# Description
Restores `bytes starts-with` so that it is able to work with byte
streams once again. For parity/consistency, this PR also adds byte
stream support to `bytes ends-with`.
# User-Facing Changes
- `bytes ends-with` now supports byte streams.
# Tests + Formatting
Re-enabled tests for `bytes starts-with` and added tests for `bytes
ends-with`.
# Description
This PR adds a few functions to `Span` for merging spans together:
- `Span::append`: merges two spans that are known to be in order.
- `Span::concat`: returns a span that encompasses all the spans in a
slice. The spans must be in order.
- `Span::merge`: merges two spans (no order necessary).
- `Span::merge_many`: merges an iterator of spans into a single span (no
order necessary).
These are meant to replace the free-standing `nu_protocol::span`
function.
The spans in a `LiteCommand` (the `parts`) should always be in order
based on the lite parser and lexer. So, the parser code sees the most
usage of `Span::append` and `Span::concat` where the order is known. In
other code areas, `Span::merge` and `Span::merge_many` are used since
the order between spans is often not known.
# Description
This PR introduces a `ByteStream` type which is a `Read`-able stream of
bytes. Internally, it has an enum over three different byte stream
sources:
```rust
pub enum ByteStreamSource {
Read(Box<dyn Read + Send + 'static>),
File(File),
Child(ChildProcess),
}
```
This is in comparison to the current `RawStream` type, which is an
`Iterator<Item = Vec<u8>>` and has to allocate for each read chunk.
Currently, `PipelineData::ExternalStream` serves a weird dual role where
it is either external command output or a wrapper around `RawStream`.
`ByteStream` makes this distinction more clear (via `ByteStreamSource`)
and replaces `PipelineData::ExternalStream` in this PR:
```rust
pub enum PipelineData {
Empty,
Value(Value, Option<PipelineMetadata>),
ListStream(ListStream, Option<PipelineMetadata>),
ByteStream(ByteStream, Option<PipelineMetadata>),
}
```
The PR is relatively large, but a decent amount of it is just repetitive
changes.
This PR fixes#7017, fixes#10763, and fixes#12369.
This PR also improves performance when piping external commands. Nushell
should, in most cases, have competitive pipeline throughput compared to,
e.g., bash.
| Command | Before (MB/s) | After (MB/s) | Bash (MB/s) |
| -------------------------------------------------- | -------------:|
------------:| -----------:|
| `throughput \| rg 'x'` | 3059 | 3744 | 3739 |
| `throughput \| nu --testbin relay o> /dev/null` | 3508 | 8087 | 8136 |
# User-Facing Changes
- This is a breaking change for the plugin communication protocol,
because the `ExternalStreamInfo` was replaced with `ByteStreamInfo`.
Plugins now only have to deal with a single input stream, as opposed to
the previous three streams: stdout, stderr, and exit code.
- The output of `describe` has been changed for external/byte streams.
- Temporary breaking change: `bytes starts-with` no longer works with
byte streams. This is to keep the PR smaller, and `bytes ends-with`
already does not work on byte streams.
- If a process core dumped, then instead of having a `Value::Error` in
the `exit_code` column of the output returned from `complete`, it now is
a `Value::Int` with the negation of the signal number.
# After Submitting
- Update docs and book as necessary
- Release notes (e.g., plugin protocol changes)
- Adapt/convert commands to work with byte streams (high priority is
`str length`, `bytes starts-with`, and maybe `bytes ends-with`).
- Refactor the `tee` code, Devyn has already done some work on this.
---------
Co-authored-by: Devyn Cairns <devyn.cairns@gmail.com>
# Description
Fixes: #12691
In `parse_short_flag`, it only checks special cases for
`SyntaxShape::Int`, `SyntaxShape::Number` to allow a flag to be a
number. This pr adds `SyntaxShape::Float` to allow a flag to be float
number.
