mirror of
https://github.com/nushell/nushell.git
synced 2025-05-16 07:54:40 +02:00
90b7b89deb
12 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Piepmatz
|
4b0b4ddce1
|
Replaced IoError::new_with_additional_context calls that still had Span::unknown() (#15056)
<!-- if this PR closes one or more issues, you can automatically link the PR with them by using one of the [*linking keywords*](https://docs.github.com/en/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword), e.g. - this PR should close #xxxx - fixes #xxxx you can also mention related issues, PRs or discussions! --> # Description <!-- Thank you for improving Nushell. Please, check our [contributing guide](../CONTRIBUTING.md) and talk to the core team before making major changes. Description of your pull request goes here. **Provide examples and/or screenshots** if your changes affect the user experience. --> In #14968 I grepped the code for `IoError::new` calls with unknown spans, but I forgot to also grep for `IoError::new_with_additional_context`, so I missed some. Hopefullly this is the last P.S. to #14968. # User-Facing Changes <!-- List of all changes that impact the user experience here. This helps us keep track of breaking changes. --> N/A # 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 toolkit.nu; toolkit test stdlib"` 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 > ``` --> - 🟢 `toolkit fmt` - 🟢 `toolkit clippy` - 🟢 `toolkit test` - 🟢 `toolkit test stdlib` # 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. --> N/A |
||
|
Piepmatz
|
66bc0542e0
|
Refactor I/O Errors (#14927)
<!--
if this PR closes one or more issues, you can automatically link the PR
with
them by using one of the [*linking
keywords*](https://docs.github.com/en/issues/tracking-your-work-with-issues/linking-a-pull-request-to-an-issue#linking-a-pull-request-to-an-issue-using-a-keyword),
e.g.
- this PR should close #xxxx
- fixes #xxxx
you can also mention related issues, PRs or discussions!
-->
# Description
<!--
Thank you for improving Nushell. Please, check our [contributing
guide](../CONTRIBUTING.md) and talk to the core team before making major
changes.
Description of your pull request goes here. **Provide examples and/or
screenshots** if your changes affect the user experience.
-->
As mentioned in #10698, we have too many `ShellError` variants, with
some even overlapping in meaning. This PR simplifies and improves I/O
error handling by restructuring `ShellError` related to I/O issues.
Previously, `ShellError::IOError` only contained a message string,
making it convenient but overly generic. It was widely used without
providing spans (#4323).
This PR introduces a new `ShellError::Io` variant that consolidates
multiple I/O-related errors (except for `ShellError::NetworkFailure`,
which remains distinct for now). The new `ShellError::Io` variant
replaces the following:
- `FileNotFound`
- `FileNotFoundCustom`
- `IOInterrupted`
- `IOError`
- `IOErrorSpanned`
- `NotADirectory`
- `DirectoryNotFound`
- `MoveNotPossible`
- `CreateNotPossible`
- `ChangeAccessTimeNotPossible`
- `ChangeModifiedTimeNotPossible`
- `RemoveNotPossible`
- `ReadingFile`
## The `IoError`
`IoError` includes the following fields:
1. **`kind`**: Extends `std::io::ErrorKind` to specify the type of I/O
error without needing new `ShellError` variants. This aligns with the
approach used in `std::io::Error`. This adds a second dimension to error
reporting by combining the `kind` field with `ShellError` variants,
making it easier to describe errors in more detail. As proposed by
@kubouch in [#design-discussion on
Discord](https://discord.com/channels/601130461678272522/615329862395101194/1323699197165178930),
this helps reduce the number of `ShellError` variants. In the error
report, the `kind` field is displayed as the "source" of the error,
e.g., "I/O error," followed by the specific kind of I/O error.
2. **`span`**: A non-optional field to encourage providing spans for
better error reporting (#4323).
3. **`path`**: Optional `PathBuf` to give context about the file or
directory involved in the error (#7695). If provided, it’s shown as a
help entry in error reports.
