# Description
The `into binary` command has a `-c` flag which strips any leading 0s in
the most significant digits to represent the minimal number of bytes,
rather than the system's complete in-memory representation of the input.
However, currently giving 0 as input results in eight 0 bytes even with
the `-c` flag, which is inconsistent with the purpose of the flag.
```nu
❯ : 345678 | into binary
Length: 8 (0x8) bytes | printable whitespace ascii_other non_ascii
00000000: 4e 46 05 00 00 00 00 00 NF•00000
❯ : 345678 | into binary -c
Length: 3 (0x3) bytes | printable whitespace ascii_other non_ascii
00000000: 4e 46 05
❯ : 0 | into binary
Length: 8 (0x8) bytes | printable whitespace ascii_other non_ascii
00000000: 00 00 00 00 00 00 00 00 00000000
❯ : 0 | into binary -c
Length: 8 (0x8) bytes | printable whitespace ascii_other non_ascii
00000000: 00 00 00 00 00 00 00 00 00000000
```
This change fixes this behavior so that if the entire input results in
all 0 bytes, only a single 0 byte is returned.
```nu
❯ : ~/src/nushell/target/aarch64-linux-android/debug/nu -c '0 | into binar
y -c'
Length: 1 (0x1) bytes | printable whitespace ascii_other non_ascii
00000000: 00
```
# User-Facing Changes
Values which result in all null bytes will be truncated to a single byte
when `-c` is given. This could potentially be considered a breaking
change if this behavior was relied upon in some way.
# Description
This PR adds the ability to parse human readable datetime strings as
part of the `into datetime` command. I added a new `-n`/`--list-human`
parameter that produces this list to give the user an idea of what is
supported.
```nushell
❯ into datetime --list-human
╭#─┬parseable human datetime examples┬───result───╮
│0 │Today 18:30 │in 8 hours │
│1 │2022-11-07 13:25:30 │a year ago │
│2 │15:20 Friday │in 3 days │
│3 │This Friday 17:00 │in 3 days │
│4 │13:25, Next Tuesday │in a week │
│5 │Last Friday at 19:45 │3 days ago │
│6 │In 3 days │in 2 days │
│7 │In 2 hours │in 2 hours │
│8 │10 hours and 5 minutes ago │10 hours ago│
│9 │1 years ago │a year ago │
│10│A year ago │a year ago │
│11│A month ago │a month ago │
│12│A week ago │a week ago │
│13│A day ago │a day ago │
│14│An hour ago │an hour ago │
│15│A minute ago │a minute ago│
│16│A second ago │now │
│17│Now │now │
╰#─┴parseable human datetime examples┴───result───╯
```
Or with `$env.config.datetime_format.table` set.
```nushell
❯ into datetime --list-human
╭#─┬parseable human datetime examples┬──────result───────╮
│0 │Today 18:30 │11/14/23 06:30:00PM│
│1 │2022-11-07 13:25:30 │11/07/22 01:25:30PM│
│2 │15:20 Friday │11/17/23 03:20:00PM│
│3 │This Friday 17:00 │11/17/23 05:00:00PM│
│4 │13:25, Next Tuesday │11/21/23 01:25:00PM│
│5 │Last Friday at 19:45 │11/10/23 07:45:00PM│
│6 │In 3 days │11/17/23 10:12:54AM│
│7 │In 2 hours │11/14/23 12:12:54PM│
│8 │10 hours and 5 minutes ago │11/14/23 12:07:54AM│
│9 │1 years ago │11/13/22 10:12:54AM│
│10│A year ago │11/13/22 10:12:54AM│
│11│A month ago │10/15/23 11:12:54AM│
│12│A week ago │11/07/23 10:12:54AM│
│13│A day ago │11/13/23 10:12:54AM│
│14│An hour ago │11/14/23 09:12:54AM│
│15│A minute ago │11/14/23 10:11:54AM│
│16│A second ago │11/14/23 10:12:53AM│
│17│Now │11/14/23 10:12:54AM│
╰#─┴parseable human datetime examples┴──────result───────╯
```
# 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
Refactors the `flatten` command to remove a bunch of cloning. This was
down by passing ownership of the `Value` to `flat_value`, removing the
lifetime on `TableInside`, and using `Vec<Record>` in `FlattenedRows`
instead of a pair of `Vec` of columns and values.