# User-Facing Changes
This is possible after this pr:
```nushell
def spam [val: float] { $val };
spam -1.4
```
# Tests + Formatting
Added 1 test
# Description
In order for `Stack::unwrap_unique` to work as intended, we currently
manually track all references to the parent stack and ensure that they
are cleared before calling `Stack::unwrap_unique` in the REPL. We also
only call `Stack::unwrap_unique` after all code from the current REPL
entry has finished executing. Since `Value`s cannot store `Stack`
references, then this should have worked in theory. However, we forgot
to account for threads. `run-external` (and maybe the plugin writers)
can spawn threads that clone the `Stack`, holding on to references of
the parent stack. These threads are not waited/joined upon, and so may
finish after the eval has already returned. This PR removes the
`Stack::unwrap_unique` function and associated debug assert that was
[causing
panics](https://gist.github.com/cablehead/f3d2608a1629e607c2d75290829354f7)
like @cablehead found.
# After Submitting
Make values cheaper to clone as a more robust solution to the
performance issues with cloning the stack.
---------
Co-authored-by: Wind <WindSoilder@outlook.com>
Fix for #12730
All of the code expected a list of floats, but the syntax shape expected
a table. Resolved by changing the syntax shape to list of floats.
cc: @maxim-uvarov
# Description
Following from #12867, this PR replaces usages of `Call::positional_nth`
with existing spans. This removes several `expect`s from the code.
Also remove unused `positional_nth_mut` and `positional_iter_mut`
# Description
So minor, but had to be fixed sometime. `help each while` used the term
"block" in the "usage", but the argument type is a closure.
# User-Facing Changes
help-only
# Description
This should fix#10155 where the `sys` command can panic due to date
math in certain cases / on certain systems.
# User-Facing Changes
The `boot_time` column now has a date value instead of a formatted date
string. This is technically a breaking change.
# Description
Fixes: #12795
The issue is caused by an empty position of `ParseError::UnexpectedEof`.
So no detailed message is displayed.
To fix the issue, I adjust the start of span to `span.end - 1`. In this
way, we can make sure that it never points to an empty position.
After lexing item, I also reorder the unclosed character checking . Now
it will be checking unclosed opening delimiters first.
# User-Facing Changes
After this pr, it outputs detailed error message for incomplete string
when running scripts.
## Before
```
❯ nu -c "'ab"
Error: nu::parser::unexpected_eof
× Unexpected end of code.
╭─[source:1:4]
1 │ 'ab
╰────
> ./target/debug/nu -c "r#'ab"
Error: nu::parser::unexpected_eof
× Unexpected end of code.
╭─[source:1:6]
1 │ r#'ab
╰────
```
## After
```
> nu -c "'ab"
Error: nu::parser::unexpected_eof
× Unexpected end of code.
╭─[source:1:3]
1 │ 'ab
· ┬
· ╰── expected closing '
╰────
> ./target/debug/nu -c "r#'ab"
Error: nu::parser::unexpected_eof
× Unexpected end of code.
╭─[source:1:5]
1 │ r#'ab
· ┬
· ╰── expected closing '#
╰────
```
# Tests + Formatting
Added some tests for incomplete string.
---------
Co-authored-by: Ian Manske <ian.manske@pm.me>
# Description
The `char` command can panic due to a failed `expect`: `char --integer
...[77 78 79]`
This PR fixes the panic for the `--integer` flag and also the
`--unicode` flag.
# After Submitting
Check other commands and places where similar bugs can occur due to
usages of `Call::positional_nth` and related methods.
# Description
There was a question in Discord today about how to remove empty rows
from a table. The user found the `compact` command on their own, but I
realized that there were no search terms on the command. I've added
'empty' and 'remove', although I subsequently figured out that 'empty'
is found in the "usage" anyway. That said, I don't think it hurts to
have good search terms behind it regardless.
# User-Facing Changes
Just the help
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
# After Submitting
- fixes#12841
# Description
Add boundary checks to ensure that the row and column chosen in
RecordView are not over the length of the possible row and columns. If
we are out of bounds, we default to Value::nothing.
# Tests + Formatting
Tests ran and formatting done
- fixes#12764
Replaced the custom logic with values_to_sql method that is already used
in crate::database.
