@fdncred:
table --collapse lines up rows and stays within --width (#19424){a,b} added (#19422)rm never lets a glob match a hidden file unless asked (#19421)source bugs (#19415)math sum size and the other math reducers sum a table's column (#19414)into datetime --formats to list the formats it reads (#19413)%$cmd dynamic builtin dispatch (#19404)older-than and newer-than operators for datetimes (#19397)into datetime on RFC 3339 strings (#19394)from dotenv and to dotenv, and had load-env and with-env convert what they set (#19391)@kronberger-droid:
@passionworkeer:
@Huang-404-Q:
@Tyarel8:
@qodelicious:
@app/dependabot:
@fdncred:
table --collapse lines up rows and stays within --width (#19424){a,b} added (#19422)rm never lets a glob match a hidden file unless asked (#19421)source bugs (#19415)math sum size and the other math reducers sum a table's column (#19414)into datetime --formats to list the formats it reads (#19413)%$cmd dynamic builtin dispatch (#19404)older-than and newer-than operators for datetimes (#19397)into datetime on RFC 3339 strings (#19394)from dotenv and to dotenv, and had load-env and with-env convert what they set (#19391)@kronberger-droid:
@passionworkeer:
@Huang-404-Q:
@Tyarel8:
@qodelicious:
@app/dependabot:
@NotTheDr01ds:
@fdncred:
@passionworkeer:
@app/dependabot:
@kronberger-droid:
*RightEnd word moves under the active caret (#1292)Signal::CtrlC (#1290)PageUp/PageDown (#1281)EditCommand::CapitalizeWord (#1280)EventStatus for out-of-tree edit modes (#1279)@Kiddo4993:
@fdncred:
@kokhlo:
@ymcx:
@pickx:
@stuartcarnie:
@fdncred:
webpage and zstd-sys and ignored the other zstd advisories to clear the RustSec audit (#19359)try hanging under pipefail and kept an external's stderr through finally and o> (#19349)ansi opt command (#19340)@patatoish:
@mirsella:
@qodelicious:
@carlos-mercado:
@Tyarel8:
@passionworkeer:
@kronberger-droid:
@bercedev:
@akasakariko:
@armadavalor:
@app/dependabot:
@fdncred:
webpage and zstd-sys and ignored the other zstd advisories to clear the RustSec audit (#19359)try hanging under pipefail and kept an external's stderr through finally and o> (#19349)ansi opt command (#19340)@patatoish:
@mirsella:
@qodelicious:
@carlos-mercado:
@Tyarel8:
@passionworkeer:
@kronberger-droid:
@bercedev:
@akasakariko:
@armadavalor:
@app/dependabot:
@passionworkeer:
@fdncred:
@taylormhall:
@qodelicious:
@kronberger-droid:
@Don-Yin:
@Nacl-HCl:
@fdncred:
@devtechedge:
@eitsupi:
@ayax79:
@fdncred:
zip extraction advisories in cargo audit, since calamine never extracts (#19273)main doesn't accept (#19272)--mcp-host help text and example (#19268)@rajaryan0x1:
@Tyarel8:
@manoj-1407:
table --abbreviated crash when the row count is too large to allocate (#19296)last crash when the count is too large to allocate (#19295)window crash when the window size is too large to allocate (#19288)chunks crash when the chunk size is too large to allocate (#19286)roll panic on empty lists and records (#19285)@ymcx:
@kronberger-droid:
@qodelicious:
tutor echo intro (#19287)decode (#19284)main in errors about a script's arguments (#19282)@intertermux-code:
@app/dependabot:
@ayax79:
@fdncred:
zip extraction advisories in cargo audit, since calamine never extracts (#19273)main doesn't accept (#19272)--mcp-host help text and example (#19268)@rajaryan0x1:
@Tyarel8:
@manoj-1407:
table --abbreviated crash when the row count is too large to allocate (#19296)last crash when the count is too large to allocate (#19295)window crash when the window size is too large to allocate (#19288)chunks crash when the chunk size is too large to allocate (#19286)roll panic on empty lists and records (#19285)@ymcx:
@kronberger-droid:
@qodelicious:
tutor echo intro (#19287)decode (#19284)main in errors about a script's arguments (#19282)@intertermux-code:
@app/dependabot:
@philocalyst:
@kronberger-droid:
@ppergame:
@qodelicious:
@fdncred:
@app/dependabot:
@qodelicious:
commandline edit in the fuzzy history menu (#1289)@DavSanchez:
@0x4D5352:
@IanHollow:
@philocalyst:
@draconacht
@fdncred
ignored and wt_renamed, link libgit2 statically (#19233)recursion_limit so infinite closure recursion errors instead of crashing (#19229)table --expand hanging when a nested table is wider than its column (#19212)@mvanhorn
@hustcer
@bfren
@eugenesvk
@Kissaki
@skyfireitdiy
@sidney3
@itsF4LCON
@kronberger-droid
@ymcx
@SyedYousufFaizan
@app/dependabot
@draconacht
@fdncred
ignored and wt_renamed, link libgit2 statically (#19233)recursion_limit so infinite closure recursion errors instead of crashing (#19229)table --expand hanging when a nested table is wider than its column (#19212)@mvanhorn
@hustcer
@bfren
@eugenesvk
@Kissaki
@skyfireitdiy
@sidney3
@itsF4LCON
@kronberger-droid
@ymcx
@SyedYousufFaizan
@app/dependabot
@kronberger-droid
@easyinplay
@fdncred
.. and ..$rest anywhere in a match list pattern, like Rust (#19172)main like bash and zsh, and fix exec (#19136)@qodelicious
@ymcx
@kronberger-droid
@Adnan-Husayn
@Sha1rholder
@rvhelden
@weslem
@feiiiiii5
@cptpiepmatz
@akshitj11
@GhostCoder6969
@app/dependabot
@fdncred
.. and ..$rest anywhere in a match list pattern, like Rust (#19172)main like bash and zsh, and fix exec (#19136)@qodelicious
@ymcx
@kronberger-droid
@Adnan-Husayn
@Sha1rholder
@rvhelden
@weslem
@feiiiiii5
@cptpiepmatz
@akshitj11
@GhostCoder6969
@app/dependabot
@kronberger-droid
@cptpiepmatz
@Dark-Alex-17
@SatvikMishra08
@glcraft
@pickx
@roberttidball
@fdncred
@fdncred
@cptpiepmatz
@Sha1rholder
@fdncred
@iibabyy
Nu 0.116.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
Nu 0.116.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
tui label --title is now tui label --titlebar To put a label in the top bar, use tui label --titlebar "My app" instead of tui label --title "My app". The old switch still works but warns on first use and is scheduled for removal in 0.118.0.
--title Use --title <text> to name the border of tui table, tui tree, tui select, tui log, tui preview, tui search, or tui textbox. The title replaces the widget's default name, while row counts and status indicators remain visible. For example:
ls | tui table --title files | tui runThe table's border now reads files (12) rather than table (12) when it has 12 rows. For tui preview, the previewed filename follows the custom title, as in files (Cargo.toml); without a file to preview, only the custom title appears. For tui search, the title also becomes the default placeholder unless --placeholder is set. An empty --title is rejected.
tui border lines with --border or $env.config.tui.border_type Bordered TUI widgets now accept --border <name> to change their outline. This includes tui table, tui tree, tui select, tui log, tui preview, tui search, tui textbox, tui box, and the dropdowns opened by tui menu. The names match table --theme options, and the flag offers completion for valid border names.
ls | tui table --border rounded | tui run
$env.config.tui.border_type = "double"
ls | tui table | tui run$env.config.tui.border_type sets the default for all TUI borders, including menu dropdowns and dialog frames. An individual widget's --border overrides it. The default is single, which preserves the previous appearance. For example, rounded draws ╭─╮, double draws ╔═╗, and heavy draws ┏━┓; styles such as basic, dots, and markdown are available too.
none and default are not valid TUI border names. Use frameless for an outline with no visible lines: it still occupies one cell, and the widget's title remains visible. When a widget has its own --border, its widgets record includes the selected border name.
Slow and endless streams now update TUI widgets while tui run is open. A log displays each row as it arrives, and a progress bar advances as its input runs instead of starting at 100%:
1.. | each {|n| sleep 1sec; $n} | tui log | tui runThis also works for streams piped into widgets in a child list, such as tui split [(tail -f app.log | tui log) (ls | tui table)].
TUI builders now always output a tui value: piping one into a non-TUI command, such as 1.. | tui table | first, no longer returns the input rows. A TUI value saved with let reads its stream on the first run. If the stream finishes, later runs display the same rows; if it stops early, later runs display no rows. When a tui run closure, tui bind, button, or menu hook returns data, that data replaces the piped data for the rest of the run, without additional rows arriving from the stream.
Streams now keep around 100,000 of their newest rows, dropping older rows in batches of 12,500. Up to 12,499 extra rows can appear between drops, and tui log --max-lines can raise the limit further. Piping data into a child list whose widgets all specify --data now produces an error instead of silently ignoring the input.
After focusing a tui split with Tab, the arrow keys or h/j/k/l move its first divider in either direction. Each press moves it one cell on splits up to 100 cells wide or tall when there is room; on larger splits, a press may move it a few cells. This works for both the default side-by-side layout and --vertical, including short splits where resizing previously got stuck.
ls | tui split [(tui table) (tui preview)] | tui run
ls | tui split [(tui table) (tui preview)] --vertical | tui runDragging a divider with the mouse now places it as close to the pointer as the surrounding panes allow, including dividers between later panes in a split. Resizing preserves at least one cell for every pane, and fixed-height splits such as a row of buttons no longer respond to keyboard resizing.
An open completion menu no longer closes when the typed text stops matching.
It shows NO RECORDS FOUND and fills again once the text matches, as it did before 0.116.
Cancelling an @interactive picker with Esc now closes its menu right away, and Tab on the same line opens the picker again.
$nu.startup-time is now measured from the start of Nushell's main function, rather than from when the operating system created the process. It no longer includes the time spent loading the executable and its libraries, and exec nu starts a fresh measurement instead of carrying over the previous program's time. This applies across platforms.
Before startup finishes, $nu.startup-time now contains the elapsed startup time instead of -1ns. It is available in config files, -e commands, and the --mcp, --lsp, --dap, and --ide-check modes. The --ide-hover, --ide-complete, --ide-goto-def, and --ide-ast modes can now also access $nu, enabling type information for constants built from it and completion of $nu fields. Config files loaded with nu --mcp now see $nu.is-mcp as true.
With --log-level perf, the "before main" line is gone. The final REPL startup line is labeled "startup (main to first prompt)", and the --mcp and --lsp startup lines no longer include the time spent serving the rest of the session.
def --wrapped rest parameter were not called for arguments starting with -, so what --<Tab> only offered declared flags. (#19130)tui records For scripts that inspect TUI values, a titled widget now has a title field in its widgets record, and a top-bar label's slot is now titlebar instead of title. When nothing is being previewed, values.<preview id>.title is now an empty string rather than preview.
Thanks to all the contributors below for helping us solve issues, improve documentation, refactor code, and more! 🙏
| author | title | link | |
]]>@kronberger-droid:
@fdncred:
--title renamed a widget's border; tui label --title became --titlebar (#19102)tui split panes in both directions (#19101)--border and $env.config.tui.border_type picked border lines by table --theme name (#19100)@cptpiepmatz:
@hustcer:
@costajohnt:
@pickx:
@app/dependabot:
@kronberger-droid:
@fdncred:
--title renamed a widget's border; tui label --title became --titlebar (#19102)tui split panes in both directions (#19101)--border and $env.config.tui.border_type picked border lines by table --theme name (#19100)@cptpiepmatz:
@hustcer:
@costajohnt:
@pickx:
@app/dependabot:
@Mrakorez:
@cptpiepmatz:
@fdncred:
@Decodetalkers:
@Bahex:
@cptpiepmatz:
@hustcer:
@app/dependabot:
tui commands for building terminal interfaces, interactive debugging, more tutor lessons, and a faster shell.
"It is with profound grief and deep love that the family of Yash Thakur announces his passing on April 2, 2026."]]>
tui commands for building terminal interfaces, interactive debugging, more tutor lessons, and a faster shell.
"It is with profound grief and deep love that the family of Yash Thakur announces his passing on April 2, 2026."
The Nushell team is deeply saddened by Yash's passing.
We found out about his passing recently and wanted to acknowledge his dedication and contributions to our project, as well as the friendship he shared with the Nushell community. Our condolences go out to his friends and family.
Yash started working on Nushell at the young age of 19 in September 2023.
He quickly learned the project and, in a short time, became a valuable contributor to the inner workings of Reedline, the parser, completions, and other components. At the time of his passing, he was also helping develop a new parser for Nushell.
Please consider reading the beautiful obituary written for Yash to learn more about his life, passions, interests, and amazing spirit.
Nu 0.116.0 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
Pressing Tab is easy. Making everything behind it agree on how completions work took a bit more effort. Thanks to @philocalyst, custom completions now share one input and output model across parameter completers, command-wide completers, external completers, and menu sources. Completers can ask for the token, the completion location, or the whole buffer, and combine their suggestions with other completion sources.
Breaking change
Existing custom completers and menu sources may need updating. See the examples and migration table in unified custom completions.
There are plenty of smaller improvements around that big rework, too. Completers can now see which command you are completing, and external completers receive expanded aliases. Sourced menus refresh properly after partial completion, warnings wait until the menu closes, and you can opt into keeping completion and history menus open while editing. Less fighting the menu, more pressing Tab.
Nushell's interactive tutor has more lessons for you. Thanks to @nos1dot618, you can now learn about command-history shortcuts and $env, write your own custom commands and work with pipeline input, and get to grips with if, match, and where.
If you have been meaning to move beyond the commands you already know, this is a good excuse. Run tutor and try a lesson right there in your shell.
Ever wanted your script to have a little interface? Thanks to @fdncred, the new tui commands let you build interactive terminal interfaces straight from Nushell. Put together tables, textboxes, menus, and more, then get the user's selections and entered values back as a record.
You probably won't replace your entire Ratatui application this afternoon, but a picker or a small form no longer needs a separate Rust project. You can even open it as a floating dialog, and your shell comes back intact when you quit. Take a look at the commands and examples here, then head over to the original PR for more. Even this changelog cannot easily cover everything these commands can do.
