* zellij: Move "populate_data_dir" to utils
and rewrite it to take a second, optional parameter. This allows
controlling whether only a specific asset should be installed. We do
this as preparation for being able to recover from a plugin version
mismatch error, where we will need to repopulate the data dir for
offending plugins.
* server/wasm_vm: Recover from PluginVersionMismatch
Adds a global plugin cache that stores, per plugin config, the wasmer
module associated with it. Make `start_plugin` take the pre-populated
module and create only the necessary modifications to the wasm env etc.
* utils: Fix formatting
* zellij: Delete non-existent module
* utils/shared: fix missing "set_permissions"
function when not on unix systems.
* server/wasm_vm: Don't populate cachedir
with serialized versions of the WASM plugins.
* utils/input/plugins: load wasm bytes from assets
for builtin plugin specifications. This foregoes any need to:
- Dump the plugin bytes to disk at all and
- subsequently read the plugin bytes from disk
* zellij: Disable default asset installation
which previously installed only the builtin plugins to disk. This is no
longer necessary because now we can load the builtin plugins directly
from the application binary.
* utils/input/plugins: Update docs
* utils/input/plugins: Add 'is_builtin' method
to `PluginConfig` that returns true if the plugin configuration refers
to a builtin plugin.
* wasm_vm: Remove plugin version mismatch handling
because a version mismatch in an internal plugin is now unfixable, with
the plugins being loaded from the running binary, and we have no control
over external plugins in the first place.
* cargo: Reintroduce feature flag
for `disable_automatic_asset_installation`
* utils/consts: Add `ASSET_MAP`
which currently contains the compiled WASM plugins.
* utils/shared: Fix clippy lint
* utils/errors: Add more `ZellijError` variants
* zellij: Make loading internal plugins optional
by reenabling the `disable_automatic_asset_installation` flag and
utilizing it for this purpose. Changes plugin search behavior to throw
better errors in case the builtin plugins cannot be found, depending on
the state of this feature.
* utils/errors: Apply rustfmt
* utils/setup: Allow dumping builtin plugins
to a specified folder on disk. This is meant to be an "escape hatch" for
users that have accidentally deleted the builtin plugins from disk (in
cases where the plugins aren't loaded from inside the zellij binary).
* utils/input/plugins: Update docs
* utils/setup: Add hint to `setup --check` output
when zellij was built without the `disable_automatic_asset_installation`
flag and will thus not read builtin plugins from the "PLUGIN DIR".
* utils/setup: Refine `setup --dump-plugins`
to dump to:
- The default "DATA DIR" when no option is provided with the argument,
or
- The provided option, if existent
Also print a message to stdout with the destination folder that the
plugins are dumped to.
* server/wasm_vm: Ignore "NotFound" errors
when attempting to delete the non-existent plugin data directories. This
silences an error message that otherwise ends up in the logs when
quitting zellij.
* utils/errors: Extend "BuiltinPluginMissing" msg
to hint the user to the `zellij setup --dump-plugins` command to fix
their issues for them!
* utils/errors: Track caller in calls to `non_fatal`
which will hopefully, once closures can be annotated, allow us to
display the location of the call to `non_fatal` in log messages.
* utils/input/plugins: Fix plugin lookup
to prefer internal assets if available. It was previously broken because
sorting the paths vector before deduping it would bring the paths into a
wrong order, looking up in the plugin folder first.
Also print a log message when a plugin is being loaded from the internal
assets but exists on disk, too.
* Apply rustfmt
* make: build-e2e depends on wasm-opt-plugins
so it updates the assets when building the binary
* server/qwasm_vm: Remove var
* utils/consts: Add plugins from target folder
and include them in the asset map from there, too. Include plugins from
debug or release builds, depending on the build type.
* utils/consts: Take release plugins from assets
instead of the target/release folder. The latter will break
installations from crates.io, because we currently rely on including the
plugins we pre-compiled and distribute along with the binary.
