument to
the compiler, so it can just be concatenated with the other string
literals, instead of being formatted with printf().
Load a couple of ASCII glyphs and check if their advance widths
matches. If not, warn the user that the font is probably not
monospaced.
This can be disabled by setting tweak.font-monospace-warn=no.
Closes#704.
As of 2021-07-31, ncurses ships its own version of foot’s terminfo.
Since:
* It doesn’t have the non-standard Sync,Tc,setrgbf,setrgbb
capabilities.
* It doesn’t set hs,fsl,dsl (statusbar).
* We want to be able to update our termminfo without waiting for an
ncurses release.
* Foot should be installable and usable on Linux systems that doesn’t
have the latest ncurses.
we still want to ship our own version. We can however not install it
to the default terminfo location (e.g. /usr/share/terminfo), since it
will collide with the ncurses provided files.
Our options are to either rename our terminfo to something else, or to
keep the name, but install our terminfo files somewhere else.
The first option would be the easy one. However, I think it makes
sense to use the same name. For example, a user that SSH’s into a
remote system that does *not* have our own version installed,
but *does* have the ncurses one, will gracefully fall back to that
one, which is better than manually having to set
e.g. TERM=xterm-256color.
Now, if we want to use the same name, we need to install it somewhere
else. But where? And how do we ensure our version is preferred over
the ncurses one?
I opted to $datadir/foot/terminfo (e.g. /usr/share/foot/terminfo) by
default. It makes it namespaced to foot (i.e. we’re not introducing a
new “standard” terminfo location), thus guaranteeing it wont collide
with ncurses.
To enable applications to find it, we export TERMINFO_DIRS. This is a
list of *additional* directories to search for terminfo files. If it’s
already defined, we *append* to it.
The nice thing with this is, if there’s no terminfo in that
location (e.g. when you SSH into a remote), the default terminfo
location is *also* searched. But only *after* having searched through
TERMINFO_DIRS.
In short: our version is preferred, but the ncurses one (or an older
version of our terminfo package!) will be used if ours cannot be
found.
To enable packagers full control over the new behavior, the existing
meson command line options have been modified, and a new option added:
-Dterminfo=disabled|enabled|auto: *build* and *install* the terminfo
files.
-Dcustom-terminfo-install-location=<path>: *where* the terminfo files
are expected to be found.
This *needs* to be set *even* if -Dterminfo=disabled. For example, if
the packaging script builds and packages the terminfo files separate
from the regular foot build. The path is *relative to $prefix*, and
defaults to $datadir/foot/terminfo.
This is the value that will be appended to TERMINFO_DIRS. Note that
you can set it to ‘no’, in which case foot will *not* set/modify
TERMINFO_DIRS. Only do this if you don’t intend to package foot’s
terminfo files at all (i.e. you plan on using the ncurses ones only).
-Ddefault-terminfo=foot. Allows overriding the default TERM
value. This should only be changed if the target platform doesn’t
support terminfo files.
Closes#671
This means that logging will be completely disabled until log_init()
has been called, which is useful to prevent log spam when running
UNITTEST{} blocks in debug builds.
Note that this doesn't change the default log level at runtime, which
was already being set to LOG_CLASS_INFO in main.c and client.c.
The new log level is also exposed to the command-line interface as
`--log-level=none`, which allows disabling logging entirely.
tllists are great when dealing with dynamically changing lists. They
are also very easy to use when building lists/arrays where the final
size is unknown.
However, this ease of use comes at a price: code size. tll-macros
expand to a lot of code.
Since things in the config are static, once the config has been
loaded, using tllists for configuration data structures doesn’t make
much sense.
This patch replaces nearly all tllists used by the configuration, with
dynamically allocated arrays.
Allow any configuration option to be overridden with -o/--override
'section.key=value' arguments, as suggested in #554
update completitions for override
slight refactoring to ease footclient support
* foot exits with -26/230
* footclient exits with -36/220
This is to give each application a range of exit codes. Currently
unused, but we may want to change this in the future.
Normally, foot and footclient uses the exit code from the client
application (i.e. the shell).
However, foot (or footclient) itself may fail to run; if run outside
of a Wayland session, or no fonts are installed, or the client
application/shell cannot be found (“foot lsdjfldsjf”) etc.
Up until now, there has been no way to differentiate these kind of
failures from the client application exiting with code 1.
This patch changes foot’s failure exit code to -27/229, and
footclient’s to -28/228. Note that footclient will exit with foot’s
-27/229 if footclient ran successfully, but the foot server failed to
instantiate a new window.
Closes#466.
Fixes error in Clang builds:
../../foot/main.c:376:19: error: implicit conversion from enumeration type 'enum log_colorize' to different enumeration type 'enum fcft_log_colorize' [-Werror,-Wenum-conversion]
fcft_log_init(log_colorize, as_server && log_syslog, log_level + 1);
~~~~~~~~~~~~~ ^~~~~~~~~~~~
Initially, these options *did not* have short options. Then, in
e813883367, the short options were added
to footclient’s getopt_long() call.
This was in a sense incorrect. But instead of reverting it, the short
options were made official in
8eaa195990, by adding the short options
to foot, documenting them in the man pages, and adding them to the
shell completions.
Though the commit message of 8eaa195990
says the options have now been included in usage(), they were in
fact *not* added to usage.
This patch does just that.
