Overhaul $in expressions (#13357)

# Description

This grew quite a bit beyond its original scope, but I've tried to make
`$in` a bit more consistent and easier to work with.

Instead of the parser generating calls to `collect` and creating
closures, this adds `Expr::Collect` which just evaluates in the same
scope and doesn't require any closure.

When `$in` is detected in an expression, it is replaced with a new
variable (also called `$in`) and wrapped in `Expr::Collect`. During
eval, this expression is evaluated directly, with the input and with
that new variable set to the collected value.

Other than being faster and less prone to gotchas, it also makes it
possible to typecheck the output of an expression containing `$in`,
which is nice. This is a breaking change though, because of the lack of
the closure and because now typechecking will actually happen. Also, I
haven't attempted to typecheck the input yet.

The IR generated now just looks like this:

```gas
collect        %in
clone          %tmp, %in
store-variable $in, %tmp
# %out <- ...expression... <- %in
drop-variable  $in
```

(where `$in` is the local variable created for this collection, and not
`IN_VARIABLE_ID`)

which is a lot better than having to create a closure and call `collect
--keep-env`, dealing with all of the capture gathering and allocation
that entails. Ideally we can also detect whether that input is actually
needed, so maybe we don't have to clone, but I haven't tried to do that
yet. Theoretically now that the variable is a unique one every time, it
should be possible to give it a type - I just don't know how to
determine that yet.

On top of that, I've also reworked how `$in` works in pipeline-initial
position. Previously, it was a little bit inconsistent. For example,
this worked:

```nushell
> 3 | do { let x = $in; let y = $in; print $x $y }
3
3
```

However, this causes a runtime variable not found error on the second
`$in`:

```nushell
> def foo [] { let x = $in; let y = $in; print $x $y }; 3 | foo
Error: nu:🐚:variable_not_found

  × Variable not found
   ╭─[entry #115:1:35]
 1 │ def foo [] { let x = $in; let y = $in; print $x $y }; 3 | foo
   ·                                   ─┬─
   ·                                    ╰── variable not found
   ╰────
```

I've fixed this by making the first element `$in` detection *always*
happen at the block level, so if you use `$in` in pipeline-initial
position anywhere in a block, it will collect with an implicit
subexpression around the whole thing, and you can then use that `$in`
more than once. In doing this I also rewrote `parse_pipeline()` and
hopefully it's a bit more straightforward and possibly more efficient
too now.

Finally, I've tried to make `let` and `mut` a lot more straightforward
with how they handle the rest of the pipeline, and using a redirection
with `let`/`mut` now does what you'd expect if you assume that they
consume the whole pipeline - the redirection is just processed as
normal. These both work now:

```nushell
let x = ^foo err> err.txt
let y = ^foo out+err>| str length
```

It was previously possible to accomplish this with a subexpression, but
it just seemed like a weird gotcha that you couldn't do it. Intuitively,
`let` and `mut` just seem to take the whole line.

- closes #13137

# User-Facing Changes
- `$in` will behave more consistently with blocks and closures, since
the entire block is now just wrapped to handle it if it appears in the
first pipeline element
- `$in` no longer creates a closure, so what can be done within an
expression containing `$in` is less restrictive
- `$in` containing expressions are now type checked, rather than just
resulting in `any`. However, `$in` itself is still `any`, so this isn't
quite perfect yet
- Redirections are now allowed in `let` and `mut` and behave pretty much
how you'd expect

# Tests + Formatting
Added tests to cover the new behaviour.

# After Submitting
- [ ] release notes (definitely breaking change)

crates/nu-cli/src/syntax_highlight.rs

@@ -429,6 +429,14 @@ fn find_matching_block_end_in_expr(
                 )
             }),
 
+            Expr::Collect(_, expr) => find_matching_block_end_in_expr(
+                line,
+                working_set,
+                expr,
+                global_span_offset,
+                global_cursor_offset,
+            ),
+
             Expr::Block(block_id)
             | Expr::Closure(block_id)
             | Expr::RowCondition(block_id)

crates/nu-cmd-lang/src/example_support.rs

@@ -4,7 +4,7 @@ use nu_protocol::{
     ast::Block,
     debugger::WithoutDebug,
     engine::{StateDelta, StateWorkingSet},
-    Range,
+    report_error_new, Range,
 };
 use std::{
     sync::Arc,
@@ -124,7 +124,10 @@ pub fn eval_block(
 
     nu_engine::eval_block::<WithoutDebug>(engine_state, &mut stack, &block, input)
         .and_then(|data| data.into_value(Span::test_data()))
-        .unwrap_or_else(|err| panic!("test eval error in `{}`: {:?}", "TODO", err))
+        .unwrap_or_else(|err| {
+            report_error_new(engine_state, &err);
+            panic!("test eval error in `{}`: {:?}", "TODO", err)
+        })
 }
 