4. **`additional_context`**: Allows adding custom messages when the
span, kind, and path are insufficient. This is rendered in the error
report at the labeled span.
5. **`location`**: Sometimes, I/O errors occur in the engine itself and
are not caused directly by user input. In such cases, if we don’t have a
span and must set it to `Span::unknown()`, we need another way to
reference the error. For this, the `location` field uses the new
`Location` struct, which records the Rust file and line number where the
error occurred. This ensures that we at least know the Rust code
location that failed, helping with debugging. To make this work, a new
`location!` macro was added, which retrieves `file!`, `line!`, and
`column!` values accurately. If `Location::new` is used directly, it
issues a warning to remind developers to use the macro instead, ensuring
consistent and correct usage.
### Constructor Behavior
`IoError` provides five constructor methods:
- `new` and `new_with_additional_context`: Used for errors caused by
user input and require a valid (non-unknown) span to ensure precise
error reporting.
- `new_internal` and `new_internal_with_path`: Used for internal errors
where a span is not available. These methods require additional context
and the `Location` struct to pinpoint the source of the error in the
engine code.
- `factory`: Returns a closure that maps an `std::io::Error` to an
`IoError`. This is useful for handling multiple I/O errors that share
the same span and path, streamlining error handling in such cases.
## New Report Look
This is simulation how the I/O errors look like (the `open crates` is
simulated to show how internal errors are referenced now):
## `Span::test_data()`
To enable better testing, `Span::test_data()` now returns a value
distinct from `Span::unknown()`. Both `Span::test_data()` and
`Span::unknown()` refer to invalid source code, but having a separate
value for test data helps identify issues during testing while keeping
spans unique.
## Cursed Sneaky Error Transfers
I removed the conversions between `std::io::Error` and `ShellError` as
they often removed important information and were used too broadly to
handle I/O errors. This also removed the problematic implementation
found here:
|
||
Jack Wright
|
d081e3386f
|
Make pipeline metadata available to plugins (#13495)
# Description Fixes an issue with pipeline metadata not being passed to plugins. |
||
Andy Gayton
|
7b82c6b482
|
feat: make ctrlc available to plugins (#13181)
# Description This PR adds a new method to `EngineInterface`: `register_ctrlc_handler` which takes a closure to run when the plugin's driving engine receives a ctrlc-signal. It also adds a mirror of the `signals` attribute from the main shell `EngineState`. This is an example of how a plugin which makes a long poll http request can end the request on ctrlc: https://github.com/cablehead/nu_plugin_http/blob/main/src/commands/request.rs#L68-L77 To facilitate the feature, a new attribute has been added to `EngineState`: `ctrlc_handlers`. This is a Vec of closures that will be run when the engine's process receives a ctrlc signal. When plugins are added to an `engine_state` during a `merge_delta`, the engine passes the ctrlc_handlers to the plugin's `.configure_ctrlc_handler` method, which gives the plugin a chance to register a handler that sends a ctrlc packet through the `PluginInterface`, if an instance of the plugin is currently running. On the plugin side: `EngineInterface` also has a ctrlc_handlers Vec of closures. Plugin calls can use `register_ctrlc_handler` to register a closure that will be called in the plugin process when the PluginInput::Ctrlc command is received. For future reference these are some alternate places that were investigated for tying the ctrlc trigger to transmitting a Ctrlc packet through the `PluginInterface`: - Directly from `src/signals.rs`: the handler there would need a reference to the Vec<Arc<RegisteredPlugins>>, which would require us to wrap the plugins in a Mutex, which we don't want to do. - have `PersistentPlugin.get_plugin` pass down the engine's CtrlcHandlers to .get and then to .spawn (if the plugin isn't already running). Once we have CtrlcHandlers in spawn, we can register a handler to write directly to PluginInterface. We don't want to double down on passing engine_state to spawn this way though, as it's unpredictable because it would depend on whether the plugin has already been spawned or not. - pass `ctrlc_handlers` to PersistentPlugin::new so it can store it on itself so it's available to spawn. - in `PersistentPlugin.spawn`, create a handler that sends to a clone of the GC event loop's tx. this has the same issues with regards to how to get CtrlcHandlers to the spawn method, and is more complicated than a handler that writes directly to PluginInterface # User-Facing Changes No breaking changes --------- Co-authored-by: Ian Manske <ian.manske@pm.me> |
||