That script is doing something strange again. Time to set a breakpoint. Thanks to @rvhelden, Nushell now has a Debug Adapter Protocol server, started with nu --dap, so you can debug your scripts from an editor that supports DAP.
Set breakpoints, step through pipelines, inspect variables, and evaluate watch expressions. You can even step backward through a recorded timeline when you realize you just stepped past the interesting bit. A lot of work behind a small flag. More details are here.
There is less waiting around this release, too. Typing and completing use in the REPL does less work, and large configs load faster, along with the plugin registry. gstat also finds the nearest tag faster in large repositories.
$nu.startup-time now measures all the way to the first prompt being ready, so you get a more honest account of where that time went. There is also faster parsing and less work when styling values, and the repeated source slowdown introduced in 0.115 is fixed. Thanks to everyone making the shell a little quicker, one less bit of unnecessary work at a time.
Custom completions now use one shared input and output model across parameter completers, command-wide completers, external completers, and menu sources. This makes completion behavior more consistent, but it means existing custom completers and menu sources may need to be updated.
Previously, different completion entry points received different arguments. Parameter completers received something like [context, offset], command-wide completers received [spans], external completers commonly received {|spans| ...}, and menu sources received {|buffer, position| ...}.
Completers now declare which completion inputs they want by naming their positional parameters. The recognized names are token, place, and buffer, and they can be declared in any order.
def complete-branch [token: record] {
git branch --format '%(refname:short)' | lines | where $it =~ $token.text
}
def complete-from-place [place: record] {
if $place.kind == "flag-value" and $place.flag == "profile" {
[debug release]
}
}
def complete-from-buffer [buffer: string] {
$buffer | split row " "
}
def complete-from-multiple [token: record, place: record, buffer: string] {
#...
}token describes the token at the cursor, with fields like text, kind, and span.place describes the resolved completion site and replacement range, including fields like cursor, target, kind, flag, index, command and shape when applicable.buffer is the exact command line from the beginning of the line through the cursor, including text before pipes, inside closures, and after separators.To demonstrate how they look we'll use the newly added commandline complete --input (explained further below) to see what a completer receives:
'git checkout mai' | commandline complete --input{
token: {
text: mai,
kind: value,
span: {
start: 13,
end: 16
}
},
place: {
cursor: 16,
target: {
start: 13,
end: 16
},
kind: positional,
index: 1,
shape: external-argument,
command: [
git,
checkout,
mai
]
},
buffer: "git checkout mai"
}Use buffer instead of calling commandline from inside a completer. buffer is always the line currently being completed; commandline can be empty in some editor states, such as after ;.
Menu sources use this same input model now. A source that used to look like this:
source: {|buffer, position|
$buffer | str substring ..<$position | split row " " | last
}should usually become either:
source: {|token| $token.text }or, if it really needs the whole line:
source: {|buffer| $buffer }External completers also need to migrate from the old spans-style input to the new named inputs. For example, a whole-line external completer should declare buffer or place:
$env.config.completions.external.completer = {|place|
carapace ($place.command | first) nushell ...$place.command | from json
}If a completer declares an unrecognized positional name, Nushell passes nothing for that parameter and writes a diagnostic to the completion log.
Custom completer output is also handled more consistently. A completer can now return null to decline and let another completion source answer, a string or list of strings, suggestion records, or an envelope record with completions, options, and fallback.
def complete-preset [token: record] {
{
completions: [
{value: main, description: "default branch"}
{value: release, description: "release branch"}
]
options: {
filter: true
completion_algorithm: "substring"
match_description: true
}
fallback: true
}
}fallback: true keeps the custom results and continues to the next completion source, so custom completions can appear beside built-in file completion or command-wide completions. Parameter completers are filtered by default; command-wide and external completers are not, because they usually do their own filtering. Set options.filter explicitly if you need the other behavior.
The old global background-completions option has been removed. Completers normally run on a background worker. If a completer needs to own the terminal, such as one that launches fzf or input list, mark the completer command with @interactive:
@interactive
def pick-file [token: record] {
ls | get name | to text | ^fzf --query $token.text | lines
}
def open-file [path: string@pick-file] {
open $path
}When completing open-file <Tab>, pick-file runs inline on the line-editor thread with stdin and the terminal available. This blocks the editor while the picker is open, then returns the selected values as completion suggestions. Calling an @interactive completer directly is not the intended test path, because the completion engine supplies its inputs.
External completers are closures and cannot carry @interactive themselves. To make an external completer interactive, have the closure call an @interactive command:
@interactive
def carapace-fzf [place: record] {
carapace ($place.command | first) nushell ...$place.command
| from json
| ^fzf
| lines
}
$env.config.completions.external.completer = {|place|
carapace-fzf $place
}commandline complete gained tools for developing completers:
commandline complete --input returns the {token, place, buffer} record a completer would receive without running the completer.commandline complete --detailed returns suggestions in the custom-completer output format.commandline complete --type directory|path|glob|command|variable|env-var runs one built-in completion source, which is useful when composing Nushell's built-in completions with custom ones.Note
--input cannot be combined with --detailed or --type.
Some completion behavior changed as part of this unification:
null to decline, allowing file completion or another source to answer.Common migrations:
| Previous form | New form | |:
]]>@fdncred:
@mrhard9090:
@kronberger-droid:
@mikehasa:
@windlandneko:
@philocalyst:
@LouisDeconinck:
@surma:
@Dorumin:
@rvhelden:
@app/dependabot:
@fdncred:
@mrhard9090:
@kronberger-droid:
@mikehasa:
@windlandneko:
@philocalyst:
@LouisDeconinck:
@surma:
@Dorumin:
@rvhelden:
@app/dependabot:
@shreeve:
@kronberger-droid:
@bhouse-nexthop:
@WindSoilder:
@fdncred:
@abrauninger:
@dannyfritz:
@kronberger-droid:
@LouisDeconinck:
@fdncred:
@KaiSforza:
@sebaschi:
@Developer1010x:
@Jorge-Polanco-Roque:
@Dheebz:
@neveroNiwe:
@app/dependabot:
@kronberger-droid:
@LouisDeconinck:
@fdncred:
@KaiSforza:
@sebaschi:
@Developer1010x:
@Jorge-Polanco-Roque:
@Dheebz:
@neveroNiwe:
@app/dependabot:
@kronberger-droid:
@shreeve:
@okuuva:
@catlover-bot:
@pimgit0:
@NotTheDr01ds:
@app/dependabot:
@ronisbr:
@fdncred:
@pickx:
@cptpiepmatz:
@zozo123:
@mkatychev:
@kronberger-droid:
@sidsri14:
@user93390:
@nos1dot618:
@philocalyst:
@app/dependabot:
@ronisbr:
@fdncred:
@pickx:
@cptpiepmatz:
@zozo123:
@mkatychev:
@kronberger-droid:
@sidsri14:
@user93390:
@nos1dot618:
@philocalyst:
@app/dependabot:
@kronberger-droid:
@Blfrg:
@Jiogo18:
@shreeve:
@catlover-bot:
@sholderbach:
@cuishuang:
@fdncred:
WORKSPACE_ROOT path (#18941)@luangucun:
@ayax79:
@cptpiepmatz:
@magnify035:
@nos1dot618:
@i-api:
@aron-intframe:
@Mrfiregem:
@catlover-bot:
@fly1d:
@dilr:
@bobisageek:
@Dheebz:
@hexbinoct:
@app/dependabot:
@cuishuang:
@fdncred:
WORKSPACE_ROOT path (#18941)@luangucun:
@ayax79:
@cptpiepmatz:
@magnify035:
@nos1dot618:
@i-api:
@aron-intframe:
@Mrfiregem:
@catlover-bot:
@fly1d:
@dilr:
@bobisageek:
@Dheebz:
@hexbinoct:
@app/dependabot:
@nos1dot618:
@NotTheDr01ds:
@app/dependabot:
@kronberger-droid:
@eitsupi:
@aron-intframe:
@fdncred:
semver | into string (#18901)fzf or input list (#18881)lines --skip-empty for string and byte stream input (#18872){ at end of format pattern and parse patterns (#18871)use std/ (#18841)@kronberger-droid:
background-completions opt-out (#18764)@ayax79:
@cptpiepmatz:
@philocalyst:
@pickx:
@aron-intframe:
@fdncred:
semver | into string (#18901)fzf or input list (#18881)lines --skip-empty for string and byte stream input (#18872){ at end of format pattern and parse patterns (#18871)use std/ (#18841)@kronberger-droid:
background-completions opt-out (#18764)@ayax79:
@cptpiepmatz:
@philocalyst:
@pickx:
@cptpiepmatz:
@catlover-bot:
@kronberger-droid:
@cptpiepmatz:
@NotTheDr01ds:
@Kissaki:
@dependabot:
$ans, SQLite table updates, and keybinding merging.
Nu 0.115.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
$ans, SQLite table updates, and keybinding merging.
Nu 0.115.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
HelixChangeMode keybinding support Keybindings can now switch helix modes with { send: HelixChangeMode, mode: normal }, taking normal, insert or select.
An until list covers vi and helix in one binding, since a mode event is inapplicable in the other editor's state machine:
$env.config.keybindings ++= [
{
name: leave_insert_mode
modifier: control
keycode: char_g
mode: [vi_insert helix_insert]
event: { until: [
{ send: ViChangeMode, mode: normal }
{ send: HelixChangeMode, mode: normal }
] }
}
]background-completions opt-out Added the background-completions experimental option (enabled by default). Disabling it makes completions block the line editor again, which lets custom completers drive interactive child processes like fzf:
nu --experimental-options '[background-completions=false]'Tab completion can run an interactive picker from $env.config.completions.external.completer (and from custom @comp / @complete closures). Closures such as fzf or input list take the TTY while the line editor waits. Completers that only print a list should still return quickly.
After input list (or a similar picker) returns from a completion or a custom menu, the prompt keeps raw mode. Backspace and other editing keys work again.
Subprocesses spawned from a background completion thread stay detached from the terminal even when their stdin is a pipe, so they cannot steal the TTY from the line editor.
When $env.config.completions.external.completer is set, Tab on an external command waits for that closure before the line editor continues. That lets a picker such as fzf or input list take the terminal. Completers that only print a list (for example Carapace) should return quickly; a slow closure freezes the prompt until it finishes.
File and command-name completion still run in the background. Custom menu source closures were already synchronous.
To make every Tab wait, not only user completers:
nu --experimental-options '[background-completions=false]'Results from the external completer and from @comp / @complete are not cached, so a picker runs again on the next Tab.
Comparing two semver values already returned a bool at runtime, but the parser treated that result as a semver. Combining it with or / and, binding it with let, or passing it to a bool parameter failed.
Those now work:
# missing or out of date
(which crush | is-empty) or ((crush --version | into semver) < ('1.0.0' | into semver))
let outdated = ('0.9.0' | into semver) < ('1.0.0' | into semver)
# $outdated is bool
true or (('1.2.3' | into semver) in ('>=1.0.0' | into semver-range))The same applies to other custom values whose comparisons return a bool (for example matrix). Custom types that return another custom value from < / == (polars expressions) are unchanged.
A raw custom value is still not a bool: true or ('1.0.0' | into semver) remains a type error.
$ans after invalid operations on $ans.last $ans stays usable after a type error on $ans.last
Invalid operations on $ans.last that embed type errors as values (for example $ans.last | str length on a table) no longer replace $ans.last or break later $ans display.
This mainly showed up with table --expand as the display_output hook.
$env.config.max_last_result_size = 1mb
ls
# table of files
$ans.last | str length
# Error: nu::shell::only_supports_this_input_type
# (still printed — the command is invalid)
$ans
# record with the original ls table in `.last` — no longer reprints the type error
$ans.last
# still the ls table$ans.exit_code, $ans.duration, and $ans.command still update for the failed line. Results that mix real values and errors still replace $ans.last. Top-level table --expand on a list of errors still shows the original diagnostic.
update closures on lazy SQLite tables update with a closure now works on a table taken from an opened SQLite database. open foo.db | get my_table is a lazy query; it already printed as a table, but update col { ... } reported a missing column instead of transforming each row.
[{ key: test, created: 1787132290545485211, expiry: 1787139010545500670, value: [7, 2, 128] }]
| into sqlite --table-name std_cache_store test.db
open test.db | get std_cache_store | update expiry { into datetime }
# table with `expiry` as datetimeopen sample.db | get ints | update z { default 0 | $in + 1 }
# each `z` value is incremented$env.config.keybindings merging for shared names Fixed an issue where two keybindings sharing a name on different keys silently dropped one of them (broke atuin's ctrl-r).
Nushell now warns once when keybindings share a name, matches mode spellings like emacs and [emacs] when merging, and $env.config.keybindings = [] clears the list again.
cd and custom completers Fixed completions for empty directory-typed arguments such as cd <tab> so they suggest directories instead of all paths.
Custom completers that error, return no result, or try to run interactive input commands now gracefully fall back to file completions instead of leaving the completion list empty or racing the line editor for terminal input.
commandline complete now includes a type: null field for completion items that do not have a more specific type.
Thanks to all the contributors below for helping us solve issues, improve documentation, refactor code, and more! 🙏
| author | change | link | |
]]>@fdncred:
@cptpiepmatz:
@kronberger-droid:
@brandondong:
@app/dependabot:
@fdncred:
@cptpiepmatz:
@kronberger-droid:
@brandondong:
@app/dependabot:
@maxim-uvarov:
@weirdan:
@cptpiepmatz:
@NotTheDr01ds:
@hustcer:
@cptpiepmatz:
@kronberger-droid:
$ans REPL variable for inspecting your last result, Helix-style editing, and a lot of internal cleanup and fixes.
Nu 0.115.0 is available as pre-built binaries or from crates.io. If you have Rust installed, you can install it using cargo install nu.
$ans REPL variable for inspecting your last result, Helix-style editing, and a lot of internal cleanup and fixes.
Nu 0.115.0 is available as pre-built binaries or from crates.io. If you have Rust installed, you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
Thanks to @cptpiepmatz, Nushell's YAML support got a proper rebuild. We now use a stable, modern, maintained implementation that no longer tries to make YAML 1.1, the scary one, and YAML 1.2 behave like one big YAML-shaped compromise.