* server/wasm_vm: Reintroduce .cache folder
to speedup subsequent application launches.
* cargo: Reorder workspace members
to improve behavior with `cargo make` with respect to compilation order.
* Makefile: restructure plugin tasks
* Makefile: Fix CI errors
* Makefile: More CI diagnosis
* github: Install wasm-opt in e2e test workflow
* Makefile: Build plugins for e2e-test target
* server/Wasm_vm: Reorder plugin folder creation
so no folders are created in the plugin cache when loading a plugin
fails due to not being present or similar.
* update plugins testcommit
* makefile: Change job order
* changelog: Add PR #1924
This commit fixes#1722.
In some environment, the SHELL environemnt variable is not exists. It
causes a panic that is reported as #1722. Use generic shell (/bin/sh) to
prevent the panic.
This is a same behavior as tmux.
https://github.com/tmux/tmux/blob/master/spawn.c#L329-L331
* utils/errors: Fix function order in `to_anyhow`
impl for `SendError`. Previously we attached the context to `anyhow!`,
which is wrong (because it doesn't create an `Err` type itself) and
leads to strange behavior where the error seemingly is immediately
panicked upon.
Instead, Wrap `anyhow!` into an `Err()` and then attach the context to
that. This achieves the intended goal and doesn't lead to premature
termination.
* server/terminal_bytes: Ignore error in `listen`
which occurs when quitting zellij with the `Ctrl+q` keybinding. At the
end of the `listen` function we break out of a loop and send a final
`Render` instruction to the Screen. However, when quitting zellij as
mentioned above, the Screen thread is likely dead already and hence we
cannot send it any Instructions. This causes an error in the async tasks
of the panes that handle reading the PTY input.
If we leave the error unhandled, we will have error messages in the log
whenever we quit zellij, even though the application exited normally.
Hence, we now send the final `Render` instruction but do not care
whether it is sent successfully or not.
This is a "workaround" for the fact that we cannot tell whether the
application is quitting or not.
* server/terminal_bytes: Add FIXME note
* changelog: Add PR #1918
don't log errors from async pane threads when quitting zellij
* utils/input/plugins: Return `Result` when loading
a plugin and failing to find it. Since we look up the plugin in multiple
locations, make sure we preserve each of these lookup failures and
report them to the user. This way the user can see what locations were
actually checked.
In addition, before performing the lookup, deduplicate the array of
locations to check. This prevents errors where we look up the same
plugin multiple times in the exact same locations, which can leave a bad
impression on anyone reading it who isn't familiar with the code.
* server/wasm_vm: Remove obsolete context
which was previouly required because the function it is attached to
returned only an `Option` instead of a `Result`.
* server/wasm_vm: Replace `log::error!`
with better error logging by means of `non_fatal`. This preserves the
original error and allows adding context information on top. Also makes
error formatting more uniform across the application.
* server/tab: Replace `log::error!`
with better error logging by means of `non_fatal`. This preserves the
original error and allows adding context information on top. Also makes
error formatting more uniform across the application.
* server/route: Replace `log::error!`
and propagate the error to the caller instead.
* server/pty: Replace `log::error!`
with better error logging by means of `non_fatal`. This preserves the
original error and allows adding context information on top. Also makes
error formatting more uniform across the application.
Also add per-instruction error context to make it clear what we tried to
accomplish when an error occured.
* server/panes/tiled_panes: Merge dependencies
and sort them into a better order.
* server/panes/tiled_panes: Replace `log::error!`
with better error logging by means of `non_fatal`. This preserves the
original error and allows adding context information on top. Also makes
error formatting more uniform across the application.
* server/os_input_output: Merge depndencies
and sort them into a better order.
* server/logging_pipe: Replace `log::error!`
with better error logging by means of `non_fatal`. This preserves the
original error and allows adding context information on top. Also makes
error formatting more uniform across the application.
