When getting the hostname to construct the legacy uvar path, if the
hostname is empty, we will create a path pointing at a directory. On
BSDs this path can be successfully open'd and we will produce errors
about invalid uvar files.
To broadcast a uvar change on Linux, we write to a named pipe, wait a bit,
and then read it back. While the pipe is readable, fish will enter a "polling
mode" where it will check for uvar changes every N msec, until the pipe is no
longer readable. If the pipe stays readable for too long (5 seconds), fish
will try to drain it; this may happen if broadcasting instance of fish is
killed before it can read back its data.
In #8209 we have a case where fish is launched in the background to set a
uvar, and then immediately exits, leaving data on the pipe. This means that
we are perpetually in a polling mode until we hit that timeout. Reduce the
timeout to 1 second and the polling interval to 10 msec.
This improves #8209; it doesn't fix it fully but I think it's the best we can
do absent some other IPC mechanism.
In rare cases this may cause the universal variable file to drop
an update, if two happen at the same time and HOME is on an nfs mount.
But this is considered better than hanging if nfs is lockless.
Fixes#7968.
env_universal_t locking discipline is now managed by env.cpp.
That is, the shared instance of env_universal_t is managed by a lock.
We no longer need to have an internal lock, so remove it.
Previously an instance of env_universal_t had to be created with a file
path. Switch to allowing it to be created as empty, and later initialized
with the file path. This will help simplify the case where universal
variables are not used; they may simply be not initialized and so just
appear empty.
In the named pipe notifier, notifications are broadcast by writing to the
pipe, waiting briefly, and then reading it back. When clients see the pipe
as readable, they report the uvars as potentially changed and fish will
sync against the uvar file.
Prior to this change, we synced repeatedly when the pipe was readable. But
we can do somewhat better by also checking the named pipe's timestamp (via
fstat). If the pipe has not changed, then we can skip the sync even if
there is currently data lingering on it.
With this change we should sync against the variable file less often
(typically once or twice per write); in the next change we refactor this
logic so it's easier to follow.
select_wrapper_t wraps up the annoying bits of using select(): keeping
track of the max fd, passing null for boring parameters, and
constructing the timeout. Introduce a wrapper struct for this and
replace the existing uses of select() with the wrapper.
If fish launches a program and that program marks stdin as O_ASYNC, then
fish will start receiving SIGIO events on Mac. This occurs even though
the file descriptor itself does not have the O_ASYNC flag set.
SIGIO is reported as interrupting select which then breaks multiple-key
bindings, especially in vi-mode.
As the SIGIO based universal notifier is disabled, remove it and the
SIGIO handler itself. This allows fish to ignore properly ignore SIGIO.
Fixes#7853
It doesn't work on WSL, Solaris and Archlinux (and presumably that
means future versions of other linux distros).
In its current state I don't trust it enough to enable it anywhere by
default, especially since I'm not aware of an actual issue with the
named pipe (other than that the code is ugly).
Fixes#7774
My bet is that the Illumos, Cygwin, and WSL are not the only Unix-like
systems where the SIGIO notifier won't work, and since we have a good
enough and portable alternative that we can use be default on other
platforms where we don't specifically know it'll work, it doesn't make
sense not to go with that one instead.
Even if this patch is reverted at some point and we go back to
blacklisting platforms that *don't* support the SIGIO strategy, this is
almost certainly the right choice for inclusion in a minor release like
3.2.0.
See discussion in #6585.
Those platforms should not be using the sigio notifier in the first
place, this just stops them from failing to be able to compile fish
altogether.
See #6585
This fixes up the SIGIO notifier in preparation for using it on BSD. It
removes the reliance on the signal's si_code, which is not available in
BSD, and it properly handles the BSD behavior where SIGIO is delivered on
a read even if the read returns EAGAIN.
On WSL1, fcntl(F_SETOWN) will fail and this would report an error.
Suppress this error message since it is not very interesting.
The effect is to disable real-time universal variable propagation.
Introduce a new strategy for notifying other fish processes of universal
variable changes, as a planned replacement for the complex
strategy_named_pipe. The new strategy still uses a named pipe, but instead
of select() on it, it arranges for SIGIO to be delivered when data is
available. If a SIGIO has been seen since the last check, it means the file
needs to be re-read.
Profiling revealed string comparison in variable lookups to be a
significant hotspot. This change causes `make test` to complete ~4.5%
faster per `hyperfine`.