The two specs are now handled separately, from yaml defaults to YAML 1.2, tags work more deliberately, multi-document streams are supported, anchors and merge keys behave properly, and to yaml is clearer about values that cannot round-trip.
Take a look at all the examples here.
$ans.last pipeline took so long to run Ever had your last pipeline take forever, only to realize you forgot to assign the result to a variable? Worry no longer. Thanks to @fdncred, Nushell now has $ans, a Nushell-specific answer variable heavily inspired by calculators.
$ans keeps track of the last REPL result, including the output, duration, exit code, and the command you typed. Storing the output is opt-in: set $env.config.max_last_result_size to a filesize of your choosing, and $ans.last will keep up to that much of the previous pipeline result.
That means if your pipeline took 15 minutes and you do not want to run it again, you can finally breathe, do a quick let took_too_long = $ans.last, and you're golden. Take a closer look here.
Alongside Vim mode, Nushell now has helix edit mode too. Set it with $env.config.edit_mode = helix, and that's it. You get a selection-first, Helix/Kakoune-style editing experience with normal, select, and insert modes.
Thanks to @kronberger-droid for the work here. More details are here.
This release does not have a huge pile of public user-facing additions, but there was a lot of work inside Nushell. Be in awe of the long list of things happening in the ✨ Hall of Fame ✨, and also check out the long list of fixes.
As always, a big shoutout to everyone contributing and making Nushell better. The internals may not always get the flashiest highlight section, but they are doing a lot of heavy lifting this time.
YAML has been a problem child for quite some time now. This release replaces the old serde_yaml-based implementation with a new one built on serde-saphyr and granit-parser. With that, we now get clearer YAML behavior, better round-tripping, proper tag support, multi-document streams, anchors, aliases, merge keys, and a few new options for configuring how the output should look like.
The big breaking change is that Nu no longer tries to parse YAML as an awkward mix of YAML 1.1 and YAML 1.2. We now default to YAML 1.2, which is usually the less crazy choice. If your files or scripts relied on old YAML 1.1 scalar magic, pass --spec 1.1.
use std/assert
# Default YAML 1.2 behavior: these stay strings.
assert equal ("yes" | from yaml) "yes"
assert equal ("off" | from yaml --spec 1.2) "off"
# YAML 1.1 keeps the classic YAML oddities.
assert equal ("yes" | from yaml --spec 1.1) true
assert equal ("off" | from yaml --spec 1.1) falseThis specifically breaks octal numbers. In the 1.1 spec, a value is interpreted as an octal number if it starts with the 0 digit. It is then parsed as an octal value. If the value is not valid octal, it is instead returned as a string.
The 1.2 spec made this much more sensible by allowing leading zeros on regular numbers and requiring octal numbers to use the 0o prefix.
use std/assert
# YAML 1.1 uses a leading zero as an octal indicator.
assert equal ("0247" | from yaml --spec 1.1) 0o247
assert equal ("0o247" | from yaml --spec 1.1) "0o247"
# YAML 1.2 replaced that with the `0o` prefix.
assert equal ("0247" | from yaml --spec 1.2) 0247
assert equal ("0o247" | from yaml --spec 1.2) 0o247Did you know that YAML 1.1 supports sexagesimal values like 190:20:30? No, well, in the 1.1 spec, these are base-60 numbers that can be used to represent time, but no one knows that, so 1.2 got rid of them. And as Nushell now defaults to 1.2, we interpret these as strings.
use std/assert
assert equal ("190:20:30" | from yaml) "190:20:30"
assert equal ("190:20:30" | from yaml --spec 1.1) 685230
assert equal ("02472256" | from yaml) 2472256
assert equal ("02472256" | from yaml --spec 1.1) 685230from yaml is stricter about mapping keys now, too. Nushell record keys are strings, so plain YAML keys that resolve to booleans, numbers, or null are rejected by default. If you want the old loose behavior, use --key-resolution verbatim and Nushell will keep the original key text.
'true: enabled' | from yaml
# Error: YAML key resolves to a boolean, but Nushell record keys are strings
'true: enabled' | from yaml --key-resolution verbatim
# => {true: enabled}Tags also mean tags now. Previously, tagged scalars weren't really handled properly; often they were just ignored and handled as plain strings, but in YAML they describe the data, usually as some types. Now unknown tags error by default, but you can use --ignore-tags to just ignore the tags and deal with them yourself.
'Key: !Sub ${AWS::StackName}' | from yaml
# Error: unknown YAML tag
'Key: !Sub ${AWS::StackName}' | from yaml --ignore-tags
# => {Key: "${AWS::StackName}"}to yaml is also more specific about values that cannot round-trip. By default it errors instead of quietly turning them into something else. If that is unwanted, choose the behavior explicitly with --non-roundtrip null or --non-roundtrip lossy. The --serialize flag is still available but will probably be deprecated in the future.
{|| $in } | to yaml
# Error: closures are not round-trippable through YAML
{|| $in } | to yaml --non-roundtrip "null"
# => null
{|| $in } | to yaml --serialize
# => !closure "{|| $in }"One more relevant breaking change: generated YAML may not look exactly like it did before. The new serializer quotes strings more carefully, writes tags for Nushell-specific values, and lets you configure indentation. If you depended on a specific YAML output, you might need to check that.
Nushell now understands standard YAML tags like !!timestamp,!!binary, !!omap, !!pairs, and !!set. It also writes Nushell-specific local tags for values that YAML did not define globally, such as filesizes, durations, ranges, globs, and cell paths.
name: Cargo.toml
size: !filesize 17515
modified: !!timestamp 2026-08-14T23:30:19.144306+02:00
path: !cell-path $.items.0.nameHere, !filesize and !cell-path are Nushell tags, while !!timestamp is a standard YAML tag. When you read this back with from yaml, those fields come back as properly typed values instead of plain strings or integers.
If you want the YAML version and Nushell tag prefix written out explicitly, use to yaml --add-directives or to yaml -d.
ls | where name == Cargo.toml | first | to yaml --add-directives%YAML 1.2
%TAG ! tag:nushell.sh,2026:@kronberger-droid:
@fdncred:
cli key to $ans record, renamed last_result_size to max_last_result_size (#18827)$ans work better with TUIs (#18820)--loose flag (#18819)--regex to use LRU (#18797)# to be passed (#18782)kdl to support --spec 1 or --spec 2 (#18779)$ans record for storing the last result from the repl (#18729)@Bahex:
@cpea2506:
@Mrfiregem:
@hexbinoct:
@latent-9:
@app/dependabot:
@kronberger-droid:
@fdncred:
cli key to $ans record, renamed last_result_size to max_last_result_size (#18827)$ans work better with TUIs (#18820)--loose flag (#18819)--regex to use LRU (#18797)# to be passed (#18782)kdl to support --spec 1 or --spec 2 (#18779)$ans record for storing the last result from the repl (#18729)@Bahex:
@cpea2506:
@Mrfiregem:
@hexbinoct:
@latent-9:
@app/dependabot:
@kronberger-droid:
@james-kassabian:
@orthonormalremy:
@idzikovsky:
@app/dependabot:
@Bahex:
@cacdu:
@cptpiepmatz:
@fdncred:
@ian-h-chamberlain:
@oh-summy:
@pheenty:
@philocalyst:
@Totara-thib:
@app/dependabot:
@Bahex:
@cacdu:
@cptpiepmatz:
@fdncred:
@ian-h-chamberlain:
@oh-summy:
@pheenty:
@philocalyst:
@Totara-thib:
@app/dependabot:
@kronberger-droid:
@matystroia:
@sim590:
@cptpiepmatz:
test() (#18748)open tests (#18740)kitest (#18727)--testbin from nu (#18719)run external tests (#18715)nu --testbin (#18713)@Bahex:
@fdncred:
doc_config.nu and tweaked the explore section (#18750)$env.config show default configuration when not set even with nu -n (#18747)ls ** when using the dc-glob experimental option (#18724)@xtqqczze:
@danielcadev:
@ayax79:
@ZayanKhan-12:
@rvhelden:
@cptpiepmatz:
test() (#18748)open tests (#18740)kitest (#18727)--testbin from nu (#18719)run external tests (#18715)nu --testbin (#18713)@Bahex:
@fdncred:
doc_config.nu and tweaked the explore section (#18750)$env.config show default configuration when not set even with nu -n (#18747)ls ** when using the dc-glob experimental option (#18724)@xtqqczze:
@danielcadev:
@ayax79:
@ZayanKhan-12:
@rvhelden:
@cptpiepmatz:
@NotTheDr01ds:
@ColumbusLabs:
@app/dependabot:
@fdncred:
is-terminal OS detection and added is-redirected for pipeline destinations (#18661)idx watch command (#18639)ls with dc-glob would fail with literal * name files (#18632)--config-home parameter also ensured xdg is absolute (#18629)export main from modules with keyword names (#18619)@pyz4:
@kronberger-droid:
@Alb-O:
@philocalyst:
@ahmojo:
@cptpiepmatz:
@Bahex:
@Tyarel8:
@cablehead:
@app/dependabot:
@fdncred:
is-terminal OS detection and added is-redirected for pipeline destinations (#18661)idx watch command (#18639)ls with dc-glob would fail with literal * name files (#18632)--config-home parameter also ensured xdg is absolute (#18629)export main from modules with keyword names (#18619)@pyz4:
@kronberger-droid:
@Alb-O:
@philocalyst:
@ahmojo:
@cptpiepmatz:
@Bahex:
@Tyarel8:
@cablehead:
@app/dependabot:
@fdncred:
@ninjack-dev:
@HigherOrderLogic:
@Mrfiregem:
@fdncred:
@kronberger-droid:
@casedami:
@philocalyst:
@easyinplay:
@eitsupi:
@nextgenthemes:
@AucaCoyan:
@fdncred:
ln command" (#18605)run from trying to parse big non-text files (#18603)stor and query db more thread-safe (#18569)mut and par-each optimizations for performance (#18566)matrix commands (#18553)@pyz4:
@pheenty:
@Alb-O:
@coyaSONG:
@sid-6581:
@Bahex:
@kobihikri:
@m-novotny:
@drbrain:
@cptpiepmatz:
@alerque:
@dmatos2012:
@Mrfiregem:
@app/dependabot:
@fdncred:
ln command" (#18605)run from trying to parse big non-text files (#18603)stor and query db more thread-safe (#18569)mut and par-each optimizations for performance (#18566)matrix commands (#18553)@pyz4:
@pheenty:
@Alb-O:
@coyaSONG:
@sid-6581:
@Bahex:
@kobihikri:
@m-novotny:
@drbrain:
@cptpiepmatz:
@alerque:
@dmatos2012:
@Mrfiregem:
@app/dependabot:
@Kissaki:
@cptpiepmatz:
@fdncred:
@KP64:
@maxim-uvarov:
@philocalyst:
@Tyarel8:
@alightgoesout:
Today, we're releasing version 0.114.1 of Nu. This patch release fixes issues found after enforce-runtime-annotations became opt-out in the last version, improving runtime type checks and error reporting.
Today, we're releasing version 0.114.1 of Nu. This patch release fixes issues found after enforce-runtime-annotations became opt-out in the last version, improving runtime type checks and error reporting.
Nu 0.114.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
In Nushell 0.114.0, we promoted enforce-runtime-annotations from an opt-in experimental option to an opt-out experimental option. That is the usual path for experimental options: more people try them, we get better feedback, and the feature gets closer to being stable.
As expected, that wider testing found some rough edges. Some came from invalid signatures, some from type checks that were too strict, and some from checks that were simply not accurate enough. We are tracking those reports in issue #18528, and this patch release fixes a number of them, mostly thanks to @Bahex. Thank you!
More issues may still turn up. Please report them if you run into anything suspicious. If enforce-runtime-annotations blocks your workflow, you can disable just that option before Nushell starts:
NU_EXPERIMENTAL_OPTIONS=enforce-runtime-annotations=false nuOr pass it as a command-line flag:
nu --experimental-options enforce-runtime-annotations=falseSee the nu-experimental user documentation for more details.
We still encourage everyone to keep as many experimental options enabled as possible. It helps us find bugs earlier and make Nushell better for future releases.
enforce-runtime-annotations With enforce-runtime-annotations enabled, assignments with incorrect types raise errors.
Previously this error was nu::shell::cant_convert originally used for unsuccessful type coercions of command arguments. Unfortunately, this error only pointed at the assigned values origin and not at the failed assignment.
This error has been switched to nu::shell::type_mismatch and now points to the failed assignment in addition to the value's origin.
| Code |
|
| Before |
|
| After |
|
| Code |
|
| Before |
|
| After |
|
With last release, pipeline input type and how it affects various expressions are being tracked in a lot more cases.
In fact, at least one too many:
def mk-add-suffix [suffix: string]: nothing -> closure {
{|| $in ++ $suffix }
}Error: nu::parser::operator_unsupported_type
× The '++' operator does not work on values of type 'nothing'.
╭─[repl_entry #8:2:5]
1 │ def mk-add-suffix [suffix: string]: nothing -> closure {
2 │ {|| $in ++ $suffix }}
· ─┬─ ─┬
· │ ╰── does not support 'nothing'
· ╰── nothing
╰────This is now fixed, and closures' input type will be inferred as any, the way it was before 0.114.0.
Allow more date commands to have a list as input -> output. These are the commands that were updated:
date humanizedate from-humandate to-timezoneformat dateinto datetimeinput listen's output type signature to match actual returned value. With the incorrect signature (and enforce-runtime-annotations on) let x = input listen would raise an error due to $x's inferred type (based on input listen's signature) not matching the value assigned to it. table and list<record> as compatible (provided the columns aren't disjoint), but that did not extend to list<oneof<record, ...>>. This has been fixed. uname's output type signature to match its actual output. Previously it was table, now it's record<...> and also includes its columns. run script.nu --some-flag could bind the value to the wrong flag of the script's main when main declared multiple flags, causing shifted positional arguments or a value being accepted into a flag of the wrong type. Flags passed to run are now matched to main parameters by name. (#18533)idx search --help example. (#18550)for loops where type of the looping variable was inferred incorrectly when iterating over a value/stream with a union type (oneof<list, table, binary>). This could manifest as a runtime type checking error with enforce-runtime-annotations enabled. (#18551)Thanks to all the contributors below for helping us solve issues, improve documentation, refactor code, and more! 🙏
| author | change | link | |
]]>@Bahex:
table and list<oneof> (#18560)input listen's output type (#18535)from xlsx/ods datetime fixes (#18516)@cptpiepmatz:
@Alb-O:
@fdncred:
@hexbinoct:
@Juhan280:
@stormasm:
@Bahex:
table and list<oneof> (#18560)input listen's output type (#18535)from xlsx/ods datetime fixes (#18516)@cptpiepmatz:
@Alb-O:
@fdncred:
@hexbinoct:
@Juhan280:
@stormasm:
@Kissaki:
@Alb-O:
@cptpiepmatz:
@hustcer:
@xav-ie:
@momomuchu:
-- option parsing, the new run command for pipeline-friendly scripts, and nice built-in support for working with Semantic Versioning.