* server/os_io: Remove uses of `log::error`
* changelog: Add PR #1881
* server/os_io: Gracefully handle failing resize
for terminals IDs that don't exist, instead of propagating the error to
the user.
* server/lib: Remove leftover log message
* server/pty: Log error cause
rather than providing a hard-coded error reason which is plain wrong in
this context.
* server/screen: Remove calls to `log::error!`
and change `get_active_tab(_mut)?` to return a `Result` instead of an
`Option`. This already makes many places in the code obsolete where
previously "failed to get active tab..." was logged manually.
Rather than logging, use the `anyhow::Error`s we have, along with all
their context information, and log these instead.
* server/os_io: Redefine `ServerOsApi` result types
to use `anyhow::Result` instead. This mostly makes the need of custom
`SpawnTerminalError` obsolete (tbd in subsequent commits) and unifies
error handling across the application.
* utils/errors: Implement new `ZellijError` type
to replace any previously defined, isolated custom error types
throughout the application. Currently implements all error variants
found in `SpawnTerminalError`.
In the long term, this will allow zellij to fall back to a single error
type for all application-specific errors, instead of having different
error types per module.
* server/unit/screen: Impl new `ServerOsApi`
with updated `Result`-types.
* server/tab/unit: Impl new `ServerOsApi`
with updated `Result`-types.
* server/os_io: Impl new `ServerOsApi`
with updated `Result`-types.
* utils/ipc: Return `anyhow::Error` in `send`
rather than a `&'static str`, which isn't compatible with
`anyhow::Context`.
* server/tab: Handle `Result` in `resize_pty!`
which is returned due to the changed return types in `ServerOsApi`.
* server/tab: Handle new `Result`s
originating in the change to the `ServerOsApi` trait definition.
* server/screen: Handle new `Result`s
originating in the change to the `ServerOsApi` trait definition.
* server/panes/tiled: Handle new `Result`s
originating in the change to the `ServerOsApi` trait definition.
* server/panes/floating: Handle new `Result`s
originating in the change to the `ServerOsApi` trait definition.
* server/lib: Unwrap on new `Result`s
originating in the change to the `ServerOsApi` trait definition. The
functions here don't return a `Result` yet, this is better left to a
follow-up PR.
* server: Remove `SpawnTerminalError`
and make use of the new `ZellijError` instead. Make use of `anyhow`s
downcast capabilities to restore the underlying original errors where
necessary, as was done previously. This gives us the flexibility to
attach context information to all errors while still allowing us to
handle specific errors in greater detail.
* server/pty: Fix vars broken in rebase
* server/os_io: Remove last `SpawnTerminalError`
* changelog: Add PR #1895
* server/tab: Don't panic in `Pane::render`
and do not crash the application on failure to receive a render update
from plugins any longer. Instead, will print a simple string with a hint
to check the application logs, where a more thorough error indication
can be found.
* utils/errors: re-export `anyhow::Error`
to create ad-hoc errors with custom error types, without having to wrap
them into a `context()` before to turn the into anyhow errors.
* plugins: Check plugin version on startup
and terminate execution with a descriptive error message in case the
plugin version is incompatible with the version of zellij being run.
* server/wasm_vm: Add plugin path in version error
so the user knows which plugin to look at in case they're using custom
plugins.
* server/wasm_vm: Check plugin version for equality
Previously we would accept cases where the plugin version was newer than
the zellij version, which doesn't make a lot of sense.
* server/wasm_vm: Prettier error handling
in call to `wasmer::Function::call` in case a plugin version mismatch
can occur.
* tile: Install custom panic handler
that will print the panic message to a plugins stdout and then call a
panic handler on the host that turns it into a real application-level
panic.
* tile: Catch errors in event deserialization
and turn them into proper panics. These errors are symptomatic of an
uncaught plugin version mismatch, for example when developing from main
and compiling zellij/the plugins from source. Normal users should never
get to see this error.