For the quick benchmark below, it seems to be twice as fast now:
```nushell
let data = ls crates | where type == dir | each { ls $'($in.name)/**/*' }
timeit { for x in 0..1000 { $data | flatten } }
```
This took 550ms on v0.86.0 and only 230ms on this PR.
But considering that
```nushell
timeit { for x in 0..1000 { $data } }
```
takes 200ms on both versions, then the difference for `flatten` itself
is really 250ms vs 30ms -- 8x faster.
# Description
Generally elide a bunch of unnecessary clones. Both globally stopping to
clone the whole input data in a bunch of places where we need to read it
but also some minor places where we currently cloned.
As part of that, we can make the overwriting with `keep-all` and
`keep-last` inplace so the items don't need to be removed and repushed
to the record.
# Benchmarking
```nu
timeit { scope commands | transpose -r }
```
Before ~24 ms now just ~5 ms
# User-Facing Changes
This can change the order of apperance in the transposed record with
`--keep-last`/`--keep-all`. Now the
order is determined by the first appearance and not by the last
appearance in the ingoing columns.
This mirrors the behavior when not passed `keep-all` or `keep-last`.
# Tests + Formatting
Sadly the `transpose` command is so far undertested for more complex
operations.
# Description
This PR refactors `drop columns` and fixes issues #10902 and #6846.
Tables with "holes" are now handled consistently, although still
somewhat awkwardly. That is, the columns in the first row are used to
determine which columns to drop, meaning that the columns displayed all
the way to the right by `table` may not be the columns actually being
dropped. For example, `[{a: 1}, {b: 2}] | drop column` will drop column
`a` instead of `b`. Before, this would give a list of empty records.
# User-Facing Changes
`drop columns` can now take records as input.
# Description
Add an extension trait `IgnoreCaseExt` to nu_utils which adds some case
insensitivity helpers, and use them throughout nu to improve the
handling of case insensitivity. Proper case folding is done via unicase,
which is already a dependency via mime_guess from nu-command.
In actuality a lot of code still does `to_lowercase`, because unicase
only provides immediate comparison and doesn't expose a `to_folded_case`
yet. And since we do a lot of `contains`/`starts_with`/`ends_with`, it's
not sufficient to just have `eq_ignore_case`. But if we get access in
the future, this makes us ready to use it with a change in one place.
Plus, it's clearer what the purpose is at the call site to call
`to_folded_case` instead of `to_lowercase` if it's exclusively for the
purpose of case insensitive comparison, even if it just does
`to_lowercase` still.
# User-Facing Changes
- Some commands that were supposed to be case insensitive remained only
insensitive to ASCII case (a-z), and now are case insensitive w.r.t.
non-ASCII characters as well.
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
---------
Co-authored-by: Stefan Holderbach <sholderbach@users.noreply.github.com>
# Description
This is pretty complementary/orthogonal to @IanManske 's changes to
`Value` cellpath accessors in:
- #10925
- to a lesser extent #10926
## Steps
- Use `R.remove` in `Value.remove_data_at_cell_path`
- Pretty sound after #10875 (tests mentioned in commit message have been
removed by that)
- Update `did_you_mean` helper to use iterator
- Change `Value::columns` to return iterator
- This is not a place of honor
- Use `Record::get` in `Value::get_data_by_key`
# User-Facing Changes
None intentional, potential edge cases on duplicated columns could
change (considered undefined behavior)
# Tests + Formatting
(-)
# Description
Based of the work and discussion in #10844, this PR adds the `exec`
command for Windows. This is done by simply spawning a
`std::process::Command` and then immediately exiting via
`std::process::exit` once the child process is finished. The child
process's exit code is passed to `exit`.
# User-Facing Changes
The `exec` command is now available on Windows, and there should be no
change in behaviour for Unix systems.
# Description
Where appropriate, this PR replaces instances of
`Value::get_data_by_key` and `Value::follow_cell_path` with
`Record::get`. This avoids some unnecessary clones and simplifies the
code in some places.