Nu 0.114.0 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
-- option parsing, the new run command for pipeline-friendly scripts, and nice built-in support for working with Semantic Versioning.
Nu 0.114.0 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
@Bahex continued tightening up Nushell's type system this release. Commands can now infer output types from pipeline input more precisely, $in gets typed from the surrounding pipeline, optional values carry their nothing-ness more honestly, and let bindings in pipelines no longer have to shrug and call everything any.
This means some mistakes move from "surprise, runtime error" to "hey, maybe fix that before running it". Very rude of the parser, but also very helpful.
There is also one important default change: enforce-runtime-annotations is now opt-out, so runtime assignment annotations get checked by default.
Read the details in the first type system entry, the second type system entry, and the note about enforce-runtime-annotations.
-- means "no really, stop parsing flags" Thanks to @fdncred, Nushell commands now understand the classic POSIX -- end-of-options delimiter. If you need to pass -weird, --looks-like-a-flag, or similar gremlins as plain positional values, you can now do that without trying to outsmart the parser.
def greet [--upper, name] { if $upper { $name | str uppercase } else { $name } }
greet -- -AliceThat -Alice is now just Alice having a dash day, not an accidental flag.
Take a look at the main -- option parsing entry and the follow-up --wrapped fix.
run The new run command from @fdncred lets a Nushell script behave like a pipeline stage. Pipe data in, let the script transform it, and pipe the result onward. Tiny reusable pipeline workers, basically.
"Hello Nushell!" | run shout.nu | str camel-caseScripts can be simple bare pipeline transforms or define a main entry point, and they run in isolation so their local bits do not leak back into your session.
See more examples here.
Nushell now has semantic version support, also from @fdncred. You can parse SemVer strings, turn them into records, build them back up, sort them correctly, and bump versions without doing string surgery with safety scissors.
'1.2.3-alpha.1' | into semver | semver bump releaseIf your scripts deal with releases, package versions, or "is 1.10.0 bigger than 1.2.0?" moments, this should make life nicer.
Read the full SemVer tour here.
Importing a module no longer implicitly imports its exported sub-modules.
export module foo {
export def bar [] {}
export module sub {
export def baz [] {}
}
}
use foo
foo bar # valid
foo sub baz # invalid, `sub` is not implicitly importedTo have your modules keep the old behavior, you will need to export the module's contents explicitly:
export module foo {
export def bar [] {}
export module sub {
export def baz [] {}
}
export use sub
}
use foo
foo bar # valid
foo sub baz # validNote
useing a module inside another module does not run the imported module's export-env block.
(See export-env runs only when the use call is evaluated)
Thus, explicit uses as recommended above will not run export-env blocks, which is inline with how the previous implicit imports worked.
This release comes with a handful of improvements to the type system, specifically the parse-time type checking. But beware! Type annotations that previously seemed fine might raise errors, as they might have been passing only because typing information was insufficient to check it at parse time.
let str = "foo" | str uppercase
let str_list = ["foo", "bar" ] | str uppercase
scope variables
| where name starts-with '$str'
| select name type╭───┬───────────┬──────╮
│ # │ name │ type │
├───┼───────────┼──────┤
│ 0 │ $str │ any │
│ 1 │ $str_list │ any │
╰───┴───────────┴──────╯╭───┬───────────┬──────────────╮
│ # │ name │ type │
├───┼───────────┼──────────────┤
│ 0 │ $str │ string │
│ 1 │ $str_list │ list<string> │
╰───┴───────────┴──────────────╯nullThe (parse time) types of optional parameters and flags without a default value are now a union of their type T and nothing
oneof<T, nothing>TDemonstrating with LSP inlay hints:
1 def foo [opt_param?: int] {
2 let var: int = $opt_param
3 }
~ 1 def foo [opt_param?: int] {
2 let var: oneof<int, nothing> = $opt_param
3 }
~ 1 def foo [opt_param?: int = 42] {
2 let var: int = $opt_param
3 }
~More type information is preserved. e.g.:
let list = [["a" "b" "c"] [1 2 3]]
let elem = $list.0Previously, $elem's type would be list<oneof<int, string>>, now $elem's type is oneof<list<int>, list<string>>.
oneof typesNushell as a language tries to strike a careful balance between correctness and convenience. One should be able to trust nu's type system will help them write correct and robust scripts, but nu shouldn't get in the way of a nice REPL experience.
Because of that, some operations that might be invalid but aren't definitely so do not raise parse errors and such checks are delegated to runtime.
For example, int's can only be added with int's and float's, yet the following code does not raise a parse-time error.
def foo [anything: any] {
10 + $anything
}But previously, despite oneof<int, nothing> being more specific than any, it would be rejected by the type checker for addition to ints.
def foo [maybe_int: oneof<int, nothing>] {
10 + $maybe_int
}Error: nu::parser::operator_unsupported_type
× The '+' operator does not work on values of type 'oneof<int, nothing>'.
╭─[repl_entry #2:2:8]
1 │ def foo [maybe_int: oneof<int, nothing>] {
2 │ 10 + $maybe_int
· ┬ ─────┬────
· │ ╰── oneof<int, nothing>
· ╰── does not support 'oneof<int, nothing>'
3 │ }
╰────Now, as long as the two operands have a possibility of being valid for an operation the parse-time type checker will accept the expression, leaving errors to runtime type checking.
enforce-runtime-annotations to opt-out In Nushell 0.108.0, enforce-runtime-annotations experimental option was added. With this release enforce-runtime-annotations is promoted to being opt-out, meaning that by default this experimental option is enabled now.
This option enforces type checking of let assignments at runtime, catching type errors that might have been missed in parse-time type checking due to insufficient type information.
If you experience any trouble with this, you can disable it via the command-line argument:
nu --experimental-options '[enforce-runtime-annotations=false]'or via the environment variable:
NU_EXPERIMENTAL_OPTIONS="enforce-runtime-annotations=false" nuVi mode now has a proper visual mode: press v to start a selection, then use motions together with d, c, y,
or r, including selections that span multiple lines.
In vi normal mode, the caret now rests on the last grapheme instead of after it,
and h and l can move across line boundaries.
Visual mode also has its own prompt indicator, which defaults to the vi-normal
indicator.
from xlsx & from ods This release comes with various improvements to from xlsx and from ods commands, and makes them behave as similarly as the file formats themselves allow.
These commands' signatures previously incorrectly indicated that they return tables. They return records that contain tables, with a table for each sheet in the file.
Also, the signature of the from ods command previously had string as its input type rather than binary. This has been fixed.
Previously, importing datetime values were supported only for xlsx files.
This release not only adds that for ods files as well, but also broadens the datetime formats that can be imported.
Aside from that, values that couldn't be properly coerced to nushell types were simply imported as null. Now those values are imported as string or float depending on the source data.
Last release changed from xlsx to use the first row of a sheet as the imported table's headers, and added a new flag, --header-row, to control which row should be used as the header or if any row should become the headers at all.
This release removes it.
Don't worry! Instead of the --header-row flag, from xlsx and from ods get two new flags that will provide finer control over how spreadsheets are imported:
--noheaders: What it says on the tin, identical to the same flag on from csv, from tsv and from ssv commands.--first-row: This is where it gets a little interesting. from xlsx and from ods skip reading all leading empty rows. This is how it has worked for a long time.
--first-row=0 allows you to turn that off and keep those empty lines. It can also be used to start reading from an arbitrary row: --first-row $nLet's see how this works with an example:
open example.xlsx --raw
| from xlsx --noheaders --first-row 0╭────────┬───────────────────────────────────────────────╮
│ │ ╭───┬─────────┬─────────┬─────────┬─────────╮ │
│ Sheet1 │ │ # │ column0 │ column1 │ column2 │ column3 │ │
│ │ ├───┼─────────┼─────────┼─────────┼─────────┤ │
│ │ │ 0 │ │ │ │ │ │
│ │ │ 1 │ │ │ │ │ │
│ │ │ 2 │ │ │ │ │ │
│ │ │ 3 │ a │ b │ c │ d │ │
│ │ │ 4 │ 0.00 │ 1.00 │ 2.00 │ 3.00 │ │
│ │ │ 5 │ 10.00 │ 11.00 │ 12.00 │ 13.00 │ │
│ │ │ 6 │ 20.00 │ 21.00 │ 22.00 │ 23.00 │ │
│ │ │ 7 │ 30.00 │ 31.00 │ 32.00 │ 33.00 │ │
│ │ ╰───┴─────────┴─────────┴─────────┴─────────╯ │
╰────────┴───────────────────────────────────────────────╯open example.xlsx
# or `open example.xlsx --raw | from xlsx`╭────────┬───────────────────────────────────────╮
│ │ ╭───┬───────┬───────┬───────┬───────╮ │
│ Sheet1 │ │ # │ a │ b │ c │ d │ │
│ │ ├───┼───────┼───────┼───────┼───────┤ │
│ │ │ 0 │ 0.00 │ 1.00 │ 2.00 │ 3.00 │ │
│ │ │ 1 │ 10.00 │ 11.00 │ 12.00 │ 13.00 │ │
│ │ │ 2 │ 20.00 │ 21.00 │ 22.00 │ 23.00 │ │
│ │ │ 3 │ 30.00 │ 31.00 │ 32.00 │ 33.00 │ │
│ │ ╰───┴───────┴───────┴───────┴───────╯ │
╰────────┴───────────────────────────────────────╯open example.xlsx --raw | from xlsx --noheaders╭────────┬───────────────────────────────────────────────╮
│ │ ╭───┬─────────┬─────────┬─────────┬─────────╮ │
│ Sheet1 │ │ # │ column0 │ column1 │ column2 │ column3 │ │
│ │ ├───┼─────────┼─────────┼─────────┼─────────┤ │
│ │ │ 0 │ a │ b │ c │ d │ │
│ │ │ 1 │ 0.00 │ 1.00 │ 2.00 │ 3.00 │ │
│ │ │ 2 │ 10.00 │ 11.00 │ 12.00 │ 13.00 │ │
│ │ │ 3 │ 20.00 │ 21.00 │ 22.00 │ 23.00 │ │
│ │ │ 4 │ 30.00 │ 31.00 │ 32.00 │ 33.00 │ │
│ │ ╰───┴─────────┴─────────┴─────────┴─────────╯ │
╰────────┴───────────────────────────────────────────────╯$in$in variable is now typed based on the pipeline input type of the surrounding expression.
Error: nu::shell::operator_incompatible_types
× Types 'int' and 'string' are not compatible for the '+' operator.
╭─[source:1:5]
1 │ 2 | $in + "foo"
· ─┬─ ┬ ──┬──
· │ │ ╰── string
· │ ╰── does not operate between 'int' and 'string'
· ╰── int
╰────Error: nu::parser::operator_incompatible_types
× Types 'int' and 'string' are not compatible for the '+' operator.
╭─[source:1:5]
1 │ 2 | $in + "foo"
· ─┬─ ┬ ──┬──
· │ │ ╰── string
· │ ╰── does not operate between 'int' and 'string'
· ╰── int
╰────let bindingslet bindings in the middle of a pipeline, or at the end of it, are now properly typed based on the pipeline input type.
You can observe this using scope variables:
42 | let x
scope variables
| where name == '$x'
| select name type╭───┬──────┬──────╮
│ # │ name │ type │
├───┼──────┼──────┤
│ 0 │ $x │ int │
╰───┴──────┴──────╯or with LSP type hints:
42 | let var: intThis type information is of course passed through to the rest of the pipeline:
def accepts-int []: int -> int {}
"foo" | let x | accepts-intError: nu::shell::only_supports_this_input_type
× Input type not supported.
╭─[source:3:9]
2 │
3 │ "foo" | let x | accepts-int
· ─┬─ ─────┬─────
· │ ╰── only int input data is supported
· ╰── input type: string
╰────definitionsType checking and inference inside def blocks are now affected by the commands input/output signatures.
def accepts-int []: int -> int {}
def foo []: string -> any {
accepts-int
}Error: nu::parser::input_type_mismatch
× Command does not support string input.
╭─[source:4:2]
3 │ def foo []: string -> any {
4 │ accepts-int
· ─────┬─────
· ╰── command doesn't support string input
5 │ }
╰────if-else, matchConditionals (if, match) are now properly typed:
output of the expression is a union of all possible output types (based on the branches)
if $x == $y {
42
} else {
"foo"
}
| let out # oneof<int, string>
match $x {
1 => 1,
2 => "foo",
_ => { {a: 1} }
}
| let out # oneof<int, string, record<a: int>>without a fallback (a final else branch for if chains, a wildcard pattern for match expressions) the output types of conditional expressions include nothing
if $x == $y {
42
}
| let out # oneof<int, nothing>
match $x {
1 => 1,
2 => "foo",
}
| let out # oneof<int, string, nothing>branch blocks get pipeline input type information
[1, 2, 3]
| if true {
let input # list<int>
$input | into string
} else {
# pass through
}
| let out # oneof<list<string>, list<int>>details and file relative spans The error record you receive in try {..} catch {..} blocks no longer has a json field. A new field named details takes its place instead. It holds the same information as the json field but requires no deserialization stop like from json.
There is one small but significant difference: the error labels now have additional location field in addition to span.
location contains the name of the file the label points to, with start and end offsets that are relative to the file itself.
use std/assert
export def wrong [] {
assert equal (2 + 2) 5
}> use example.nu
> wrong
Error: nu::shell::error
× These are not equal.