This will ensure that handling of parameters is the same between sqlite
and stor.
# Description
In this PR I added two new methods to `Stack`, `stdout_file` and
`stderr_file`. These two modify the inner `StackOutDest` and set a
`File` into the `stdout` and `stderr` respectively. Different to the
`push_redirection` methods, these do not require to hold a guard up all
the time but require ownership of the stack.
This is primarly useful for applications that use `nu` as a language but
not the `nushell`.
This PR replaces my first attempt #12851 to add a way to capture
stdout/-err of external commands. Capturing the stdout without having to
write into a file is possible with crates like
[`os_pipe`](https://docs.rs/os_pipe), an example for this is given in
the doc comment of the `stdout_file` command and can be executed as a
doctest (although it doesn't validate that you actually got any data).
This implementation takes `File` as input to make it easier to implement
on different operating systems without having to worry about
`OwnedHandle` or `OwnedFd`. Also this doesn't expose any use `os_pipe`
to not leak its types into this API, making it depend on it.
As in my previous attempt, @IanManske guided me here.
# User-Facing Changes
This change has no effect on `nushell` and therefore no user-facing
changes.
# Tests + Formatting
This only exposes a new way of using already existing code and has
therefore no further testing. The doctest succeeds on my machine at
least (x86 Windows, 64 Bit).
# After Submitting
All the required documentation is already part of this PR.
# Description
It's commonly forgotten or overlooked that a lot of `std repeat`
functionality can be handled with the built-in `fill`. Added 'repeat` as
a search term for `fill` to improve discoverability.
Also replaced one of the existing examples with one `fill`ing an empty
string, a la `repeat`. There were 6 examples already, and 3 of them
pretty much were variations on the same theme, so I repurposed one of
those rather than adding a 7th.
# User-Facing Changes
Changes to `help` only
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
# After Submitting
I assume the "Commands" doc is auto-generated from the `help`, but I'll
double-check that assumption.
# Description
This PR resolves an inconsistency between different `str` subcommands,
notably `str contains`, `str starts-with` and `str ends-with`. Only the
`str contains` command has the `--not` flag and a desicion was made in
this #12781 PR to remove the `--not` flag and use the `not` operator
instead.
Before:
`"blob" | str contains --not o`
After:
`not ("blob" | str contains o)` OR `"blob" | str contains o | not $in`
> Note, you can currently do all three, but the first will be broken
after this PR is merged.
# User-Facing Changes
- remove `--not(-n)` flag from `str contains` command
- This is a breaking change!
# Tests + Formatting
- [x] Added tests
- [x] Ran `cargo fmt --all`
- [x] Ran `cargo clippy --workspace -- -D warnings -D
clippy::unwrap_used`
- [x] Ran `cargo test --workspace`
- [ ] Ran `cargo run -- -c "use std testing; testing run-tests --path
crates/nu-std"`
- I was unable to get this working.
```
Error: nu::parser::export_not_found
× Export not found.
╭─[source:1:9]
1 │ use std testing; testing run-tests --path crates/nu-std
· ───┬───
· ╰── could not find imports
╰────
```
^ I still can't figure out how to make this work 😂
# After Submitting
Requires update of documentation
# Description
Fixes#12796 where a combined out and err pipe redirection (`o+e>|`)
into `complete` still provides separate `stdout` and `stderr` columns in
the record. Now, the combined output will be in the `stdout` column.
This PR also fixes a similar error with the `e>|` pipe redirection.
# Tests + Formatting
Added two tests.
This PR has two parts. The first part is the addition of the
`Stack::set_pwd()` API. It strips trailing slashes from paths for
convenience, but will reject otherwise bad paths, leaving PWD in a good
state. This should reduce the impact of faulty code incorrectly trying
to set PWD.