Devyn Cairns
|
d618fd0527
|
Fix bad method links in docstrings (#13471)
# Description Seems like I developed a bit of a bad habit of trying to link ```rust /// [`.foo()`] ``` in docstrings, and this just doesn't work automatically; you have to do ```rust /// [`.foo()`](Self::foo) ``` if you want it to actually link. I think I found and replaced all of these. # User-Facing Changes Just docs. |
||
Ian Manske
|
399a7c8836
|
Add and use new Signals struct (#13314)
# Description This PR introduces a new `Signals` struct to replace our adhoc passing around of `ctrlc: Option<Arc<AtomicBool>>`. Doing so has a few benefits: - We can better enforce when/where resetting or triggering an interrupt is allowed. - Consolidates `nu_utils::ctrl_c::was_pressed` and other ad-hoc re-implementations into a single place: `Signals::check`. - This allows us to add other types of signals later if we want. E.g., exiting or suspension. - Similarly, we can more easily change the underlying implementation if we need to in the future. - Places that used to have a `ctrlc` of `None` now use `Signals::empty()`, so we can double check these usages for correctness in the future. |
||
|
Jack Wright
|
0d060aeae8
|
Use pipeline data for http post|put|patch|delete commands. (#13254)
# Description Provides the ability to use http commands as part of a pipeline. Additionally, this pull requests extends the pipeline metadata to add a content_type field. The content_type metadata field allows commands such as `to json` to set the metadata in the pipeline allowing the http commands to use it when making requests. This pull request also introduces the ability to directly stream http requests from streaming pipelines. One other small change is that Content-Type will always be set if it is passed in to the http commands, either indirectly or throw the content type flag. Previously it was not preserved with requests that were not of type json or form data. # User-Facing Changes * `http post`, `http put`, `http patch`, `http delete` can be used as part of a pipeline * `to text`, `to json`, `from json` all set the content_type metadata field and the http commands will utilize them when making requests. |
||
|
Devyn Cairns
|
c61075e20e
|
Add string/binary type color to ByteStream (#12897)
# Description
This PR allows byte streams to optionally be colored as being
specifically binary or string data, which guarantees that they'll be
converted to `Binary` or `String` appropriately on `into_value()`,
making them compatible with `Type` guarantees. This makes them
significantly more broadly usable for command input and output.
There is still an `Unknown` type for byte streams coming from external
commands, which uses the same behavior as we previously did where it's a
string if it's UTF-8.
A small number of commands were updated to take advantage of this, just
to prove the point. I will be adding more after this merges.
# User-Facing Changes
- New types in `describe`: `string (stream)`, `binary (stream)`
- These commands now return a stream if their input was a stream:
- `into binary`
- `into string`
- `bytes collect`
- `str join`
- `first` (binary)
- `last` (binary)
- `take` (binary)
- `skip` (binary)
- Streams that are explicitly binary colored will print as a streaming
hexdump
- example:
```nushell
1.. | each { into binary } | bytes collect
```
# Tests + Formatting
I've added some tests to cover it at a basic level, and it doesn't break
anything existing, but I do think more would be nice. Some of those will
come when I modify more commands to stream.
# After Submitting
There are a few things I'm not quite satisfied with:
- **String trimming behavior.** We automatically trim newlines from
streams from external commands, but I don't think we should do this with
internal commands. If I call a command that happens to turn my string
into a stream, I don't want the newline to suddenly disappear. I changed
this to specifically do it only on `Child` and `File`, but I don't know
if this is quite right, and maybe we should bring back the old flag for
`trim_end_newline`
- **Known binary always resulting in a hexdump.** It would be nice to
have a `print --raw`, so that we can put binary data on stdout
explicitly if we want to. This PR doesn't change how external commands
work though - they still dump straight to stdout.