╭─[/tmp/example.nu:4:16]
3 │ export def wrong [] {
4 │ assert equal (2 + 2) 5
· ───┬─── ┬
· │ ╰── right: 5
· ╰── left: 4
5 │ }
╰────
> try { wrong } catch { get details }
╭─────────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────╮
│ msg │ These are not equal. │
│ │ ╭───┬──────────┬────────────────────┬────────────────────────────────────────────────────╮ │
│ labels │ │ # │ text │ span │ location │ │
│ │ ├───┼──────────┼────────────────────┼────────────────────────────────────────────────────┤ │
│ │ │ 0 │ left: 4 │ ╭───────┬────────╮ │ ╭───────┬────────────────────────────────────────╮ │ │
│ │ │ │ │ │ start │ 170639 │ │ │ file │ D:\Projects\nushell\scratch\example.nu │ │ │
│ │ │ │ │ │ end │ 170646 │ │ │ start │ 56 │ │ │
│ │ │ │ │ ╰───────┴────────╯ │ │ end │ 63 │ │ │
│ │ │ │ │ │ ╰───────┴────────────────────────────────────────╯ │ │
│ │ │ 1 │ right: 5 │ ╭───────┬────────╮ │ ╭───────┬────────────────────────────────────────╮ │ │
│ │ │ │ │ │ start │ 170647 │ │ │ file │ D:\Projects\nushell\scratch\example.nu │ │ │
│ │ │ │ │ │ end │ 170648 │ │ │ start │ 64 │ │ │
│ │ │ │ │ ╰───────┴────────╯ │ │ end │ 65 │ │ │
│ │ │ │ │ │ ╰───────┴────────────────────────────────────────╯ │ │
│ │ ╰───┴──────────┴────────────────────┴────────────────────────────────────────────────────╯ │
│ code │ │
│ url │ │
│ help │ │
│ inner │ [list 0 items] │
╰─────────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────╯With these file-relative spans, making external tools for reporting nushell errors should be much easier.
grid record input support In the last release, the grid command was modified to accept a column name argument rather than implicitly displaying the name column. The implicit assumption of name was deprecated, along with accepting a single record rather than a table.
These deprecated behaviors of the grid command are removed with this release.
# This will no longer work
{ name: test } | grid
# …and should be updated to receive tables and include the column name as an argument
[{ name: test }] | grid nameFollowing general type system improvements, standard library commands now also have more accurate type annotations.
std/iter find-index returns null instead of -1 when failing to find a suitable element
std/iter scan now has the same signature as reduceWarning
This is a breaking change.
Previously iter scan always needed an initial value to start off of (equivalent to reduce's --fold), and to compensate for it a --noinit flag to remove that initial value from the output.
Now its signature is identical to reduce:
[1 2 3] | iter scan 0 {|x, y| $x + $y}[1 2 3] | iter scan 0 --noinit {|x, y| $x + $y}[1 2 3] | iter scan --fold 0 {|x, y| $x + $y}[1 2 3] | iter scan {|x, y| $x + $y}[0, 1, 3, 6][1, 3, 6]The following commands now stream:
std/iter:
intersperseflat-mapstd-rfc/conversion
into liststd-rfc/tables
select column-slicesreject column-slicesThese commands don't really "stream", in that they don't return a stream. Instead, they now avoid eagerly collecting input streams:
std-rfc/conversions
name-values: strictly an internal change with possible performance benefitsstd-rfc/iter
only: will at most consume 2 items from the inputidx find and idx search. (#18477)This release adds support for KDL v2.0.0 with to kdl and from kdl commands.
> '
hello 1 2 3
// Comment
world prop=string-value {
child 1
child 2
child #inf
}
'
| from kdl
╭───┬───────┬────────────────┬─────────────────────────┬──────────────────────────────────────────────────────────────────╮
│ # │ name │ args │ props │ children │
├───┼───────┼────────────────┼─────────────────────────┼──────────────────────────────────────────────────────────────────┤
│ 0 │ hello │ ╭───┬───╮ │ {record 0 fields} │ [list 0 items] │
│ │ │ │ 0 │ 1 │ │ │ │
│ │ │ │ 1 │ 2 │ │ │ │
│ │ │ │ 2 │ 3 │ │ │ │
│ │ │ ╰───┴───╯ │ │ │
│ 1 │ world │ [list 0 items] │ ╭──────┬──────────────╮ │ ╭───┬───────┬─────────────┬───────────────────┬────────────────╮ │
│ │ │ │ │ prop │ string-value │ │ │ # │ name │ args │ props │ children │ │
│ │ │ │ ╰──────┴──────────────╯ │ ├───┼───────┼─────────────┼───────────────────┼────────────────┤ │
│ │ │ │ │ │ 0 │ child │ ╭───┬───╮ │ {record 0 fields} │ [list 0 items] │ │
│ │ │ │ │ │ │ │ │ 0 │ 1 │ │ │ │ │
│ │ │ │ │ │ │ │ ╰───┴───╯ │ │ │ │
│ │ │ │ │ │ 1 │ child │ ╭───┬───╮ │ {record 0 fields} │ [list 0 items] │ │
│ │ │ │ │ │ │ │ │ 0 │ 2 │ │ │ │ │
│ │ │ │ │ │ │ │ ╰───┴───╯ │ │ │ │
│ │ │ │ │ │ 2 │ child │ ╭───┬─────╮ │ {record 0 fields} │ [list 0 items] │ │
│ │ │ │ │ │ │ │ │ 0 │ inf │ │ │ │ │
│ │ │ │ │ │ │ │ ╰───┴─────╯ │ │ │ │
│ │ │ │ │ ╰───┴───────┴─────────────┴───────────────────┴────────────────╯ │
╰───┴───────┴────────────────┴─────────────────────────┴──────────────────────────────────────────────────────────────────╯polars map-batches Provides a new polars command polars map-batches that provides the ability to map a custom Nushell closure over one or more dataframe columns.
> [[a b]; [1 4] [2 5] [3 6]]
| polars into-df
| polars map-batches --name out { [10 20 30] } a
╭───┬─────╮
│ # │ out │
├───┼─────┤
│ 0 │ 10 │
│ 1 │ 20 │
│ 2 │ 30 │
╰───┴─────╯> [[a b]; [1 4] [2 5] [3 6]]
| polars into-df
| polars map-batches {|cols| $cols | first | polars get a | each { $in * 2 }} a
╭───┬───╮
│ # │ a │
├───┼───┤
│ 0 │ 2 │
│ 1 │ 4 │
│ 2 │ 6 │
╰───┴───╯> [[a b]; [1 4] [2 5] [3 6]]
| polars into-df
| polars map-batches --name a_plus_b { |cols|
let a = $cols | get 0 | polars get a
let b = $cols | get 1 | polars get b
$a | zip $b | each { |pair| $pair.0 + $pair.1 }
} a b
╭───┬──────────╮
│ # │ a_plus_b │
├───┼──────────┤
│ 0 │ 5 │
│ 1 │ 7 │
│ 2 │ 9 │
╰───┴──────────╯commandline complete to invoke nushell completions Add commandline complete command. This can be used to obtain the suggestions that nushell itself would normally suggest. Some examples:
commandline complete — return completions based on the current commandline contents (i.e. what would be suggested when pressing Tab)'./a' | commandline complete --type directory — return directory paths relative to the current directory that start with a'%ls -' | commandline complete --detailed — return all flags the builtin ls command accepts, including their descriptions.These completions can be used in custom command completions, for example to wrap a builtin command with additional options, or to obtain a list of suggested paths for further filtering or processing.
--pretty flag and align table columns in to nuon to nuon now aligns table columns when using --indent, --tabs, or the new --pretty (-p) flag, making output easier to read. --pretty is shorthand for --indent 2.
> [[name, age]; [Alice, 30], [Bob, 25]] | to nuon --pretty
[
[name, age];
[Alice, 30],
[Bob, 25]
]run command for using scripts in pipelines There's a new command named run that allows scripts to participate in a nushell pipeline with input and output.
It keeps execution isolated so script artifacts do not leak into the parent scope. Script resolution remains aligned with existing behavior (cwd / NU_LIB_DIRS / explicit paths); PATH-based lookup is not included at this time.
The script body is evaluated directly as a pipeline transform (using implicit or explicit $in as usual).
str uppercase> "Hello Nushell!" | run run-script1.nu
HELLO NUSHELL!def main entry pointIf a top-level def main is defined, run invokes that entrypoint.
def main [] {
str length
}> "Hello Nushell!" | run run-script2.nu
14> "Hello Nushell!" | run run-script1.nu | str camel-case
helloNushell
> "Hello Nushell!" | run run-script2.nu | $in + 5
19-- option parsing for Nushell commands Nushell now supports the POSIX -- end-of-options delimiter for built-in commands, custom commands, and def --wrapped commands.
When -- appears in a command invocation, it stops all flag parsing. Any arguments after -- are treated as positional operands, even if they start with - or --. The -- token itself is consumed and does not appear in the command's arguments.
> def greet [--upper, name] {
if $upper { $name | str uppercase } else { $name }
}
> greet -- -Alice # -Alice is now a positional, not an unknown flag
-Alice> def process [...args] { $args }
> process -- --verbose -x foo # -- consumed; args = ["--verbose", "-x", "foo"]
╭───┬───────────╮
│ 0 │ --verbose │
│ 1 │ -x │
│ 2 │ foo │
╰───┴───────────╯> def --wrapped my-git [...args] { print ...$args }
> my-git commit -- -m "my message" # -- NOT consumed; -m passed as operand
commit
--
-m
my messageFor extern-declared known external commands, -- continues to be passed through to the external binary unchanged (as those programs manage their own argument parsing).
ignore Update the ignore command with flags to control whether stderr, stdout, or both get consumed and if errors are show which update $env.LAST_EXIT_CODE. This should work on internal and external commands.
Existing behavior is the default when no new flags are provided to avoid breaking changes.
--stderr (-e) to consume stderr while allowing stdout to pass through.--stdout (-o) to consume stdout while allowing stderr output through.--show-errors (-x) to show external/internal errors and set $env.LAST_EXIT_CODE (external exit code for external failures, 1 for internal failures).--context support to idx search Add context to your idx search queries by using --context/-c and a nushell range. The range has to be in the form of negative_number..positive_number where negative_number is the before lines to show and the positive_number is the after lines to show in the with_context record key.
> idx search foo --context -3..5 | first
╭───────────────┬─────────────────────────────────────────────────────────────────────────────────────────────╮
│ relative_path │ crates\nu-utils\src\default_files\doc_config.nu │
│ line_number │ 249 │
│ column │ 51 │
│ byte_offset │ 10991 │
│ line │ # Example: If a directory contains only "forage", "food", and "forest", │
│ │ ╭───┬───────┬───────╮ │
│ matches │ │ # │ start │ end │ │
│ │ ├───┼───────┼───────┤ │
│ │ │ 0 │ 11042 │ 11045 │ │
│ │ ╰───┴───────┴───────╯ │
│ │ ╭───┬─────────────────────────────────────────────────────────────────────────────────────╮ │
│ with_context │ │ 0 │ # Default: true │ │
│ │ │ 1 │ $env.config.completions.partial = true │ │
│ │ │ 2 │ │ │
│ │ │ 3 │ # Example: If a directory contains only "forage", "food", and "forest", │ │
│ │ │ 4 │ # typing "ls " and pressing Tab will partially complete the first matching letters. │ │
│ │ │ 5 │ # If the directory also includes "faster", only "f" would be partially completed. │ │
│ │ │ 6 │ │ │
│ │ │ 7 │ # completions.use_ls_colors (bool): Apply LS_COLORS to file/path completions. │ │
│ │ │ 8 │ # true: Use LS_COLORS for styling file completions. │ │
│ │ ╰───┴─────────────────────────────────────────────────────────────────────────────────────╯ │
╰───────────────┴─────────────────────────────────────────────────────────────────────────────────────────────╯idx search --context to accept counts or ranges This change allows idx search --context to be supplied as either an int or as a range.
> idx search foo --context 2 | first
╭───────────────┬─────────────────────────────────────────────────────────────────────────────────────────────╮
│ relative_path │ crates\nu-utils\src\default_files\doc_config.nu │
│ line_number │ 249 │
│ column │ 51 │
│ byte_offset │ 10991 │
│ line │ # Example: If a directory contains only "forage", "food", and "forest", │
│ │ ╭───┬───────┬───────╮ │
│ matches │ │ # │ start │ end │ │
│ │ ├───┼───────┼───────┤ │
│ │ │ 0 │ 11042 │ 11045 │ │
│ │ ╰───┴───────┴───────╯ │
│ │ ╭───┬─────────────────────────────────────────────────────────────────────────────────────╮ │
│ with_context │ │ 0 │ $env.config.completions.partial = true │ │
│ │ │ 1 │ │ │
│ │ │ 2 │ # Example: If a directory contains only "forage", "food", and "forest", │ │
│ │ │ 3 │ # typing "ls " and pressing Tab will partially complete the first matching letters. │ │
│ │ │ 4 │ # If the directory also includes "faster", only "f" would be partially completed. │ │
│ │ ╰───┴─────────────────────────────────────────────────────────────────────────────────────╯ │
╰───────────────┴─────────────────────────────────────────────────────────────────────────────────────────────╯Introducing polars bitwise commands:
polars math bitwise-and - Perform an aggregation of bitwise ANDs over a column expression.polars math bitwise-count-ones - Compute the number of set bits for each element in an integer column expression.polars math bitwise-count-zeros - Compute the number of unset bits for each element in an integer column expression.polars math bitwise-leading-ones - Compute the number of leading set bits for each element in an integer column expression.polars math bitwise-leading-zeros - Compute the number of leading unset bits for each element in an integer column expression.polars math bitwise-or - Perform an aggregation of bitwise ORs over a column expression.polars math bitwise-trailing-ones - Compute the number of trailing set bits for each element in an integer column expression.polars math bitwise-trailing-zeros - Compute the number of trailing unset bits for each element in an integer column expression.polars math bitwise-xor - Perform an aggregation of bitwise XORs over a column expression.url encode/url decode url encode can now encode binary input too, adding support for encoding non-UTF-8 texts
> '£ rates' | encode iso-8859-1 | url encode
%A3%20ratesurl decode can now decode inputs that do not evaluate to UTF-8 texts by returning a binary value (only with the --binary switch)
> '%A3%20rates' | url decode --binary | decode iso-8859-1
£ ratesReedline menus can now be configured with input_mode (diff, cursor_prefix, full_buffer) and output_mode (suggested_span, full_buffer, extend_to_end) to control what text the menu reads from the buffer and how an accepted suggestion is written back.