# Description
Since #10841 the goal is to remove the implementation details of
`Record` outside of core operations.
To this end use Record iterators and map-like accessors in a bunch of
places. In this PR I try to collect the boring cases where I don't
expect any dramatic performance impacts or don't have doubts about the
correctness afterwards
- Use checked record construction in `nu_plugin_example`
- Use `Record::into_iter` in `columns`
- Use `Record` iterators in `headers` cmd
- Use explicit record iterators in `split-by`
- Use `Record::into_iter` in variable completions
- Use `Record::values` iterator in `into sqlite`
- Use `Record::iter_mut` for-loop in `default`
- Change `nu_engine::nonexistent_column` to use iterator
- Use `Record::columns` iter in `nu-cmd-base`
- Use `Record::get_index` in `nu-command/network/http`
- Use `Record.insert()` in `merge`
- Refactor `move` to use encapsulated record API
- Use `Record.insert()` in `explore`
- Use proper `Record` API in `explore`
- Remove defensiveness around record in `explore`
- Use encapsulated record API in more `nu-command`s
# User-Facing Changes
None intentional
# Tests + Formatting
(-)
# Description
- Simplify `table` record highlight with `.get_mut`
- pretty straight forward
- Use record iterators in `table` abbreviation logic
- This required some rework if we go from guaranted contiguous arrays to
iterators
- Refactor `nu-table` internals to new record API
# User-Facing Changes
None intened
# Tests + Formatting
(-)
# Description
Rewrite `find` internals with the same principles as in #10927.
Here we can remove an unnecessary lookup accross all columns when not
narrowing find to particular columns
- Change `find` internal fns to use iterators
- Remove unnecessary quadratic lookup in `find`
- Refactor `find` record highlight logic
# User-Facing Changes
Should provide a small speedup when not providing `find --columns`
# Tests + Formatting
(-)
# Description
Changes the `captures` field in `Closure` from a `HashMap` to a `Vec`
and makes `Stack::captures_to_stack` take an owned `Vec` instead of a
borrowed `HashMap`.
This eliminates the conversion to a `Vec` inside `captures_to_stack` and
makes it possible to avoid clones altogether when using an owned
`Closure` (which is the case for most commands). Additionally, using a
`Vec` reduces the size of `Value` by 8 bytes (down to 72).
# User-Facing Changes
Breaking API change for `nu-protocol`.
# Description
This is easy to do with rust-analyzer, but I didn't want to just pump
these all out without feedback.
Part of #10700
# User-Facing Changes
None
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
# After Submitting
N/A
---------
Co-authored-by: Stefan Holderbach <sholderbach@users.noreply.github.com>
# Description
`split-by` only works on a `Record`, the error type was updated to
match, and now uses a more-specific type. (Two type fixes for the price
of one!)
The `usage` was updated to say "record" as well
# User-Facing Changes
* Providing the wrong type to `split-by` now gives an error messages
with the correct required input type
Previously:
```
❯ ls | get name | split-by type
Error: × unsupported input
╭─[entry #267:1:1]
1 │ ls | get name | split-by type
· ─┬─
· ╰── requires a table with one row for splitting
╰────
```
With this PR:
```
❯ ls | get name | split-by type
Error: nu:🐚:type_mismatch
× Type mismatch.
╭─[entry #1:1:1]
1 │ ls | get name | split-by type
· ─┬─
· ╰── requires a record to split
╰────
```
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
# After Submitting
Only generated commands need to be updated
---------
Co-authored-by: Darren Schroeder <343840+fdncred@users.noreply.github.com>
# Description
Limit the test `-p nu-command --test main
commands::run_external::redirect_combine` which uses `sh` to running on
`not(Windows)` like is done for other tests assuming unixy CLI items;
`sh` doesn't exist on Windows.
# User-Facing Changes
None; this is a change to tests only.
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- 🟢 `toolkit test`
- 🟢 `toolkit test stdlib`
# Description
@jntrnr discovered that `items` wasn't properly setting the
`eval_block_with_early_return()` block settings. This change fixes that
which allows `echo` to be redirected and therefore pass data through the
pipeline.