* utils/errors: Improve output in `to_stdout`
for anyhow errors. The default anyhow error formatting of `{:?}` is
already very good, and we just made it worse by trying to invent our own
formatting.
* tile: Reword plugin mismatch error message
* zellij: Apply rustfmt
* changelog: Add PR #1838
Improve error handling on plugin version mismatch.
* server/wasm_vm: Rephrase error in passive voice
* server/wasm_vm: Compact module imports
* utils/errors: Impl `to_anyhow` for PoisonError
which is returned by calls to `lock` on various types of locks from
`std`. In our case, some of the locks we try to acquire in `wasm_vm` can
contain an `mpsc::Sender`, which is `!Send` and hence doesn't work with
`anyhow`. Turn the `PoisonError` into an error string instead and
returns that as `anyhow::Err`.
* wasm_vm: Remove calls to `unwrap`
in the WASM VM codes server API. Note that this doesn't include the
Plugin APIs. Mark the error as `fatal` in `server/lib`, where the wasm
thread is created.
This will cause zellij to report a proper error (and log it) when any of
the plugin-related functions fails. Unfortunately, this closes the
channel to the WASM thread. Hence, when loading the plugins upon startup
fails, the error reported in the terminal (visible to the user) hints
towards a call in `plugin_pane` being the culprit. However, the real
error will be contained in the logs.
Also add an error message and print it to the user in case that the
plugin failure was caused by a plugin version mismatch.
* server/wasm_vm: Restore panic on failure to load
plugins.
* server/wasm_vm: Add fix to plugin mismatch error
* server/panes/plugin_pane: Hint to logs
when failing to receive a message from the plugins for rendering pane
contents.
* feat(cli): move command to the end of the cli arguments
* feat(cli): allow naming panes from the command line
* fix(cli): adjust actions after pane rename
* feat(cli): zellij run completions for fish
* feat(cli): zellij run completions for bash and zsh
* style(fmt): rustfmt
* fix(e2e): fix run test and snapshot
* style(fmt): rustfmt
* feat(layouts): allow defining a global cwd
* feat(layouts): allow passing global cwd from cli
* style(fmt): rustfmt
* fix(layouts): error on mixed cwd and pane children
* fix(layouts): error on non-bare children node
* refactor(layout): consolidate split direction parsing
* refactor(layout): remove unused import
* fix(layouts): log error when there is no room for layout
* fix(layout): error on size 0
* feat(layouts): allow pane templates to override template command attributes
* style(fmt): rustfmt
* chore(config): default kdl keybindings config
* tests
* work
* refactor(config): move stuff around
* work
* tab merge layout
* work
* work
* layouts working
* work
* layout tests
* work
* work
* feat(parsing): kdl layouts without config
* refactor(kdl): move stuff around
* work
* tests(layout): add cases and fix bugs
* work
* fix(kdl): various bugs
* chore(layouts): move all layouts to kdl
* feat(kdl): shared keybidns
* fix(layout): do not count fixed panes toward percentile
* fix(keybinds): missing keybinds and actions
* fix(config): adjust default tips
* refactor(config): move stuff around
* fix(tests): make e2e tests pass
* fix(kdl): add verbose parsing errors
* fix(kdl): focused tab
* fix(layout): corret default_tab_template behavior
* style(code): fix compile warnings
* feat(cli): send actions through the cli
* fix(cli): exit only when action is done
* fix(cli): open embedded pane from floating pane
* fix(cli): send actions to other sessions
* feat(cli): command alias
* feat(converter): convert old config
* feat(converter): convert old layout and theme files
* feat(kdl): pretty errors
* feat(client): convert old YAML files on startup
* fix: various bugs and styling issues
* fix: e2e tests
* fix(screen): propagate errors after merge
* style(clippy): lower clippy level
* fix(tests): own session_name variable
* style(fmt): rustfmt
* fix(cli): various action fixes
* style(fmt): rustfmt
* fix(themes): loading of theme files
* style(fmt): rustfmt
* fix(tests): theme fixtures
* fix(layouts): better errors on unknown nodes
* fix(kdl): clarify valid node terminator error
* fix(e2e): adjust close tab test
* fix(e2e): adjust close tab test again
* style(code): cleanup some comments
* get command panes not to exit on command exit
* separate terminal pane_ids from raw_fds
* render frame according to exit status
* re-run command on enter and close pane on ctrl-c
* proper error when command is not found
* make ui nicer
* initial pane title for command panes
* fix pane override bug
* reap terminal_ids from os_input_output on pane close
* bool floating flag
* some ui tweaks
* fix tests
* make rustfmt happy
* e2e test for command pane
* fix various concurrency issues
* rename command to run in the cli
* rustfmt
* style(fmt): rustfmt
* fix(e2e): command => run
* fix(e2e): command => run in snapshot too!