Otherwise, here's the normal checklist:
- [ ] release notes
- [ ] docs update for plugin protocol changes (added `type` field)
---------
Co-authored-by: Ian Manske <ian.manske@pm.me>
|
||
|
Ian Manske
|
6fd854ed9f
|
Replace ExternalStream with new ByteStream type (#12774)
# 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>
|
||
|
Ian Manske
|
e879d4ecaf
|
ListStream touchup (#12524)
# 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. |
||
|
Devyn Cairns
|
ad6deadf24
|
Flush on every plugin Data message (#12728)
# Description This helps to ensure data produced on a stream is immediately available to the consumer of the stream. The BufWriter introduced for performance reasons in 0.93 exposed the behavior that data messages wouldn't make it to the other side until they filled the buffer in @cablehead's [`nu_plugin_from_sse`](https://github.com/cablehead/nu_plugin_from_sse). I had originally not flushed on every `Data` message because I figured that it isn't really critical that the other side sees those messages immediately, since they're not used for control and they are flushed when waiting for acknowledgement or when the buffer is too full anyway. Increasing the amount of data that can be sent with a single underlying write increases performance, but this interferes with some plugins that want to use streams in a more real-time way. In the future I would like to make this configurable, maybe even per-command, so that a command can decide what the priority is. But for now I think this is reasonable. In the worst case, this decreases performance by about 40%, when sending very small values (just numbers). But for larger values, this PR actually increases performance by about 20%, because I've increased the buffer size about 2x to 16,384 bytes. The previous value of 8,192 bytes was too small to fit a full buffer coming from an external command, so doubling it makes sense, and now a write of a buffer from an external command can be done in exactly one write call, which I think makes sense. I'm doing this at the same time because flushing each data message would make it very likely that each individual data message from an external stream would require exactly two writes rather than approximately one (amortized). Again, hopefully the tradeoff isn't too bad, and if it is I'll just make it configurable. # User-Facing Changes - Performance of plugin streams will be a bit different - Plugins that expect to send streams in real-time will work again # Tests + Formatting - 🟢 `toolkit fmt` - 🟢 `toolkit clippy` - 🟢 `toolkit test` - 🟢 `toolkit test stdlib` |
||
|
Devyn Cairns
|
0c4d5330ee
|
Split the plugin crate (#12563)
# Description This breaks `nu-plugin` up into four crates: - `nu-plugin-protocol`: just the type definitions for the protocol, no I/O. If someone wanted to wire up something more bare metal, maybe for async I/O, they could use this. - `nu-plugin-core`: the shared stuff between engine/plugin. Less stable interface. - `nu-plugin-engine`: everything required for the engine to talk to plugins. Less stable interface. - `nu-plugin`: everything required for the plugin to talk to the engine, what plugin developers use. Should be the most stable interface. No changes are made to the interface exposed by `nu-plugin` - it should all still be there. Re-exports from `nu-plugin-protocol` or `nu-plugin-core` are used as required. Plugins shouldn't ever have to use those crates directly. This should be somewhat faster to compile as `nu-plugin-engine` and `nu-plugin` can compile in parallel, and the engine doesn't need `nu-plugin` and plugins don't need `nu-plugin-engine` (except for test support), so that should reduce what needs to be compiled too. The only significant change here other than splitting stuff up was to break the `source` out of `PluginCustomValue` and create a new `PluginCustomValueWithSource` type that contains that instead. One bonus of that is we get rid of the option and it's now more type-safe, but it also means that the logic for that stuff (actually running the plugin for custom value ops) can live entirely within the `nu-plugin-engine` crate. # User-Facing Changes - New crates. - Added `local-socket` feature for `nu` to try to make it possible to compile without that support if needed. # Tests + Formatting - 🟢 `toolkit fmt` - 🟢 `toolkit clippy` - 🟢 `toolkit test` - 🟢 `toolkit test stdlib` |