Without `echo`
```nushell
❯ { new: york, san: francisco } | items {|key, value| $'($key) ($value)' }
╭─┬─────────────╮
│0│new york │
│1│san francisco│
╰─┴─────────────╯
```
With `echo`
```nushell
❯ { new: york, san: francisco } | items {|key, value| echo $'($key) ($value)' }
╭─┬─────────────╮
│0│new york │
│1│san francisco│
╰─┴─────────────╯
```
# 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
This PR updates the `items` example so that it doesn't use `echo`.
`echo` now works like print unless it's being redirected, so it doesn't
send values through the pipeline anymore like the example showed.
# 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.
-->
- Replaced `start`/`end` with span.
- Fixed standard library.
- Add `help` option.
- Add a couple more errors for invalid record types.
Resolve#10914
# Description
# User-Facing Changes
- **BREAKING CHANGE:** `error make` now takes in `span` instead of
`start`/`end`:
```Nushell
error make {
msg: "Message"
label: {
text: "Label text"
span: (metadata $var).span
}
}
```
- `error make` now has a `help` argument for custom error help.
# Description
After talking to @CAD97, I decided to change these unwraps to expects.
See the comments. The bigger question is, how did unwrap pass the CI?
# 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.
-->
Now the `input list` command, when nothing is selected, will return a
null instead of empty string or an empty list.
Resolves#10909.
# User-Facing Changes
`input list` now returns a `null` when nothing is selected.
# Description
Pretty much all operations/commands in Nushell assume that the column
names/keys in a record and thus also in a table (which consists of a
list of records) are unique.
Access through a string-like cell path should refer to a single column
or key/value pair and our output through `table` will only show the last
mention of a repeated column name.
```nu
[[a a]; [1 2]]
╭─#─┬─a─╮
│ 0 │ 2 │
╰───┴───╯
```
While the record parsing already either errors with the
`ShellError::ColumnDefinedTwice` or silently overwrites the first
occurence with the second occurence, the table literal syntax `[[header
columns]; [val1 val2]]` currently still allowed the creation of tables
(and internally records with more than one entry with the same name.
This is not only confusing, but also breaks some assumptions around how
we can efficiently perform operations or in the past lead to outright
bugs (e.g. #8431 fixed by #8446).
This PR proposes to make this an error.
After this change another hole which allowed the construction of records
with non-unique column names will be plugged.
## Parts
- Fix `SE::ColumnDefinedTwice` error code
- Remove previous tests permitting duplicate columns
- Deny duplicate column in table literal eval
- Deny duplicate column in const eval
- Deny duplicate column in `from nuon`
# User-Facing Changes
`[[a a]; [1 2]]` will now return an error:
```
Error: nu:🐚:column_defined_twice
× Record field or table column used twice
╭─[entry #2:1:1]
1 │ [[a a]; [1 2]]
· ┬ ┬
· │ ╰── field redefined here
· ╰── field first defined here
╰────
```
this may under rare circumstances block code from evaluating.
Furthermore this makes some NUON files invalid if they previously
contained tables with repeated column names.
# Tests + Formatting
Added tests for each of the different evaluation paths that materialize
tables.
# Description
This change allows `compact` to also compact things with empty strings,
empty lists, and empty records if the `--empty` switch is used. Let's
add a quality-of-life improvement here to just compact all this mess. If
this is a bad idea, please cite examples demonstrating why.
```
❯ [[name position]; [Francis Lead] [Igor TechLead] [Aya null]] | compact position
╭#┬─name──┬position╮
│0│Francis│Lead │
│1│Igor │TechLead│
╰─┴───────┴────────╯
❯ [[name position]; [Francis Lead] [Igor TechLead] [Aya ""]] | compact position --empty
╭#┬─name──┬position╮
│0│Francis│Lead │
│1│Igor │TechLead│
╰─┴───────┴────────╯
❯ [1, null, 2, "", 3, [], 4, {}, 5] | compact
╭─┬─────────────────╮
│0│ 1│
│1│ 2│
│2│ │
│3│ 3│
│4│[list 0 items] │
│5│ 4│
│6│{record 0 fields}│
│7│ 5│
╰─┴─────────────────╯
❯ [1, null, 2, "", 3, [], 4, {}, 5] | compact --empty
╭─┬─╮
│0│1│
│1│2│
│2│3│
│3│4│
│4│5│
╰─┴─╯
```
# 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
Changes `FromValue` to take owned `Value`s instead of borrowed `Value`s.