* utils/errors: Add `ToAnyhow` trait
for converting `Result` types that don't satisfy `anyhow`s trait
constraints (`Display + Send + Sync + 'static`) conveniently.
An example of such a Result is the `SendError` returned from
`send_to_plugins`, which sends `PluginInstruction`s as message type.
One of the enum variants can contain a `mpsc::Sender`, which is `!Sync`
and hence makes the whole `SendError` be `!Sync` in this case. Add an
implementation for this case that takes the message and converts it into
an error containing the message formatted as string, with the additional
`ErrorContext` as anyhow context.
* server/tab: Remove calls to `unwrap()`
and apply error reporting via `anyhow` instead. Make all relevant
functions return `Result`s where previously a panic could occur and
attach error context.
* server/screen: Modify `update_tab!`
to accept an optional 4th parameter, a literal "?". If present, this
will append a `?` to the given closure verbatim to handle/propagate
errors from within the generated macro code.
* server/screen: Handle new `Result`s from `Tab`
and apply appropriate error context and propagate errors further up.
* server/tab/unit: `unwrap` on new `Result`s
* server/unit: Unwrap `Results` in screen tests
* server/tab: Better message for ad-hoc errors
created with `anyhow!`. Since these errors don't have an underlying
cause, we describe the cause in the macro instead and then attach the
error context as usual before `?`ing the error back up.
* utils/cargo: Activate `anyhow`s "backtrace" feature
to capture error backtraces at the error origins (i.e. where we first
receive an error and convert it to a `anyhow::Error`). Since we
propagate error back up the call stack now, the place where we `unwrap`
on errors doesn't match the place where the error originated. Hence, the
callstack, too, is quite misleading since it contains barely any
references of the functions that triggered the error.
As a consequence, we have 2 backtraces now when zellij crashes: One from
`anyhow` (that is implicitly attached to anyhows error reports), and one
from the custom panic handler (which is displayed through `miette`).
* utils/errors: Separate stack traces
in the output of miette. Since we record backtraces with `anyhow` now,
we end up having two backtraces in the output: One from the `anyhow`
error and one from the actual call to `panic`. Adds a comment explaining
the situation and another "section" to the error output of miette: We
print the backtrace from anyhow as "Stack backtrace", and the output
from the panic handler as "Panic backtrace". We keep both for the
(hopefully unlikely) case that the anyhow backtrace isn't existent, so
we still have at least something to work with.
* server/screen: Remove calls to `fatal`
and leave the `panic`ing to the calling function instead.
* server/screen: Remove needless macro
which extended `active_tab!` by passing the client IDs to the closure.
However, this isn't necessary because closures capture their environment
already, and the client_id needn't be mutable.
* server/screen: Handle unused result
* server/screen: Reintroduce arcane macro
that defaults to some default client_id if it isn't valid (e.g. when the
ScreenInstruction is sent via CLI).
* server/tab/unit: Unwrap new results