 pub fn check_example_evaluates_to_expected_output(

crates/nu-command/src/example_test.rs

@@ -31,7 +31,7 @@ mod test_examples {
         check_example_evaluates_to_expected_output,
         check_example_input_and_output_types_match_command_signature,
     };
-    use nu_cmd_lang::{Break, Echo, If, Let, Mut};
+    use nu_cmd_lang::{Break, Describe, Echo, If, Let, Mut};
     use nu_protocol::{
         engine::{Command, EngineState, StateWorkingSet},
         Type,
@@ -81,6 +81,7 @@ mod test_examples {
             working_set.add_decl(Box::new(Break));
             working_set.add_decl(Box::new(Date));
             working_set.add_decl(Box::new(Default));
+            working_set.add_decl(Box::new(Describe));
             working_set.add_decl(Box::new(Each));
             working_set.add_decl(Box::new(Echo));
             working_set.add_decl(Box::new(Enumerate));

crates/nu-command/tests/commands/redirection.rs

@@ -439,3 +439,37 @@ fn no_duplicate_redirection() {
         );
     });
 }
+
+#[rstest::rstest]
+#[case("let", "out>")]
+#[case("let", "err>")]
+#[case("let", "out+err>")]
+#[case("mut", "out>")]
+#[case("mut", "err>")]
+#[case("mut", "out+err>")]
+fn file_redirection_in_let_and_mut(#[case] keyword: &str, #[case] redir: &str) {
+    Playground::setup("file redirection in let and mut", |dirs, _| {
+        let actual = nu!(
+            cwd: dirs.test(),
+            format!("$env.BAZ = 'foo'; {keyword} v = nu --testbin echo_env_mixed out-err BAZ BAZ {redir} result.txt")
+        );
+        assert!(actual.status.success());
+        assert!(file_contents(dirs.test().join("result.txt")).contains("foo"));
+    })
+}
+
+#[rstest::rstest]
+#[case("let", "err>|", "foo3")]
+#[case("let", "out+err>|", "7")]
+#[case("mut", "err>|", "foo3")]
+#[case("mut", "out+err>|", "7")]
+fn pipe_redirection_in_let_and_mut(
+    #[case] keyword: &str,
+    #[case] redir: &str,
+    #[case] output: &str,
+) {
+    let actual = nu!(
+        format!("$env.BAZ = 'foo'; {keyword} v = nu --testbin echo_env_mixed out-err BAZ BAZ {redir} str length; $v")
+    );
+    assert_eq!(actual.out, output);
+}

crates/nu-engine/src/compile/builder.rs

@@ -204,6 +204,7 @@ impl BlockBuilder {
             Instruction::Drain { src } => allocate(&[*src], &[]),
             Instruction::LoadVariable { dst, var_id: _ } => allocate(&[], &[*dst]),
             Instruction::StoreVariable { var_id: _, src } => allocate(&[*src], &[]),
+            Instruction::DropVariable { var_id: _ } => Ok(()),
             Instruction::LoadEnv { dst, key: _ } => allocate(&[], &[*dst]),
             Instruction::LoadEnvOpt { dst, key: _ } => allocate(&[], &[*dst]),
             Instruction::StoreEnv { key: _, src } => allocate(&[*src], &[]),

crates/nu-engine/src/compile/expression.rs

@@ -171,6 +171,27 @@ pub(crate) fn compile_expression(
                 Err(CompileError::UnsupportedOperatorExpression { span: op.span })
             }
         }
+        Expr::Collect(var_id, expr) => {
+            let store_reg = if let Some(in_reg) = in_reg {
+                // Collect, clone, store
+                builder.push(Instruction::Collect { src_dst: in_reg }.into_spanned(expr.span))?;
+                builder.clone_reg(in_reg, expr.span)?
+            } else {
+                // Just store nothing in the variable
+                builder.literal(Literal::Nothing.into_spanned(Span::unknown()))?
+            };
+            builder.push(
+                Instruction::StoreVariable {
+                    var_id: *var_id,
+                    src: store_reg,
+                }
+                .into_spanned(expr.span),
+            )?;
+            compile_expression(working_set, builder, expr, redirect_modes, in_reg, out_reg)?;
+            // Clean it up afterward
+            builder.push(Instruction::DropVariable { var_id: *var_id }.into_spanned(expr.span))?;
+            Ok(())
+        }
         Expr::Subexpression(block_id) => {
             let block = working_set.get_block(*block_id);
             compile_block(working_set, builder, block, redirect_modes, in_reg, out_reg)

crates/nu-engine/src/eval.rs

@@ -10,8 +10,8 @@ use nu_protocol::{
     debugger::DebugContext,
     engine::{Closure, EngineState, Redirection, Stack, StateWorkingSet},
     eval_base::Eval,
-    ByteStreamSource, Config, FromValue, IntoPipelineData, OutDest, PipelineData, ShellError, Span,
+    ByteStreamSource, Config, DataSource, FromValue, IntoPipelineData, OutDest, PipelineData,
-    Spanned, Type, Value, VarId, ENV_VARIABLE_ID,
+    PipelineMetadata, ShellError, Span, Spanned, Type, Value, VarId, ENV_VARIABLE_ID,
 };
 use nu_utils::IgnoreCaseExt;
 use std::{fs::OpenOptions, path::PathBuf, sync::Arc};
@@ -259,6 +259,10 @@ pub fn eval_expression_with_input<D: DebugContext>(
             input = eval_external(engine_state, stack, head, args, input)?;
         }
 