This eliminates some unnecessary clones (e.g., in `call_ext.rs`).
# User-Facing Changes
Breaking API change for `nu_protocol`.
# Description
Reuses the existing `Closure` type in `Value::Closure`. This will help
with the span refactoring for `Value`. Additionally, this allows us to
more easily box or unbox the `Closure` case should we chose to do so in
the future.
# User-Facing Changes
Breaking API change for `nu_protocol`.
# Description
This PR restores and old functionality that must of been broken with the
input_output_types() updating. It allows commands like this to work
again.
```nushell
open $nu.history-path |
get history.command_line |
split column ' ' cmd |
group-by cmd --to-table |
update items {|u| $u.items | length} |
sort-by items -r |
first 10 |
table -n 1
```
output
```
╭#─┬group─┬items╮
│1 │exit │ 3004│
│2 │ls │ 2591│
│3 │git │ 1678│
│4 │help │ 1549│
│5 │open │ 1374│
│6 │cd │ 1186│
│7 │cargo │ 944│
│8 │let │ 784│
│9 │source│ 755│
│10│z │ 486│
╰#─┴group─┴items╯
```
# 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
Previously `group-by` returned a record containing each group as a
column. This data layout is hard to work with for some tasks because you
have to further manipulate the result to do things like determine the
number of items in each group, or the number of groups. `transpose` will
turn the record returned by `group-by` into a table, but this is
expensive when `group-by` is run on a large input.
In a discussion with @fdncred [several
workarounds](https://github.com/nushell/nushell/discussions/10462) to
common tasks were discussed, but they seem unsatisfying in general.
Now when `group-by --to-table` is used a table is returned with the
columns "groups" and "items" making it easier to do things like count
the number of groups (`| length`) or count the number of items in each
group (`| each {|g| $g.items | length`)
# User-Facing Changes
* `group-by` returns a `table` with "group" and "items" columns instead
of a `record` with one column per group name
# Tests + Formatting
Tests for `group-by` were updated
# After Submitting
* No breaking changes were made. The new `--to-table` switch should be
added automatically to the [`group-by`
documentation](https://www.nushell.sh/commands/docs/group-by.html)
# Description
Use `record!` macro instead of defining two separate `vec!` for `cols`
and `vals` when appropriate.
This visually aligns the key with the value.
Further more you don't have to deal with the construction of `Record {
cols, vals }` so we can hide the implementation details in the future.
## State
Not covering all possible commands yet, also some tests/examples are
better expressed by creating cols and vals separately.
# User/Developer-Facing Changes
The examples and tests should read more natural. No relevant functional
change
# Bycatch
Where I noticed it I replaced usage of `Value` constructors with
`Span::test_data()` or `Span::unknown()` to the `Value::test_...`
constructors. This should make things more readable and also simplify
changes to the `Span` system in the future.
# Description
as we can see in the [documentation of
`str.to_lowercase`](https://doc.rust-lang.org/std/primitive.str.html#method.to_lowercase),
not only ASCII symbols have lower and upper variants.
- `str upcase` uses the correct method to convert the string
7ac5a01e2f/crates/nu-command/src/strings/str_/case/upcase.rs (L93)
- `str downcase` incorrectly converts only ASCII characters
7ac5a01e2f/crates/nu-command/src/strings/str_/case/downcase.rs (L124)
this PR uses `str.to_lower_case` instead of `str.to_ascii_lowercase` in
`str downcase`.