List menus additionally accept description_position (before / after). The existing only_buffer_difference flag still works and is treated as a shorthand (true = diff, false = cursor_prefix).
random pass for generating passwords New command random pass generates random passwords.
> 0..10 | each {random pass}
╭────┬──────────────╮
│ 0 │ WR9+*T4em#&2 │
│ 1 │ >~z$0I-E<3&x │
│ 2 │ "%uF9qafB%+{ │
│ 3 │ VMI5h+T4,XcX │
│ 4 │ B,diE+Z`#O6g │
│ 5 │ 9fKehf4|F,xr │
│ 6 │ 6d?nH1|HV^ck │
│ 7 │ iH>5l?`P7q2x │
│ 8 │ w>w2I(v67=9E │
│ 9 │ QYqc}g8ux#!U │
│ 10 │ #Jla4UAoJ+<G │
╰────┴──────────────╯random pass supports these flags.
Flags:
-h, --help: Display the help message for this command
-c, --chars <int>: Length of the generated password (default 12).
-u, --no-uppercase: Exclude uppercase letters A-Z.
-l, --no-lowercase: Exclude lowercase letters a-z.
-n, --no-numbers: Exclude numbers 0-9.
-s, --no-symbols: Exclude symbols like !@#$%.
--include-ambiguous: Include ambiguous characters O, 0, l, 1.
--include-similar: Include similar characters i, l, 1.
--require-each-type: Guarantee at least one char from each enabled character type.is-terminal to detect redirection The is-terminal command can detect redirection now. It now defaults to --stdout.
> is-terminal
true
> is-terminal | $in
false
> is-terminal o> file
> open file
false
> let x = (is-terminal)
> $x
false--right to split row and split column Both split row and split column now support --right, which modifies where splits are skipped, e.g.:
> 'some-package-1.0' | split row '-' --number 2
╭───┬─────────────╮
│ 0 │ some │
│ 1 │ package-1.0 │
╰───┴─────────────╯
> 'some-package-1.0' | split row '-' --number 2 --right
╭───┬──────────────╮
│ 0 │ some-package │
│ 1 │ 1.0 │
╰───┴──────────────╯Added union, intersect, difference, combinations, and permutations as built-in filter commands. These provide immutable, functional list operations.
union: Returns a deduplicated list of unique elements present in either the input list or the given list.
> [1 2 3 4] | union [3 4 5 6]
╭───┬───╮
│ 0 │ 1 │
│ 1 │ 2 │
│ 2 │ 3 │
│ 3 │ 4 │
│ 4 │ 5 │
│ 5 │ 6 │
╰───┴───╯
> [{a:1} {a:2}] | union [{a:2} {a:3}]
╭───┬───╮
│ # │ a │
├───┼───┤
│ 0 │ 1 │
│ 1 │ 2 │
│ 2 │ 3 │
╰───┴───╯intersect: Returns a deduplicated list of unique elements present in both the input list and the given list.
> [1 2 3 4] | intersect [3 4 5 6]
╭───┬───╮
│ 0 │ 3 │
│ 1 │ 4 │
╰───┴───╯
> [1 2 3] | intersect [4 5 6]
╭────────────╮
│ empty list │
╰────────────╯difference: Returns a deduplicated list of unique elements present in the input list but not in the given list.
> [1 2 3 4] | difference [3 4 5 6]
╭───┬───╮
│ 0 │ 1 │
│ 1 │ 2 │
╰───┴───╯
> [{a:1} {a:2} {a:3}] | difference [{a:2} {a:4}]
╭───┬───╮
│ # │ a │
├───┼───┤
│ 0 │ 1 │
│ 1 │ 3 │
╰───┴───╯combinations: Generates all combinations of a given size k from the input list, streamed lazily via ListStream. If k > n, returns an empty list.
> [1 2 3] | combinations 2
╭───┬───────────╮
│ 0 │ ╭───┬───╮ │
│ │ │ 0 │ 1 │ │
│ │ │ 1 │ 2 │ │
│ │ ╰───┴───╯ │
│ 1 │ ╭───┬───╮ │
│ │ │ 0 │ 1 │ │
│ │ │ 1 │ 3 │ │
│ │ ╰───┴───╯ │
│ 2 │ ╭───┬───╮ │
│ │ │ 0 │ 2 │ │
│ │ │ 1 │ 3 │ │
│ │ ╰───┴───╯ │
╰───┴───────────╯
> [1 2] | combinations 3
╭────────────╮
│ empty list │
╰────────────╯permutations: Generates all permutations of the input list using Heap's algorithm, streamed lazily via ListStream.
> [1 2 3] | permutations
╭───┬───────────╮
│ 0 │ ╭───┬───╮ │
│ │ │ 0 │ 1 │ │
│ │ │ 1 │ 2 │ │
│ │ │ 2 │ 3 │ │
│ │ ╰───┴───╯ │
│ 1 │ ╭───┬───╮ │
│ │ │ 0 │ 2 │ │
│ │ │ 1 │ 1 │ │
│ │ │ 2 │ 3 │ │
│ │ ╰───┴───╯ │
│ 2 │ ╭───┬───╮ │
│ │ │ 0 │ 3 │ │
│ │ │ 1 │ 1 │ │
│ │ │ 2 │ 2 │ │
│ │ ╰───┴───╯ │
│ 3 │ ╭───┬───╮ │
│ │ │ 0 │ 1 │ │
│ │ │ 1 │ 3 │ │
│ │ │ 2 │ 2 │ │
│ │ ╰───┴───╯ │
│ 4 │ ╭───┬───╮ │
│ │ │ 0 │ 2 │ │
│ │ │ 1 │ 3 │ │
│ │ │ 2 │ 1 │ │
│ │ ╰───┴───╯ │
│ 5 │ ╭───┬───╮ │
│ │ │ 0 │ 3 │ │
│ │ │ 1 │ 2 │ │
│ │ │ 2 │ 1 │ │
│ │ ╰───┴───╯ │
╰───┴───────────╯
> [1 2] | permutations
╭───┬───────────╮
│ 0 │ ╭───┬───╮ │
│ │ │ 0 │ 1 │ │
│ │ │ 1 │ 2 │ │
│ │ ╰───┴───╯ │
│ 1 │ ╭───┬───╮ │
│ │ │ 0 │ 2 │ │
│ │ │ 1 │ 1 │ │
│ │ ╰───┴───╯ │
╰───┴───────────╯Nushell now understands and allows you to work with Semantic Versioning. All functionality works through idiomatic Nushell patterns:
> '1.2.3' | into semver
1.2.3
> {major: 1, minor: 2, patch: 3} | into semver
1.2.3
> {major: 1, minor: 2, patch: 3} | into semver | describe
semver> '1.2.3-alpha.1+build.2' | into semver | into record
╭───────────────────┬───────────────╮
│ major │ 1 │
│ minor │ 2 │
│ patch │ 3 │
│ pre │ alpha.1 │
│ build │ build.2 │
│ │ ╭───┬───────╮ │
│ pre_identifiers │ │ 0 │ alpha │ │
│ │ │ 1 │ 1 │ │
│ │ ╰───┴───────╯ │
│ │ ╭───┬───────╮ │
│ build_identifiers │ │ 0 │ build │ │
│ │ │ 1 │ 2 │ │
│ │ ╰───┴───────╯ │
╰───────────────────┴───────────────╯> {major: 1, minor: 2, patch: 3, pre: "alpha.1"} | into semver
1.2.3-alpha.1> '1.2.3' | into semver | semver bump major
2.0.0
> '1.2.3' | into semver | semver bump minor
1.3.0
> '1.2.3' | into semver | semver bump patch
1.2.4
> '1.2.3' | into semver | semver bump alpha
1.2.3-alpha.1
> '1.2.3-alpha.1' | into semver | semver bump alpha
1.2.3-alpha.2
> '1.2.3-alpha' | into semver | semver bump release
1.2.3
> '1.2.3' | semver bump major
2.0.0
> '1.2.3' | semver bump minor
1.3.0| '1.2.3' | semver bump
alpha beta major minor
patch rc release> let v = '1.2.3' | into semver
| $v.
build major minor patch
pre> ['2.0.0', '1.0.0', '1.2.3'] | each { into semver } | sort
╭───┬───────╮
│ 0 │ 1.0.0 │
│ 1 │ 1.2.3 │
│ 2 │ 2.0.0 │
╰───┴───────╯
> ['2.0.0', '1.0.0', '1.2.3'] | each { into semver } | sort --reverse
╭───┬───────╮
│ 0 │ 2.0.0 │
│ 1 │ 1.2.3 │
│ 2 │ 1.0.0 │
╰───┴───────╯> try { '1.2.3' | into semver } | is-not-empty
true
> try { 'not-valid' | into semver } | is-not-empty
false> '1.2.3' | into semver | $in in ('>=1.0.0' | into semver-range)
true
> '1.0.0' | into semver | $in in ('>=2.0.0' | into semver-range)
false
> '1.3.0-alpha' | into semver | $in in ('>=1.2.3' | into semver-range)
false> '>=1.0.0' | into semver-range
>=1.0.0
> '^1.2.3' | into semver-range
^1.2.3
> '~1.2' | into semver-range
~1.2> let v = '1.2.3' | into semver
> $v | describe
semver
> $v | describe -d
╭───────────────┬───────────────────────────────────────────────────────────────────────╮
│ type │ custom │
│ detailed_type │ semver │
│ subtype │ semver │
│ rust_type │ &alloc::boxed::Box<dyn nu_protocol::value::custom_value::CustomValue> │
│ value │ 1.2.3 │
╰───────────────┴───────────────────────────────────────────────────────────────────────╯> [
[string, semver, semver-range];
['1.2.3', ('1.2.3' | into semver), ('>=1.0.0' | into semver-range)]
]
╭───┬────────┬────────┬──────────────╮
│ # │ string │ semver │ semver-range │
├───┼────────┼────────┼──────────────┤
│ 0 │ 1.2.3 │ 1.2.3 │ >=1.0.0 │
╰───┴────────┴────────┴──────────────╯$env.config.color_config.semver = "cyan_bold"
$env.config.color_config.semver-range = "cyan_bold"from xlsx/from ods gets a --prefer-integers flag xlsx and ods files store numbers as floats by default, even if they are not display with decimal points.
from xlsx and from ods now has a --prefer-integers (-i) flag that imports whole-number floats as integers.
This has no effect on non-whole floats.
> open --raw tests/fixtures/formats/sample_data.xlsx
| from xlsx
| get SalesOrders
| first 5
╭───┬─────────────┬─────────┬─────────┬────────┬───────┬───────────┬────────╮
│ # │ OrderDate │ Region │ Rep │ Item │ Units │ Unit Cost │ Total │
├───┼─────────────┼─────────┼─────────┼────────┼───────┼───────────┼────────┤
│ 0 │ 8 years ago │ East │ Jones │ Pencil │ 95.00 │ 1.99 │ 189.05 │
│ 1 │ 8 years ago │ Central │ Kivell │ Binder │ 50.00 │ 19.99 │ 999.50 │
│ 2 │ 8 years ago │ Central │ Jardine │ Pencil │ 36.00 │ 4.99 │ 179.64 │
│ 3 │ 8 years ago │ Central │ Gill │ Pen │ 27.00 │ 19.99 │ 539.73 │
│ 4 │ 8 years ago │ West │ Sorvino │ Pencil │ 56.00 │ 2.99 │ 167.44 │
╰───┴─────────────┴─────────┴─────────┴────────┴───────┴───────────┴────────╯> open --raw tests/fixtures/formats/sample_data.xlsx
| from xlsx --prefer-integers
| get SalesOrders
| first 5
╭───┬─────────────┬─────────┬─────────┬────────┬───────┬───────────┬────────╮
│ # │ OrderDate │ Region │ Rep │ Item │ Units │ Unit Cost │ Total │
├───┼─────────────┼─────────┼─────────┼────────┼───────┼───────────┼────────┤
│ 0 │ 8 years ago │ East │ Jones │ Pencil │ 95 │ 1.99 │ 189.05 │
│ 1 │ 8 years ago │ Central │ Kivell │ Binder │ 50 │ 19.99 │ 999.50 │
│ 2 │ 8 years ago │ Central │ Jardine │ Pencil │ 36 │ 4.99 │ 179.64 │
│ 3 │ 8 years ago │ Central │ Gill │ Pen │ 27 │ 19.99 │ 539.73 │
│ 4 │ 8 years ago │ West │ Sorvino │ Pencil │ 56 │ 2.99 │ 167.44 │
╰───┴─────────────┴─────────┴─────────┴────────┴───────┴───────────┴────────╯math cbrt (cube root) function. (#18473)append can now take multiple values to append as rest parameters (#18218)LS_COLORS support to commandline complete --detailed. (#18325)idx init command uses content indexing by default. (#18332)hash md5 and hash sha256 help examples to show hashing binary data. (#18338)ansi gradient's --fgnamed and --bgnamed flags. (#18342)nu-highlight now clears content_type metadata (#18266)edit events now support reedline's verb-based edit commands (move, extend, cut, copy, change, erase) combined with a motion and its parameters, enabling vi-style operator/motion keybindings to be expressed directly in config. The new commands and their fields are discoverable via keybindings list. (#18396)http commands now use the response body as error message when getting an error response. (#18387)NU_EXPERIMENTAL_OPTIONS environment variable is set, it will be used when running nushell with the --login or --execute flags since they also load config. (#18392)explore config command: hl to collapse/expand, and jk/ctrl+p/n to go up and down. (#18440)--log-level called perf that shows performance information. (#18468)str upcase/str downcase and add str uppercase/str lowercase Added str uppercase command to convert text to uppercase (replaces str upcase)
Added str lowercase command to convert text to lowercase (replaces str downcase)
Deprecated str upcase, using it shows a warning recommending str uppercase.
