This commit does two related things, from opposite ends of the spectrum:
* Sets the sticky bit on pid files and unix sockets to avoid tmpwatch
deleting them in cleanup scenarios
* Falls back to looking at the changed time if the filesystem doesn't
support reporting creation time when wezterm does its own liveness
and cleanup checks for unix domain sockets in the runtime dir
* Allow any wezterm instance to perform that cleanup
refs: #1601
This action causes the active workspace for the gui to change.
If the name is omitted a random name will be generated.
If the workspace doesn't exist, it will be be created.
The optional spawn parameter can be used to launch a specific
program into the new workspace; if omitted, the default prog
will be used.
The gui only supports a single active workspace. Switching workspaces
will repurpose existing gui windows and re-assign them to windows
in the new workspace, adjusting their size to fit those windows,
spawning new windows or closing unused windows as required.
The gui now exits when there are no panes in the active workspace,
rather than no panes at all.
refs: #1531
This got lost somewhere along the way.
Importantly, we need to explicitly set the pty to use invalid stdio
handles for the spawned child in order to avoid a weird situation
where eg: cmd or powershell would end up writing to the log file
despite it being spawned into its own PTY.
refs: #1358
The first problem in #1222 is that the environment in the spawned
mux server has inherited the unix socket path from the GUI that
is the client, rather than the mux server.
That means that `wezterm cli` is operating on the mux namespace
of the client (which is a layered/wrapped view of the mux server),
rather than the mux server.
This fixes `wezterm cli split-pane` in the case where unix domains
are in use together with `connect_automatically = true`, although
there is still a really weird issue with tabs that needs to be
run down.
refs: https://github.com/wez/wezterm/issues/1225 (sort of)
refs: https://github.com/wez/wezterm/issues/1222
This has been a commonly requested feature in the past week,
and it's a reasonable one. The mux server inherited the
close-when-done behavior from when it used to be an alternate
front-end in the same executable as the gui, but it doesn't
need to be that way any more.
We also need to accomodate that case in the client: if the
newly attached domain doesn't result in any panes being imported,
we need to spawn a new command there in order to keep the client
alive. The pre-existing check for whether the mux was empty had
false positives because the local mux may still reference the
pane from the connection UI, which would finish closing out shortly
after we had decided not to spawn anything, and then the client
would close.
refs: https://github.com/wez/wezterm/issues/631
refs: https://github.com/wez/wezterm/issues/507
Rust 1.51 allows addressesing a long-standing TODO which was that we
shouldn't need to build a vendored copy of openssl on most sensible unix
systems.
We do require a vendored copy on macOS and Windows, but due to the way
that Cargo's feature resolver worked, it wasn't possible for this
requirement to be respected.
Rust 1.51 introduces `resolver="2"` which can deal with this feature
resolution!
https://doc.rust-lang.org/nightly/cargo/reference/features.html#feature-resolver-version-2
The upshot of this is that building wezterm on real unix systems that
are not macos will now link against the system libssl, resulting in both
a shorter compile time and less headaches arising from having a slightly
different openssl used by wezterm than the rest of the system.
cc: @jsgf
`wezterm`, `wezterm-gui` and `wezterm-mux-server` now all support
a new `--config name=value` CLI option that can be specified
multiple times to supply config overrides.
Since there isn't a simple, direct way to update arbitrary fields
of a struct in Rust (there's no runtime reflection), we do this
work in lua.
The config file returns a config table. Before that is mapped
to the Rust Config type, we have a new phase that takes each
of the `--config` values and applies it to the config table.
For example, you can think of configuration as working like this
if wezterm is started as `wezterm --config foo="bar"`:
```lua
config = load_config_file();
config.foo = "bar";
return config;
```
The `--config name=value` option is split into `name` and `value`
parts. The name part is literally concatenated with `config` in
the generated lua code, so the name MUST be valid in that context.
The `value` portion is literally inserted verbatim as the rvalue in the
assignment. Not quoting or other processing is done, which means
that you can (and must!) use the same form that you would use in
the config file for the RHS. Strings must be quoted. This allows
you to use more complicated expressions on the right hand side,
such as:
```
wezterm --config 'font=wezterm.font("Fira Code")'
```
The overrides stick for the lifetime of the process; even if
you change the config file and reload, then the value specified
by the override will take precedence.
refs: https://github.com/wez/wezterm/issues/469
refs: https://github.com/wez/wezterm/issues/499
* Add cli option --config-file
* Update cli arg doc & make it conflict with skipping config
* When the config is given explicitly (either using --config-file or via WEZTERM_CONFIG_FILE), failing to load this file will use the default config.
* Otherwise the config file is searched one by one in a few directories.
80214319ae broke the use of RUST_LOG to
turn up trace logging.
This commit refactors logger initialization into the env-bootstrap crate
so that it is centralized, and adopts the use of `WEZTERM_LOG` to
override the default logging filters, rather than `RUST_LOG`.
Revise logging so that we use info level for things that we want
to always log, and adjust the logger config to always log info
level messages.
That means shifting some warning level logs down lower to debug level so
that they aren't noisy.
closes: https://github.com/wez/wezterm/issues/388
One of the updates in the past week made the reactor/async-io
stuff sensitive to the fork performed by daemonizing.
We can "simply" re-exec ourselves post-fork to resolve this.
This builds on the new lua event handler plumbing added
in ccea650a93 to co-opt
the default URI opening action:
```lua
wezterm.on("open-uri", function(uri)
if uri:find("jira") then
wezterm.log_error("do something with jira")
wezterm.run_child_process({
"wezterm",
"start",
"--",
"jira",
"view",
extract_task_from_uri(uri)
})
-- prevent the default action from opening in a browser
return false
else
-- log but allow the uri to be opened in the browser
wezterm.log_error("clicken " .. uri)
end
end)
```
This doesn't allow exactly the sketched out option from
issue #223 to be implemented, but may be close enough
to be useful.
refs: #223
refs: #225
* Taught wezterm-mux-server how to `--daemonize` on windows
* Removed pty based command spawn used to spawn unix domain servers.
This was present because it was the only way to successfully spawn
wsl in the past. What I'm finding today is that it doesn't work
at all for me, generating an `0xc0000142` Application failed to
initialize error. A plain command builder spawn seems to work,
so that's what we're going with.
* Ensure that we put `.exe` on executable name on windows, otherwise
the spawn may fail.
* `Path::exists()` always returns false for unix domain sockets on
Windows, so unconditionally try to remove the socket before binding,
otherwise the bind will falsely fail, claiming that another process
is already bound.
The docs for mux will need to be updated to show how to revise them
for the new mux server invocation:
```lua
unix_domains = {
{
name = "wsl",
serve_command = {"wsl", "wezterm-mux-server", "--daemonize"}
},
}
```
This version no longer generates invalid CA certificates and
allows TLS connections using the internal PKI to advance
further. Still need to debug an early disconnect.
kindof a lot going on in this commit, unintentionally:
* Need the lua context set to be moved into the config crate
otherwise configs cannot be parsed by the server and we end
up with the default configs
* Make the server use smol for async io
* Drop the use of the daemonize crate, which I had forked anyway.
Just inline our own tighter daemonize module
* Improve daemon spawning synchronization, however, it still needs
work for windows to avoid blocking forever where we don't do
daemonizing.