# User-Facing Changes
- upcase still works fine
```nushell
~ l> "ὀδυσσεύς" | str upcase
ὈΔΥΣΣΕΎΣ
```
- downcase now works
👉 before
```nushell
~ l> "ὈΔΥΣΣΕΎΣ" | str downcase
ὈΔΥΣΣΕΎΣ
```
👉 after
```nushell
~ l> "ὈΔΥΣΣΕΎΣ" | str downcase
ὀδυσσεύς
```
# Tests + Formatting
- 🟢 `toolkit fmt`
- 🟢 `toolkit clippy`
- ⚫ `toolkit test`
- ⚫ `toolkit test stdlib`
adds two tests
- `non_ascii_upcase`
- `non_ascii_downcase`
# After Submitting
# Description
looking at the [Wax documentation about
`wax::Walk.not`](https://docs.rs/wax/latest/wax/struct.Walk.html#examples),
especially
> therefore does not read directory trees from the file system when a
directory matches an [exhaustive glob
expression](https://docs.rs/wax/latest/wax/trait.Pattern.html#tymethod.is_exhaustive)
> **Important**
> in the following of this PR description, i talk about *pruning* and a
`--prune` option, but this has been changed to *exclusion* and
`--exclude` after a discussion with @fdncred.
this looks like a *pruning* operation to me, right? 😮
i wanted to make the `glob` option `--not` clearer about that, because
> -n, --not <List(String)> - Patterns to exclude from the results
from `help glob` is not very explicit about whether the search is pruned
when entering a directory matching a pattern in `--not` or just removing
it from the output 😕
## changelog
this PR proposes to rename the `glob --not` option to `glob --prune` and
make it's documentation more explicit 😋
## benchmarking
to support the *pruning* behaviour put forward above, i've run a
benchmark
1. define two closures to compare the behaviour between removing
patterns manually or using `--not`
```nushell
let where = {
[.*/\.local/.*, .*/documents/.*, .*/\.config/.*]
| reduce --fold (glob **) {|pat, acc| $acc | where $it !~ $pat}
| length
}
```
```nushell
let not = { glob ** --not [**/.local/**, **/documents/**, **/.config/**] | length }
```
2. run the two to make sure they give similar results
```nushell
> do $where
33424
```
```nushell
> do $not
33420
```
👌
3. measure the performance
```nushell
use std bench
```
```nushell
> bench --verbose --pretty --rounds 25 $not
44ms 52µs 285ns +/- 977µs 571ns
```
```nushell
> bench --verbose --pretty --rounds 5 $where
1sec 250ms 187µs 99ns +/- 8ms 538µs 57ns
```
👉 we can see that the results are (almost) the same but
`--not` is much faster, looks like pruning 😋
# User-Facing Changes
- `--not` will give a warning message but still work
- `--prune` will work just as `--not` without warning and with a more
explicit doc
- `--prune` and `--not` at the same time will give an error
# Tests + Formatting
this PR fixes the examples of `glob` using the `--not` option.
# After Submitting
prepare the removal PR and mention in release notes.
<!--
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.
-->
Implements `whoami` using the `whoami` command from uutils as backend.
This is a draft because it depends on
https://github.com/uutils/coreutils/pull/5310 and a new release of
uutils needs to be made (and the paths in `Cargo.toml` should be
updated). At this point, this is more of a proof of concept 😄
Additionally, this implements a (simple and naive) conversion from the
uutils `UResult` to the nushell `ShellError`, which should help with the
integration of other utils, too. I can split that off into a separate PR
if desired.
I put this command in the "platform" category. If it should go somewhere
else, let me know!
The tests will currently fail, because I've used a local path to uutils.
Once the PR on the uutils side is merged, I'll update it to a git path
so that it can be tested and runs on more machines than just mine.
# User-Facing Changes
<!-- List of all changes that impact the user experience here. This
helps us keep track of breaking changes. -->
New `whoami` command. This might break some users who expect the system
`whoami` command. However, the result of this new command should be very
close, just with a nicer help message, at least for Linux users. The
default `whoami` on Windows is quite different from this implementation:
https://learn.microsoft.com/en-us/windows-server/administration/windows-commands/whoami
# 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.
-->
---------
Co-authored-by: Darren Schroeder <343840+fdncred@users.noreply.github.com>
# Description
Currently the following command is broken:
```nushell
echo a o+e> 1.txt
```
It's because we don't redirect output of `echo` command. This pr is
trying to fix it.
# Description
- this PR should close#10819
# User-Facing Changes
Behaviour is similar to pre 0.86.0 behaviour of the cp command and
should as such not have a user-facing change, only compared to the
current version, were the option is readded.
