* Pass completion options to each fetch() call
* Add MatchAlgorithm to CompletionOptions
* Add unit test for MatchAlgorithm
* Pass completion options to directory completer
Changes to reedline since `v0.4.0`:
- vi normal mode `I` for inserting at line beginning
- `InsertNewline` edit command that can be bound to `Alt-Enter` if
desired to have line breaks without relying on the `Validator`
- `ClearScreen` will directly clear the visible screen. `Signal::CtrlL` has been
removed.
- `ClearScrollback` will clear the screen and scrollback. Can be used to
mimic macOS `Cmd-K` screen clearing. Helps with #5089
* nu-cli: added tests for file completions
* test adding extra sort
* Feature/refactor completion options (#5228)
* Copy completion filter to custom completions
* Remove filter function from completer
This function was a no-op for FileCompletion and CommandCompletion.
Flag- and VariableCompletion just filters with `starts_with` which
happens in both completers anyway and should therefore also be a no-op.
The remaining use case in CustomCompletion was moved into the
CustomCompletion source file.
Filtering should probably happen immediately while fetching completions
to avoid unnecessary memory allocations.
* Add get_sort_by() to Completer trait
* Remove CompletionOptions from Completer::fetch()
* Fix clippy lints
* Apply Completer changes to DotNuCompletion
* add os to $nu based on rust's understanding (#5243)
* add os to $nu based on rust's understanding
* add a few more constants
Co-authored-by: Richard <Tropid@users.noreply.github.com>
Co-authored-by: Darren Schroeder <343840+fdncred@users.noreply.github.com>
* Copy completion filter to custom completions
* Remove filter function from completer
This function was a no-op for FileCompletion and CommandCompletion.
Flag- and VariableCompletion just filters with `starts_with` which
happens in both completers anyway and should therefore also be a no-op.
The remaining use case in CustomCompletion was moved into the
CustomCompletion source file.
Filtering should probably happen immediately while fetching completions
to avoid unnecessary memory allocations.
* Add get_sort_by() to Completer trait
* Remove CompletionOptions from Completer::fetch()
* Fix clippy lints
* Apply Completer changes to DotNuCompletion
Prior to this change we would recover the names for known
externals by looking up the span in the engine state. This would fail
when using an alias for two reasons:
1. In cases where we don't have a subcommand, like this:
```
>>> extern bat [filename: string]
>>> alias b = bat
>>> bat some_file
'b' is not recognized as an internal or external command,
operable program or batch file.
```
The problem is that after alias expansion, we replace the span of the
expanded name with the original alias (this is done to alleviate
non-related issues). The span contents we look up therefore contain `b`,
the alias, instead of the expanded command name.
2. In cases where there's a subcommand:
```
>>> alias g = git
>>> g push
thread 'main' panicked at 'internal error: span missing in file contents cache', crates\nu-protocol\src\engine\engine_state.rs:474:9
note: run with `RUST_BACKTRACE=1` environment variable to display a
backtrace
```
In this case, the span in call starts where the expansion for the `g`
alias is defined and end after `push` on the last command entered. This
is not a proper span and causes a panic when we try to look it up. Note
that this is the case for all expanded aliases that involve a
subcommand, but we never actually try to retrieve the contents for that
span in other cases.
Anyway, the new way of looking up the name is arguably cleaner
regardless of the issues mentioned above. But it's nice that it fixes
them too.
Co-authored-by: Hristo Filaretov <h.filaretov@protonmail.com>
This commit renders ANSI chars in order to provide shell integrations
such Kitty's opening feature that captures the output of the last
command in a pager such as less.
Fixes#5138
- With a change to reedline hints can now be hidden. This is useful when
no ANSI coloring is available as hints become indistinguishable from the
actual buffer
- remove commented out code
- order the logging calls according to the implementation
I don't personally agree with this; I'd prefer less magic,
and not expanding _anything_ except `~` as an initial path element
Co-authored-by: nicole mazzuca <mazzucan@outlook.com>
To remove a default keybinding for a particular edit mode, set the `event: null`:
e.g. to disable screen clearing with Ctrl-L
```
let $config = {keybindings: [{
modifier: control
keycode: char_l
mode: [emacs, vi_normal, vi_insert]
event: null
} ]}
```
* Remove unused packages from base Cargo.toml
* Remove unused crossterm_winapi from nu-cli
* Remove unused dependencies from nu-system
* Remove unused dependencies from nu-test-support
* Initial implementation of ordered call args
* Run cargo fmt
* Fix some clippy lints
* Add positional len and nth
* Cargo fmt
* Remove more old nth calls
* Good ole rustfmt
* Add named len
Co-authored-by: Hristo Filaretov <h.filaretov@protonmail.com>
* Include license text in all crates
Three crates already have license texts, so I'm keeping them, but
symlinking the `LICENSE` from the top level to the rest of the crate
directories. This works as long as `cargo publish` is done on a Unix-y
system and not Windows.
Also bump the copyright year to end in 2022.
Signed-off-by: Michel Alexandre Salim <salimma@fedoraproject.org>
* Replace symlinks
Co-authored-by: sholderbach <sholderbach@users.noreply.github.com>
* Add timestamp flag to `touch` command
* Add modify flag to `touch` command
* Add date flag to `touch` command
* Remove unnecessary `touch` test and fix tests setups
* Change `touch` flags descriptions
* Update `touch` example
* Add reference flag to `touch` command
* Add access flag to `touch` command
* Add no-create flag to `touch` command
* Replace `unwrap` with `expect`