(https://github.com/nushell/nushell/pull/12760#issuecomment-2095393012)
The second part is implementing a PWD recovery mechanism. PWD can become
bad even when we did nothing wrong. For example, Unix allows you to
remove any directory when another process might still be using it, which
means PWD can just "disappear" under our nose. This PR makes it possible
to use `cd` to reset PWD into a good state. Here's a demonstration:
```sh
mkdir /tmp/foo
cd /tmp/foo
# delete "/tmp/foo" in a subshell, because Nushell is smart and refuse to delete PWD
nu -c 'cd /; rm -r /tmp/foo'
ls # Error: × $env.PWD points to a non-existent directory
# help: Use `cd` to reset $env.PWD into a good state
cd /
pwd # prints /
```
Also, auto-cd should be working again.
# Description
Refactors the code in `nu-cli`, `main.rs`, `run.rs`, and few others.
Namely, I added `EngineState::generate_nu_constant` function to
eliminate some duplicate code. Otherwise, I changed a bunch of areas to
return errors instead of calling `std::process::exit`.
# User-Facing Changes
Should be none.
# Description
Fixes#12813 where a panic occurs when syntax highlighting `not`. Also
fixes#12814 where syntax highlighting for `not` no longer works.
# User-Facing Changes
Bug fix.
# Description
Adds an additional `&Stack` parameter to `Completer::fetch` so that the
completers don't have to store a `Stack` themselves. I also removed
unnecessary `EngineState`s from the completers, since the same
`EngineState` is available in the `working_set.permanent_state` also
passed to `Completer::fetch`.
# Description
Changes the iterator in `rm` to be an iterator over
`Result<Option<String>, ShellError>` (an optional message or error)
instead of an iterator over `Value`. Then, the iterator is consumed and
each message is printed. This allows the
`PipelineData::print_not_formatted` method to be removed.
# Description
On 64-bit platforms the current size of `Value` is 56 bytes. The
limiting variants were `Closure` and `Range`. Boxing the two reduces the
size of Value to 48 bytes. This is the minimal size possible with our
current 16-byte `Span` and any 24-byte `Vec` container which we use in
several variants. (Note the extra full 8-bytes necessary for the
discriminant or other smaller values due to the 8-byte alignment of
`usize`)
This is leads to a size reduction of ~15% for `Value` and should overall
be beneficial as both `Range` and `Closure` are rarely used compared to
the primitive types or even our general container types.
# User-Facing Changes
Less memory used, potential runtime benefits.
(Too late in the evening to run the benchmarks myself right now)
# Description
This PR is a continuation of #12629 and meant to address [Reilly's
stated
issue](https://github.com/nushell/nushell/pull/12629#issuecomment-2099660609).
With this PR, nushell should work more consistently with WezTerm on
Windows. However, that means continued scrolling with typing if osc133
is enabled. If it's possible to run WezTerm inside of vscode, then
having osc633 enabled will also cause the display to scroll with every
character typed. I think the cause of this is that reedline paints the
entire prompt on each character typed. We need to figure out how to fix
that, but that's in reedline.
For my purposes, I keep osc133 and osc633 set to true and don't use
WezTerm on Windows.
Thanks @rgwood for reporting the issue. I found several logic errors.
It's often good to come back to PRs and look at them with fresh eyes. I
think this is pretty close to logically correct now. However, I'm
approaching burn out on ansi escape codes so i could've missed
something.
Kudos to [escape-artist](https://github.com/rgwood/escape-artist) for
helping me debug an ansi escape codes that are actually being sent to
the terminal. It was an invaluable tool.
# User-Facing Changes
<!-- List of all changes that impact the user experience here. This
helps us keep track of breaking changes. -->
# Tests + Formatting
<!--
Don't forget to add tests that cover your changes.