Deprecated str downcase, using it shows a warning recommending str lowercase.
use std/clip copy and paste commands; they are now called copy52 and paste52 for the ones that use the terminal OSC 52, or you can use clip copy and clip paste from the experimental option native-clip that access the system clipboard directly. (#18403)Info
The next release will have a complete rework of the YAML implementation. So this is only relevant for this release.
The previous emitter always quoted string values, which produced valid YAML but was more conservative than necessary and less aligned with YAML 1.2 plain-scalar rules.
This change makes output:
value: 'off'
path: '/dev/stdout'
listen: '0.0.0.0:8444,0.0.0.0:8445 ssl'
name: 'kong'
kind: 'Deployment'value: 'off'
path: /dev/stdout
listen: 0.0.0.0:8444,0.0.0.0:8445 ssl
name: kong
kind: DeploymentMultiline strings now emit as:
string: |-
Hello
worldAdded support for column width-priority hints in table rendering. Values can now carry --table-width-priority-columns metadata so the table command gives selected columns more space when fitting output to the terminal width.
Providing the metadata will cause the table command to render the output differently
> ps -l | select name command pid | first 2 | table
╭───┬──────────────────────────────────┬───────────────────────────────────────────────────────────────────────────┬─────╮
│ # │ name │ command │ ... │
├───┼──────────────────────────────────┼───────────────────────────────────────────────────────────────────────────┼─────┤
│ 0 │ GameInputRedistService.exe │ C:\Program Files\Microsoft GameInput\x64\GameInputRedistService.exe │ ... │
│ │ │ /session=\\.\pipe\GameInputServiceSession-002fb4627cd24645-00000001 │ │
│ 1 │ nvcontainer.exe │ C:\Program Files\NVIDIA Corporation\NvContainer\nvcontainer.exe -f │ ... │
│ │ │ C:\ProgramData\NVIDIA Corporation\NVIDIA │ │
│ │ │ App\NvContainer\NvContainerUser%d.log -d C:\Program Files\NVIDIA │ │
│ │ │ Corporation\NvContainer\plugins\User -r -l 3 -p 30000 -c │ │
╰───┴──────────────────────────────────┴───────────────────────────────────────────────────────────────────────────┴─────╯
> ps -l | select name command pid | first 2 | metadata set -w [pid] | table
╭───┬────────────────────────────┬───────────────────────────────────────────────────────────────────────────────┬───────╮
│ # │ name │ command │ pid │
├───┼────────────────────────────┼───────────────────────────────────────────────────────────────────────────────┼───────┤
│ 0 │ GameInputRedistService.exe │ C:\Program Files\Microsoft GameInput\x64\GameInputRedistService.exe │ 12744 │
│ │ │ /session=\\.\pipe\GameInputServiceSession-002fb4627cd24645-00000001 │ │
│ 1 │ nvcontainer.exe │ C:\Program Files\NVIDIA Corporation\NvContainer\nvcontainer.exe -f │ 12792 │
│ │ │ C:\ProgramData\NVIDIA Corporation\NVIDIA │ │
│ │ │ App\NvContainer\NvContainerUser%d.log -d C:\Program Files\NVIDIA │ │
│ │ │ Corporation\NvContainer\plugins\User -r -l 3 -p 30000 -c │ │
╰───┴────────────────────────────┴───────────────────────────────────────────────────────────────────────────────┴───────╯polar implode now supports the flag --maintain-orderpolars is-in now supports the flag --maintain-orderpolars pivot now supports the flag --always-combine-namespolars replace requires a --default <value> parameter when used with --strict and --return-dtype if the dtype has changed from the original dtypeThis release introduces a new experimental feature named dc-glob (named for Devyn Cairn's glob).
Currently, Nushell has two separate globbing engines with varying features and platform support.
This engine is intended to replace the existing globbing engines and standardize on something that works well on all platforms. In addition, this globbing engine seems to perform faster than the existing engines.
sqrt will suggest math sqrt. (#18498)plugin list --help to show the correct output type and example for the commands column, which contains each plugin command's name and description. (#18357)@attrs to the toolkit.nu commands so that they show up in category, search terms, and examples. I also homogenized help casing and punctuation. (#18350)--features lsp. (#18148)http post/put/patch/delete now respects --content-type for JSON-variant MIME types like application/json-patch+json. (#18496)idx dependency update and command reliability improvements Update the idx fff-search dependency and fix bugs.
idx init . --wait now properly blocksidx init now has a --follow-links switchidx status reports accurate counts after indexingidx import fully restores snapshot for queries — The idx import command now properly restores snapshots to enable all query operations. Imported snapshots remain queryable even if the original project files have been deleted or moved.idx files works on imported snapshotsidx files has a query parameter now for easy searching (renamed from path for consistency)idx dirs works on imported snapshotsidx dirs has a query parameter now for easy searchingidx find (fuzzy search) works on imported snapshotsidx search (content search) works on imported snapshotsFiles are ignored during indexing due to the use of the ignore crate. These are the rules for ignore https://docs.rs/ignore/latest/ignore/struct.WalkBuilder.html#ignore-rules. There's no way to override this in the current version.
which Tab completions for which and attr complete/@complete commands will now properly quote command names containing spaces.
> which 'path par
> which path parse
╭───┬─────────┬──────┬──────────╮
│ # │ command │ path │ type │
├───┼─────────┼──────┼──────────┤
│ 0 │ path │ │ built-in │
│ 1 │ parse │ │ built-in │
╰───┴─────────┴──────┴──────────╯> which 'path par
> which "path parse"
╭───┬────────────┬──────┬──────────╮
│ # │ command │ path │ type │
├───┼────────────┼──────┼──────────┤
│ 0 │ path parse │ │ built-in │
╰───┴────────────┴──────┴──────────╯This fixes a bug where update with a nested column path and closure (for example rss_item.pubDate) could evaluate the closure against a projected list and write the same result back to every row.
> {
rss_item: [
[pubDate];
[1773429600],
[1774325700],
[1775448000]
]
} | update rss_item.pubDate { into datetime -f "%s" }
╭──────────┬──────────────────────────────╮
│ │ ╭───┬──────────────────────╮ │
│ rss_item │ │ # │ pubDate │ │
│ │ ├───┼──────────────────────┤ │
│ │ │ 0 │ ╭───┬──────────────╮ │ │
│ │ │ │ │ 0 │ 3 months ago │ │ │
│ │ │ │ │ 1 │ 3 months ago │ │ │
│ │ │ │ │ 2 │ 2 months ago │ │ │
│ │ │ │ ╰───┴──────────────╯ │ │
│ │ │ 1 │ ╭───┬──────────────╮ │ │
│ │ │ │ │ 0 │ 3 months ago │ │ │
│ │ │ │ │ 1 │ 3 months ago │ │ │
│ │ │ │ │ 2 │ 2 months ago │ │ │
│ │ │ │ ╰───┴──────────────╯ │ │
│ │ │ 2 │ ╭───┬──────────────╮ │ │
│ │ │ │ │ 0 │ 3 months ago │ │ │
│ │ │ │ │ 1 │ 3 months ago │ │ │
│ │ │ │ │ 2 │ 2 months ago │ │ │
│ │ │ │ ╰───┴──────────────╯ │ │
│ │ ╰───┴──────────────────────╯ │
╰──────────┴──────────────────────────────╯> {
rss_item: [
[pubDate];
[1773429600],
[1774325700],
[1775448000]
]
} | update rss_item.pubDate { into datetime -f "%s" }
╭──────────┬──────────────────────╮
│ │ ╭───┬──────────────╮ │
│ rss_item │ │ # │ pubDate │ │
│ │ ├───┼──────────────┤ │
│ │ │ 0 │ 3 months ago │ │
│ │ │ 1 │ 3 months ago │ │
│ │ │ 2 │ 2 months ago │ │
│ │ ╰───┴──────────────╯ │
╰──────────┴──────────────────────╯> { w: { x: [ { y: "1" } { y: "2" } ] } }
| update w.x.y { into float }
| to yaml
w:
x:
- 'y':
- 1.0
- 2.0
- 'y':
- 1.0
- 2.0
> { w: { x: [ { y: "1" } { y: "2" } ] } }
| update w.x.y { into float }
| to yaml
w:
x:
- 'y': 1.0
- 'y': 2.0
> [ [[a]; [1] [2]] ]
| update a {into float}
| to yaml
-
- a:
- 1.0
- 2.0
- a:
- 1.0
- 2.0
> [ [[a]; [1] [2]] ]
| update a {into float}
| to yaml
-
- a: 1.0
- a: 2.0
nu-mcp evaluation output The MCP evaluate tool keeps evaluated Nushell values structured in output, so output data is fully unescaped instead of a wrapped string. Successful and error responses are also mirrored as MCP structuredContent JSON for clients that support structured tool output.
{
"jsonrpc": "2.0",
"id": 3,
"result": {
"content": [
{
"type": "text",
"text": "{cwd:/Users/fdncred/src/nushell,history_index:0,timestamp:2026-05-21T15:10:18.619935+00:00,output:\"\\\"[\n {\n \\\\\\\"name\\\\\\\": \\\\\\\"assets\\\\\\\",\n \\\\\\\"type\\\\\\\": \\\\\\\"dir\\\\\\\",\n \\\\\\\"size\\\\\\\": 160,\n \\\\\\\"modified\\\\\\\": \\\\\\\"2024-09-14T07:06:47.316979108-05:00\\\\\\\"\n },\n {\n \\\\\\\"name\\\\\\\": \\\\\\\"ast-grep\\\\\\\",\n \\\\\\\"type\\\\\\\": \\\\\\\"dir\\\\\\\",\n \\\\\\\"size\\\\\\\": 160,\n \\\\\\\"modified\\\\\\\": \\\\\\\"2026-01-07T06:22:12.488582817-06:00\\\\\\\"\n },\n {\n \\\\\\\"name\\\\\\\": \\\\\\\"benches\\\\\\\",\n \\\\\\\"type\\\\\\\": \\\\\\\"dir\\\\\\\",\n \\\\\\\"size\\\\\\\": 128,\n \\\\\\\"modified\\\\\\\": \\\\\\\"2026-05-21T08:31:29.486599444-05:00\\\\\\\"\n }\n]\\\"\"}"
}
],
"isError": false
}
}{
"jsonrpc": "2.0",
"id": 3,
"result": {
"content": [
{
"type": "text",
"text": "{cwd:/Users/fdncred/src/nushell,history_index:0,timestamp:2026-06-30T06:52:27.000000+00:00,output:{a:1,b:[2,3]}}"
}
],
"structuredContent": {
"cwd": "/Users/fdncred/src/nushell",
"history_index": 0,
"timestamp": "2026-06-30T06:52:27.000000+00:00",
"output": {
"a": 1,
"b": [2, 3]
}
},
"isError": false
}
}Improved error messages when negative indices are used in cell paths.
Previously, negative indices could produce misleading "Row number too large" errors (e.g., [["foo", "bar"], ["foo", "baz"]] | get 0.-1) or generic "NeedsPositiveValue" errors (e.g., ["foo", "bar", "baz"] | get (-1)).
Now Nushell reports clear messages such as "negative index is not supported in cell path" or "can't convert negative number to cell path".
oneof with a closure without pipe oneof<..., table>In some specific cases, the parser could reject values provided to oneof typed parameters, such as rejecting table literal syntax for oneof<.., table>:
> def cmd [--flag: oneof<record, table>] {}
> cmd --flag [[foo]; 1]
Error: nu::parser::parse_mismatch_with_full_string_msg
× Parse mismatch during operation.
╭─[repl_entry #17:1:12]
1 │ cmd --flag [[foo]; 1]
· ─────┬────
· ╰── expected oneof<record, table>
╰────
The parser no longer considers such code invalid.
oneof<..., closure>Arguments of type oneof<..., closure> could only be parsed as closures if they had a parameter list ({|| }).
> def cmd [--flag: oneof<list, closure>] {}
> cmd --flag {|| ls }
> cmd --flag { ls }
Error: nu::parser::parse_mismatch_with_full_string_msg
× Parse mismatch during operation.
╭─[repl_entry #21:1:12]
1 │ cmd --flag { ls }
· ───┬──
· ╰── expected non-block value: oneof<list<any>, closure()>
╰────
The parser now accepts calls like cmd --flag { ls }
header_on_separator Alignment issues in table when header_on_separator is on are fixed:
╭─#─┬─align──┬────val─────╮
│ 0 │ _ │ __________ │
│ 1 │ left │ a │
│ 2 │ right │ 0 │
│ 3 │ left │ a │
│ 4 │ center │ ∅ │
│ 5 │ left │ a │
│ 6 │ center │ ∅ │
│ 7 │ right │ 0 │
╰───┴────────┴────────────╯╭─#─┬─align──┬────val─────╮
│ 0 │ _ │ __________ │
│ 1 │ left │ a │
│ 2 │ right │ 0 │
│ 3 │ left │ a │
│ 4 │ center │ ∅ │
│ 5 │ left │ a │
│ 6 │ center │ ∅ │
│ 7 │ right │ 0 │
╰───┴────────┴────────────╯--full-reparse/-f to run command The run command now supports --full-reparse/-f, making workflows like watch . -g *.nu | where path ends-with test.nu | each -f { run -f ./test.nu } work as expected. Running scripts repeatedly is also fixed: starting with nu -n, then running run toolkit.nu twice now works, whereas previously a caching bug prevented the same file from being run more than once.
Float ranges without an explicit step now use more natural fractional steps, so 0.1..0.3 yields 0.1, 0.2, 0.3 instead of just 0.1. This also means 0.1..0.3 | last now correctly returns 0.3. Float range values are now rounded to match the step's decimal precision, which removes floating-point display artifacts like 0.30000000000000004 from serialized output.
> 0.1..0.3 | last
0.1
> 0.1..0.3 | to json -r
[0.1]> 0.1..0.3 | last
0.3
> 0.1..0.3 | to json -r
[0.1,0.2,0.3]
> 0.001..0.005 | to json -r
[0.001,0.002,0.003,0.004,0.005]
> 0.0..0.5 | to json -r
[0.0,0.1,0.2,0.3,0.4,0.5]stor open returns the custom value SQLiteDatabase, yet its signature had sqlite-in-memory as the return type. This causes problems with the enforce-runtime-annotations experimental option:
> let db = stor open
Error: nu::shell::cant_convert
× Can't convert to sqlite-in-memory.