# After Submitting
I guess the documentation will be automatically updated and as this
feature is no further highlighted, probably, no more work will be needed
here.
# Considerations
coreutils actually allows a third option:
```
pub enum UpdateMode {
// --update=`all`,
ReplaceAll,
// --update=`none`
ReplaceNone,
// --update=`older`
// -u
ReplaceIfOlder,
}
```
namely `ReplaceNone`, which I have not added. Also I think that
specifying `--update 'abc'` is non functional.
# Description
Fixes: #10830
The issue happened during lite-parsing, when we want to put a
`LiteElement` to a `LitePipeline`, we do nothing if relative redirection
target is empty.
So the command `echo aaa o> | ignore` will be interpreted to `echo aaa |
ignore`.
This pr is going to check and return an error if redirection target is
empty.
# User-Facing Changes
## Before
```
❯ echo aaa o> | ignore # nothing happened
```
## After
```nushell
❯ echo aaa o> | ignore
Error: nu::parser::parse_mismatch
× Parse mismatch during operation.
╭─[entry #1:1:1]
1 │ echo aaa o> | ignore
· ─┬
· ╰── expected redirection target
╰────
```
# Description
Support pattern matching against the `null` literal. Fixes#10799
### Before
```nushell
> match null { null => "success", _ => "failure" }
failure
```
### After
```nushell
> match null { null => "success", _ => "failure" }
success
```
# User-Facing Changes
<!-- List of all changes that impact the user experience here. This
helps us keep track of breaking changes. -->
Users can pattern match against a `null` literal.
# Description
`from tsv` and `from csv` both support a `--flexible` flag. This flag
can be used to "allow the number of fields in records to be variable".
Previously, a record's invariant that `rec.cols.len() == rec.vals.len()`
could be broken during parsing. This can cause runtime errors as in
#10693. Other commands, like `select` were also affected.
The inconsistencies are somewhat hard to see, as most nushell code
assumes an equal number of columns and values.
# Before
### Fewer values than columns
```nushell
> let record = (echo "one,two\n1" | from csv --flexible | first)
# There are two columns
> $record | columns | to nuon
[one, two]
# But only one value
> $record | values | to nuon
[1]
# And printing the record doesn't show the second column!
> $record | to nuon
{one: 1}
```
### More values than columns
```nushell
> let record = (echo "one,two\n1,2,3" | from csv --flexible | first)
# There are two columns
> $record | columns | to nuon
[one, two]
# But three values
> $record | values | to nuon
[1, 2, 3]
# And printing the record doesn't show the third value!
> $record | to nuon
{one: 1, two: 2}
```
# After
### Fewer values than columns
```nushell
> let record = (echo "one,two\n1" | from csv --flexible | first)
# There are two columns
> $record | columns | to nuon
[one, two]
# And a matching number of values
> $record | values | to nuon
[1, null]
# And printing the record works as expected
> $record | to nuon
{one: 1, two: null}
```
### More values than columns
```nushell
> let record = (echo "one,two\n1,2,3" | from csv --flexible | first)
# There are two columns
> $record | columns | to nuon
[one, two]
# And a matching number of values
> $record | values | to nuon
[1, 2]
# And printing the record works as expected
> $record | to nuon
{one: 1, two: 2}
```
# User-Facing Changes
Using the `--flexible` flag with `from csv` and `from tsv` will not
result in corrupted record state.
# 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.
-->
Reverts nushell/nushell#10812
This goes back to a version of `regex` and its dependencies that is
shared with a lot of our other dependencies. Before this we did not
duplicate big dependencies of `regex` that affect binary size and
compile time.
As there is no known bug or security problem we suffer from, we can wait
on receiving the performance improvements to `regex` with the rest of
our `regex` dependents.
r? @fdncred
Last one, I hope. At least short of completely redesigning `registry
query`'s interface. (Which I wouldn't implement without asking around
first.)