Make sure you've run and fixed any issues with these commands:
- `cargo fmt --all -- --check` to check standard code formatting (`cargo
fmt --all` applies these changes)
- `cargo clippy --workspace -- -D warnings -D clippy::unwrap_used` to
check that you're using the standard code style
- `cargo test --workspace` to check that all tests pass (on Windows make
sure to [enable developer
mode](https://learn.microsoft.com/en-us/windows/apps/get-started/developer-mode-features-and-debugging))
- `cargo run -- -c "use std testing; testing run-tests --path
crates/nu-std"` to run the tests for the standard library
> **Note**
> from `nushell` you can also use the `toolkit` as follows
> ```bash
> use toolkit.nu # or use an `env_change` hook to activate it
automatically
> toolkit check pr
> ```
-->
# After Submitting
<!-- If your PR had any user-facing changes, update [the
documentation](https://github.com/nushell/nushell.github.io) after the
PR is merged, if necessary. This will help us keep the docs up to date.
-->
# Description
- Switches the `excess` in the `ParserStream` and
`ParseStreamerExternal` types from a `Vec` to a `VecDeque`
- Removes unnecessary clones to `stream_helper`
- Other simplifications and loop restructuring
- Merges the `ParseStreamer` and `ParseStreamerExternal` types into a
common `ParseIter`
- `parse` now streams for list values
# Description
This changes the message that shows up when running a plugin executable
directly rather than as a plugin to direct the user to run `plugin add
--help`, which should have enough information to figure out what's going
on. The message previously just vaguely suggested that the user needs to
run the plugin "from within Nushell", which is not really enough - it
has to be added with `plugin add` to be used as a plugin.
Also fix docs for `plugin add` to mention `plugin use` rather than
`register` (oops)
Refer to #12603 for part 1.
We need to be careful when migrating to the new API, because the new API
has slightly different semantics (PWD can contain symlinks). This PR
handles the "obviously safe" part of the migrations. Namely, it handles
two specific use cases:
* Passing PWD into `canonicalize_with()`
* Passing PWD into `EngineState::merge_env()`
The first case is safe because symlinks are canonicalized away. The
second case is safe because `EngineState::merge_env()` only uses PWD to
call `std::env::set_current_dir()`, which shouldn't affact Nushell. The
commit message contains detailed stats on the updated files.
Because these migrations touch a lot of files, I want to keep these PRs
small to avoid merge conflicts.
# Description
Add a new `sys users` command which returns a table of the users of the
system. This is the same table that is currently present as
`(sys).host.sessions`. The same table has been removed from the recently
added `sys host` command.
# User-Facing Changes
Adds a new command. (The old `sys` command is left as is.)
# Description
Refactors `describe` a bit. Namely, I added a `Description` enum to get
rid of `compact_primitive_description` and its awkward `Value` pattern
matching.
# Description
Adds subcommands to `sys` corresponding to each column of the record
returned by `sys`. This is to alleviate the fact that `sys` now returns
a regular record, meaning that it must compute every column which might
take a noticeable amount of time. The subcommands, on the other hand,
only need to compute and return a subset of the data which should be
much faster. In fact, it should be as fast as before, since this is how
the lazy record worked (it would compute only each column as necessary).
I choose to add subcommands instead of having an optional cell-path
parameter on `sys`, since the cell-path parameter would:
- increase the code complexity (can access any value at any row or
nested column)
- prevents discovery with tab-completion
- hinders type checking and allows users to pass potentially invalid
columns
# User-Facing Changes
Deprecates `sys` in favor of the new `sys` subcommands.
This moves to predominantly supporting only lazy dataframes for most
operations. It removes a lot of the type conversion between lazy and
eager dataframes based on what was inputted into the command.
For the most part the changes will mean:
* You will need to run `polars collect` after performing operations
* The into-lazy command has been removed as it is redundant.
* When opening files a lazy frame will be outputted by default if the
reader supports lazy frames
A list of individual command changes can be found
[here](https://hackmd.io/@nucore/Bk-3V-hW0)
---------
Co-authored-by: Ian Manske <ian.manske@pm.me>
# Description
Fixes: #12744
This pr is moving raw string lex logic into `lex_item` function, so we
can use raw string inside subexpression, list, closure.
```nushell
> [r#'abc'#]
╭───┬─────╮
│ 0 │ abc │
╰───┴─────╯
> (r#'abc'#)
abc
> do {r#'aa'#}
aa
```
# Tests + Formatting
Done
# After Submitting
NaN
# Description
Fixes: #12429
To fix the issue, we need to pass the `input pattern` itself to
`glob_from` function, but currently on latest main, nushell pass
`expanded path of input pattern` to `glob_from` function.