╭─[repl_entry #2:1:10]
1 │ let db = stor open
· ────┬────
· ╰── can't convert SQLiteDatabase to sqlite-in-memory
╰────
The return type has been corrected to SQLiteDatabase.
Descriptions of module completions for use were previously not formatted. Instead of what you would get from help modules ..., the descriptions were the modules' doc comment as it appeared in the file (# and leading indent not removed).
Before:
> use c
╭───────╮ ╭─────────────────────────────────────────────╮
│clip │ │# Commands for interacting with the system │
╰───────╯ │clipboard # # > These commands require your │
│terminal to support OSC 52 # > Terminal │
│multiplexers such as screen, tmux, zellij etc│
│may interfere with this command │
╰─────────────────────────────────────────────╯After:
> use c
╭───────╮ ╭────────────────────────────────────────╮
│clip │ │Commands for interacting with the system│
╰───────╯ │clipboard │
╰────────────────────────────────────────╯Shadowing variables used in aliases could cause the alias to stop working:
> let x = 10
> alias xx = print $x
> xx
10
> let x = 20
> xx
Error: nu::shell::variable_not_found
× Variable not found
╭─[repl_entry #32:1:18]
1 │ alias xx = print $x
· ─┬
· ╰── variable not found
╰────
Now aliases continue working with the original value of the variable:
> let x = 10
> alias xx = print $x
> xx
10
> let x = 20
> xx
10-- option parsing Fixed a bug with how -- was handled with --wrapped custom commands. Now you can do things like this:
> def --wrapped example [--my-flag: string ...rest] {
let rest = $rest | skip 1 # skip the '--' delimiter
$rest | each { {type: ($in | describe) value: $in} } | print
print $"--my-flag='($my_flag)'"
}
> example --my-flag="hi" -- true false 001 --my-flag="goodbye"
╭───┬────────┬───────────────────╮
│ # │ type │ value │
├───┼────────┼───────────────────┤
│ 0 │ string │ true │
│ 1 │ string │ false │
│ 2 │ string │ 001 │
│ 3 │ string │ --my-flag=goodbye │
╰───┴────────┴───────────────────╯
--my-flag='hi'from json Ctrl+C is more responsive in from json, from kdl, and from xml. They now respond promptly to Ctrl+C when parsing large files. Pressing Ctrl+C during parsing will interrupt the command and return to the prompt. It will also produce better parse-error messages and should now show a focused snippet of the source around the error location instead of dumping the entire file content.
flatten This pr fixes flatten renaming column behavior
> [[b, a]; [[[a]; [9]], 1]]
| flatten -a b
╭───┬───╮
│ # │ a │
├───┼───┤
│ 0 │ 1 │
╰───┴───╯> [[b, a]; [[[a]; [9]], 1]]
| flatten -a b
╭───┬─────┬───╮
│ # │ b_a │ a │
├───┼─────┼───┤
│ 0 │ 9 │ 1 │
╰───┴─────┴───╯{} as closure, not record Empty braces ({}) as row conditions are now parsed as empty closures rather than empty records.
> 1..3 | where {}
Error: nu::parser::type_mismatch
× Type mismatch.
╭─[repl_entry #39:1:14]
1 │ 1..3 | where {}
· ─┬
· ╰── expected bool, found record
╰────
> 1..3 | where {}
╭────────────╮
│ empty list │
╰────────────╯idx search Allow idx search 'Lyrics[' to work by taking [ literally instead of assuming it's a glob. Also added a couple examples explaining how idx search should work.
idx search 'arr[0]'idx search pattern */tests/*idx search pattern *.{rs,js}Respect the table mode when passed in on the cli when running script so you can do things like this.
> '1..3 | each { { number: $in, comment: "hello" } }' | save script.nu
> nu -m "none" script.nu
# number comment
0 1 hello
1 2 hello
2 3 helloinput list streaming and rendering input list no longer waits for streamed input to be fully collected before showing the first paint. It now stays responsive while items stream in and shows a live-updating count of collected items in the footer. In table mode, header cells now expand properly along with their columns when scrolling through the list, and control characters like tabs no longer throw off UI alignment during rendering.
nu --plugins handling The nushell cli parsing now supports the --plugins parameter in these variations
nu --plugins /path/to/nu_plugin_one --plugins /path/to/nu_plugin_two # multiple --plugins params
nu --plugins '[/path/to/nu_plugin_one, /path/to/nu_plugin_two]' # comma separation
nu --plugins '[/path/to/nu_plugin_one /path/to/nu_plugin_two]' # no comma separation
nu --plugins '["/path/to plugin/nu_plugin_one", "/path/to plugin/nu_plugin_two"]' # quoted paths, with or without commascommandline edit did not update the commandline in REPL on 0.113.0. commandline edit now correctly updates the visible prompt buffer again. (#18301)math abs would crash Nushell when given the minimum 64-bit integer or duration; it now reports an overflow error instead. (#18275)bytes index-of doesn't panic on empty patterns anymore (#18254)explore config if you supply --output <file> when you're editing the config i.e. not using explore config in json mode, it will fail fast and let you know that isn't possible instead of you finding that out later. (#18327)$"($env.pwd)/foo", ~/($bar)) to a flag or argument of type glob they will be converted automatically just like with regular strings instead of throwing a type error. (#18263)lines failed on input containing invalid UTF-8. By default, lines now replaces invalid UTF-8 bytes and continues processing, and lines --strict can be used to fail on invalid UTF-8. (#18261)input list --fuzzy truncating ANSI-styled options too early because formatting escape sequences were counted as visible width. (#18340)uniq-by silently collapsed every value into a single row when given a list whose elements are not records/tables. uniq-by now reports an error in that case (e.g. [1 2 2] | uniq-by foo), and it also reports a missing column when any record in the input lacks the requested column, not just the first one. (#18309)bits shl and bits shr now default to an 8-byte word size when --number-bytes is not provided, so shifts such as 1 | bits shl 20 no longer fail because the input was auto-sized to one byte. (#18277)[.foons] will cause parser error. (#18363)glob c:\apps\*. (#18367)--wrapped where numbers and strings passed to externals as glob now everything is passed as string or glob or creates an error. (#18372)dc-glob, glob expands ~ as expected now. (#18373)glob .cargo/bin/nu* nor %ls .cargo/bin/nu* wouldn't work. (#18375)--ignore-case/-i when using the dc-glob experimental-option. (#18376)glob ~/.cargo/bin/nu* from within ~/Downloads and making sure the results start with /Users/<username>/.cargo/bin/nu. Previously they incorrectly started with /Users/<username>/Downloads/nu* which was completely wrong. Thanks @Tyarel8 ! (#18378)nu --mcp agent tool instructions. (#18413)str index-of --grapheme-clusters when the search string matched inside a multi–code-point grapheme cluster (flag emoji, ZWJ sequence, skin-tone modifier, or combining mark); it now returns -1. (#18418)exit in the repl will no longer use a different logic and will call the exit command normally. Furthermore, any invocation of exit (e.g. inside a keybind of custom command) will restore the terminal cursor to the state it was before rather than only in the aforementioned case. (#18389)[] | insert 0.0 1, produced a confusing error mentioning 18446744073709551615. It now reports a clear "Row number too large (empty content)" error. (#18463)register_uninitialized) when the right-hand side of an and/or/comparison collects $in, for example ... | where $in.x > 0 and not ($in.x > 5). (#18465)--log-include or --log-exclude now accept module or target name correctly. Previously it accepted log level instead. (#18464)math avg, math sum, and other math commands on tables produced a generic "Unable to give a result with this input" error instead of the real underlying error. List and table inputs now produce consistent error messages. (#18480)help aliases now correctly shows the alias name (not the target command name) for aliases to internal commands. (#18483)Thanks to all the contributors below for helping us solve issues, improve documentation, refactor code, and more! 🙏
| author | change | link | |
]]>@cptpiepmatz:
@fdncred:
nu --plugins handling (#18500)idx find and idx search to show hits relative to cwd (#18477)semver as a custom value type with support commands (#18473)@Alb-O:
@pheenty:
@yertto:
@Himanshu121865:
@Bahex:
@cacdu:
@kronberger-droid:
@jlcrochet:
@Juhan280:
@app/dependabot:
@cptpiepmatz:
@fdncred:
nu --plugins handling (#18500)idx find and idx search to show hits relative to cwd (#18477)semver as a custom value type with support commands (#18473)@Alb-O:
@pheenty:
@yertto:
@Himanshu121865:
@Bahex:
@cacdu:
@kronberger-droid:
@jlcrochet:
@Juhan280:
@app/dependabot:
@kronberger-droid:
@MerlijnW70:
@tonythethompson:
@rossnomann:
@fdncred:
perf so we can more easily see only performance logging (#18468)idx search 'Lyrics[' bug (#18447)is-terminal to detect redirection (#18443)@maximilize:
@Juhan280:
@hexbinoct:
@flying-sheep:
@cptpiepmatz:
@Tyarel8:
@Bahex:
@stuartcarnie:
@mkatychev:
@app/dependabot:
@fdncred:
perf so we can more easily see only performance logging (#18468)idx search 'Lyrics[' bug (#18447)is-terminal to detect redirection (#18443)@maximilize:
@Juhan280:
@hexbinoct:
@flying-sheep:
@cptpiepmatz:
@Tyarel8:
@Bahex:
@stuartcarnie:
@mkatychev:
@app/dependabot:
@NotTheDr01ds:
@sharpchen:
@kiil:
@app/dependabot:
@fdncred:
@reubeno:
@ldsands:
@kaascevich:
@DeadBranches:
@cptpiepmatz:
@Bahex:
@leeewee:
@fdncred:
@Alb-O:
@WindSoilder:
@kronberger-droid:
@Tyarel8:
@madjar:
@blindFS:
@app/dependabot:
@cptpiepmatz:
@Bahex:
@leeewee:
@fdncred:
@Alb-O:
@WindSoilder:
@kronberger-droid:
@Tyarel8:
@madjar:
@blindFS:
@app/dependabot:
@casedami:
@NotTheDr01ds:
@ethan-hawksley:
@blindFS:
@stormasm:
@kronberger-droid:
@philocalyst:
@ymcx:
@ayax79:
@fdncred:
/some/path/wildcard* (#18375)--wrapped to not always convert to glob (#18372)--context to be one of int or range (#18368)idx search command (#18366)plugin list (#18357)@casedami:
@cptpiepmatz:
@Ashif-107:
@blindFS:
@Bahex:
@zelshahawy:
@app/dependabot:
@ayax79:
@fdncred:
/some/path/wildcard* (#18375)--wrapped to not always convert to glob (#18372)--context to be one of int or range (#18368)idx search command (#18366)plugin list (#18357)@casedami:
@cptpiepmatz:
@Ashif-107:
@blindFS:
@Bahex:
@zelshahawy:
@app/dependabot:
@kronberger-droid:
@Alb-O:
@casedami:
@reubeno:
@a-rookie-of-C-language:
@guluo2016:
@NotTheDr01ds:
@Bahex:
expand and header_on_separator (#18344)get_locale_from_env_vars in nu-utils (#18313)@puneetdixit200:
@fdncred:
--full-reparse/-f to run command (#18339)--output from being used in config_mode in explore config command (#18327)uu_* deps to 0.9.0 (#18322)ignore to really ignore 😃 (#18306)-- end-of-options delimiter (#18299)@LeonidasZhak:
@dmtrKovalenko:
@Mrfiregem:
@ian-h-chamberlain:
@zhiburt:
@stormasm:
@ayax79:
@cptpiepmatz:
@hustcer:
@sholderbach:
@guluo2016:
@cho-minsung:
@app/dependabot:
@Bahex:
expand and header_on_separator (#18344)get_locale_from_env_vars in nu-utils (#18313)@puneetdixit200:
@fdncred:
--full-reparse/-f to run command (#18339)--output from being used in config_mode in explore config command (#18327)uu_* deps to 0.9.0 (#18322)ignore to really ignore 😃 (#18306)-- end-of-options delimiter (#18299)@LeonidasZhak:
@dmtrKovalenko:
@Mrfiregem:
@ian-h-chamberlain:
@zhiburt:
@stormasm:
@ayax79:
@cptpiepmatz:
@hustcer:
@sholderbach:
@guluo2016:
@cho-minsung:
@app/dependabot:
@kronberger-droid:
@skull-squadron:
@kaascevich:
@Hentioe:
@cptpiepmatz:
@NotTheDr01ds:
@samoylovfp:
idx command family, and a REPL regression in commandline edit.
Nu 0.113.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
idx command family, and a REPL regression in commandline edit.
Nu 0.113.1 is available as pre-built binaries or from crates.io. If you have Rust installed you can install it using cargo install nu.
As part of this release, we also publish a set of optional plugins you can install and use with Nushell.
The previous emitter always quoted string values, which produced valid YAML but was more conservative than necessary and less aligned with YAML 1.2 plain-scalar rules.
This change makes output:
value: 'off'
path: '/dev/stdout'
listen: '0.0.0.0:8444,0.0.0.0:8445 ssl'
name: 'kong'
kind: 'Deployment'value: 'off'
path: /dev/stdout
listen: 0.0.0.0:8444,0.0.0.0:8445 ssl
name: kong
kind: Deploymentstring: |-
Hello
worldidx dependency update and command reliability improvements Update the idx fff-search dependency and fix bugs.
idx init . --wait now properly blocksidx init now has a --follow-links switchidx status reports accurate counts after indexingidx import fully restores snapshot for queries — The idx import command now properly restores snapshots to enable all query operations. Imported snapshots remain queryable even if the original project files have been deleted or moved.idx files works on imported snapshotsidx files has a query parameter now for easy searching (renamed from path for consistency)idx dirs works on imported snapshotsidx dirs has a query parameter now for easy searchingidx find (fuzzy search) works on imported snapshotsidx search (content search) works on imported snapshotsFiles are ignored during indexing due to the use of the ignore crate. These are the rules for ignore. There's no way to override this in the current version.
commandline edit did not update the commandline in REPL on 0.113.0. commandline edit now correctly updates the visible prompt buffer again. (#18301)Thanks to all the contributors below for helping us solve issues, improve documentation, refactor code, and more! 🙏
| author | change | link | |
]]>