# Description
User-Facing Changes has the general overview. Inline comments provide a
lot of justification on specific choices. Most of the type conversions
should be reasonably noncontroversial, but expanding `REG_EXPAND_SZ`
needs some justification. First, an example of the behavior there:
```shell
> # release nushell:
> version | select version commit_hash | to md --pretty
| version | commit_hash |
| ------- | ---------------------------------------- |
| 0.85.0 | a6f62e05ae |
> registry query --hkcu Environment TEMP | get value
%USERPROFILE%\AppData\Local\Temp
> # with this patch:
> version | select version commit_hash | to md --pretty
| version | commit_hash |
| ------- | ---------------------------------------- |
| 0.86.1 | 0c5a4c991f |
> registry query --hkcu Environment TEMP | get value
C:\Users\CAD\AppData\Local\Temp
> # Microsoft CLI tooling behavior:
> ^pwsh -c `(Get-ItemProperty HKCU:\Environment).TEMP`
C:\Users\CAD\AppData\Local\Temp
> ^reg query HKCU\Environment /v TEMP
HKEY_CURRENT_USER\Environment
TEMP REG_EXPAND_SZ %USERPROFILE%\AppData\Local\Temp
```
As noted in the inline comments, I'm arguing that it makes more sense to
eagerly expand the %EnvironmentString% placeholders, as none of
Nushell's path functionality will interpret these placeholders. This
makes the behavior of `registry query` match the behavior of pwsh's
`Get-ItemProperty` registry access, and means that paths (the most
common use of `REG_EXPAND_SZ`) are actually usable.
This does *not* break nu_script's
[`update-path`](https://github.com/nushell/nu_scripts/blob/main/sourced/update-path.nu);
it will just be slightly inefficient as it will not find any
`%Placeholder%`s to manually expand anymore. But also, note that
`update-path` is currently *wrong*, as a path including
`%LocalAppData%Low` is perfectly valid and sometimes used (to go to
`Appdata\LocalLow`); expansion isn't done solely on a path segment
basis, as is implemented by `update-path`.
I believe that the type conversions implemented by this patch are
essentially always desired. But if we want to keep `registry query`
"pure", we could easily introduce a `registry get`[^get] which does the
more complete interpretation of registry types, and leave `registry
query` alone as doing the bare minimum. Or we could teach `path expand`
to do `ExpandEnvironmentStringsW`. But REG_EXPAND_SZ being the odd one
out of not getting its registry type semantics decoded by `registry
query` seems wrong.
[^get]: This is the potential redesign I alluded to at the top. One
potential change could be to make `registry get Environment` produce
`record<Path: string, TEMP: string, TMP: string>` instead of `registry
query`'s `table<name: string, value: string, type: string>`, the idea
being to make it feel as native as possible. We could even translate
between Nu's cell-path and registry paths -- cell paths with spaces do
actually work, if a bit awkwardly -- or even introduce lazy records so
the registry can be traversed with normal data manipulation ... but that
all seems a bit much.
# User-Facing Changes
- `registry query`'s produced `value` has changed. Specifically:
- ❗ Rows `where type == REG_EXPAND_SZ` now expand `%EnvironmentVarable%`
placeholders for you. For example, `registry query --hkcu Environment
TEMP | get value` returns `C:\Users\CAD\AppData\Local\Temp` instead of
`%USERPROFILE%\AppData\Local\Temp`.
- You can restore the old behavior and preserve the placeholders by
passing a new `--no-expand` switch.
- Rows `where type == REG_MULTI_SZ` now provide a `list<string>` value.
They previously had that same list, but `| str join "\n"`.
- Rows `where type == REG_DWORD_BIG_ENDIAN` now provide the correct
numeric value instead of a byte-swapped value.
- Rows `where type == REG_QWORD` now provide the correct numeric
value[^sign] instead of the value modulo 2<sup>32</sup>.
- Rows `where type == REG_LINK` now provide a string value of the link
target registry path instead of an internal debug string representation.
(This should never be visible, as links should be transparently
followed.)
- Rows `where type =~ RESOURCE` now provide a binary value instead of an
internal debug string representation.
[^sign]: Nu's `int` is a signed 64-bit integer. As such, values >=
2<sup>63</sup> will be reported as their negative two's compliment
value. This might sometimes be the correct interpretation -- the
registry does not distinguish between signed and unsigned integer values
-- but regedit and pwsh display all values as unsigned.