It causes globbing failed if expanded path includes `[]` brackets.
It's a pity that I have to duplicate `nu_engine::glob_from` function
into `ls`, because `ls` might convert from `NuGlob::NotExpand` to
`NuGlob::Expand`, in that case, `nu_engine::glob_from` won't work if
user want to ls for a directory which includes tilde:
```
mkdir "~abc"
ls "~abc"
```
So I need to duplicate `glob_from` function and pass original
`expand_tilde` information.
# User-Facing Changes
Nan
# Tests + Formatting
Done
# After Submitting
Nan
# Description
Does some misc changes to `ListStream`:
- Moves it into its own module/file separate from `RawStream`.
- `ListStream`s now have an associated `Span`.
- This required changes to `ListStreamInfo` in `nu-plugin`. Note sure if
this is a breaking change for the plugin protocol.
- Hides the internals of `ListStream` but also adds a few more methods.
- This includes two functions to more easily alter a stream (these take
a `ListStream` and return a `ListStream` instead of having to go through
the whole `into_pipeline_data(..)` route).
- `map`: takes a `FnMut(Value) -> Value`
- `modify`: takes a function to modify the inner stream.
# Description
this PR
- moves the documentation from `lib.rs` to `README.md` while still
including it in the lib file, so that both the [crates.io
page](https://crates.io/crates/nuon) and the
[documentation](https://docs.rs/nuon/latest/nuon/) show the top-level
doc
- mention that comments are allowed in NUON
- add a JSON-NUON example
- put back the formatting of NOTE blocks in the doc
# User-Facing Changes
# Tests + Formatting
# After Submitting
# Description
Our current flattening code creates a bunch of intermediate `Vec`s for
each function call. These intermediate `Vec`s are then usually appended
to the current `output` `Vec`. By instead passing a mutable reference of
the `output` `Vec` to each flattening function, this `Vec` can be
reused/appended to directly thereby eliminating the need for
intermediate `Vec`s in most cases.
# Description
Make typos config more strict: ignore false positives where they occur.
1. Ignore only files with typos
2. Add regexp-s with context
3. Ignore variable names only in Rust code
4. Ignore only 1 "identifier"
5. Check dot files
🎁 Extra bonus: fix typos!!
# Description
Spaces were causing an issue with into_sqlite when they appeared in
column names.
This is because the column names were not properly wrapped with
backticks that allow sqlite to properly interpret the column.
The issue has been addressed by adding backticks to the column names of
into sqlite. The output of the column names when using open is
unchanged, and the column names appear without backticks as expected.
fixes#12700
# User-Facing Changes
N/A
# Tests + Formatting
Formatting has been respected.
Repro steps from the issue have been done, and ran multiple times. New
values get added to the correct columns as expected.
# Description
Bumps `base64` to 0.22.1 which fixes the alphabet used for binhex
encoding and decoding. This required updating some test expected output.
Related to PR #12469 where `base64` was also bumped and ran into the
failing tests.
# User-Facing Changes
Bug fix, but still changes binhex encoding and decoding output.
# Tests + Formatting
Updated test expected output.
PR https://github.com/nushell/nushell/pull/12603 made it so that PWD can
never contain a trailing slash. However, a root path (such as `/` or
`C:\`) technically counts as "having a trailing slash", so now `cd /`
doesn't work.
I feel dumb for missing such an obvious edge case. Let's just merge this
quickly before anyone else finds out...
EDIT: It appears I'm too late.
# Description
Fixes#12758.
#12662 introduced a bug where calling `cd` with a path with a trailing
slash would cause `PWD` to be set to a path including a trailing slash,
which is not allowed. This adds a helper to `nu_path` to remove this,
and uses it in the `cd` command to clean it up before setting `PWD`.
# Tests + Formatting
I added some tests to make sure we don't regress on this in the future.
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