+        Expr::Collect(var_id, expr) => {
+            input = eval_collect::<D>(engine_state, stack, *var_id, expr, input)?;
+        }
+
         Expr::Subexpression(block_id) => {
             let block = engine_state.get_block(*block_id);
             // FIXME: protect this collect with ctrl-c
@@ -605,6 +609,44 @@ pub fn eval_block<D: DebugContext>(
     Ok(input)
 }
 
+pub fn eval_collect<D: DebugContext>(
+    engine_state: &EngineState,
+    stack: &mut Stack,
+    var_id: VarId,
+    expr: &Expression,
+    input: PipelineData,
+) -> Result<PipelineData, ShellError> {
+    // Evaluate the expression with the variable set to the collected input
+    let span = input.span().unwrap_or(Span::unknown());
+
+    let metadata = match input.metadata() {
+        // Remove the `FilePath` metadata, because after `collect` it's no longer necessary to
+        // check where some input came from.
+        Some(PipelineMetadata {
+            data_source: DataSource::FilePath(_),
+            content_type: None,
+        }) => None,
+        other => other,
+    };
+
+    let input = input.into_value(span)?;
+
+    stack.add_var(var_id, input.clone());
+
+    let result = eval_expression_with_input::<D>(
+        engine_state,
+        stack,
+        expr,
+        // We still have to pass it as input
+        input.into_pipeline_data_with_metadata(metadata),
+    )
+    .map(|(result, _failed)| result);
+
+    stack.remove_var(var_id);
+
+    result
+}
+
 pub fn eval_subexpression<D: DebugContext>(
     engine_state: &EngineState,
     stack: &mut Stack,
@@ -729,6 +771,18 @@ impl Eval for EvalRuntime {
         eval_external(engine_state, stack, head, args, PipelineData::empty())?.into_value(span)
     }
 
+    fn eval_collect<D: DebugContext>(
+        engine_state: &EngineState,
+        stack: &mut Stack,
+        var_id: VarId,
+        expr: &Expression,
+    ) -> Result<Value, ShellError> {
+        // It's a little bizarre, but the expression can still have some kind of result even with
+        // nothing input
+        eval_collect::<D>(engine_state, stack, var_id, expr, PipelineData::empty())?
+            .into_value(expr.span)
+    }
+
     fn eval_subexpression<D: DebugContext>(
         engine_state: &EngineState,
         stack: &mut Stack,

crates/nu-engine/src/eval_ir.rs

@@ -6,9 +6,9 @@ use nu_protocol::{
     debugger::DebugContext,
     engine::{Argument, Closure, EngineState, ErrorHandler, Matcher, Redirection, Stack},
     ir::{Call, DataSlice, Instruction, IrAstRef, IrBlock, Literal, RedirectMode},
-    record, ByteStreamSource, DeclId, ErrSpan, Flag, IntoPipelineData, IntoSpanned, ListStream,
+    record, ByteStreamSource, DataSource, DeclId, ErrSpan, Flag, IntoPipelineData, IntoSpanned,
-    OutDest, PipelineData, PositionalArg, Range, Record, RegId, ShellError, Signals, Signature,
+    ListStream, OutDest, PipelineData, PipelineMetadata, PositionalArg, Range, Record, RegId,
-    Span, Spanned, Type, Value, VarId, ENV_VARIABLE_ID,
+    ShellError, Signals, Signature, Span, Spanned, Type, Value, VarId, ENV_VARIABLE_ID,
 };
 use nu_utils::IgnoreCaseExt;
 
@@ -345,6 +345,10 @@ fn eval_instruction<D: DebugContext>(
             ctx.stack.add_var(*var_id, value);
             Ok(Continue)
         }
+        Instruction::DropVariable { var_id } => {
+            ctx.stack.remove_var(*var_id);
+            Ok(Continue)
+        }
         Instruction::LoadEnv { dst, key } => {
             let key = ctx.get_str(*key, *span)?;
             if let Some(value) = get_env_var_case_insensitive(ctx, key) {
@@ -1341,9 +1345,19 @@ fn get_env_var_name_case_insensitive<'a>(ctx: &mut EvalContext<'_>, key: &'a str
 }
 
 /// Helper to collect values into [`PipelineData`], preserving original span and metadata
+///
+/// The metadata is removed if it is the file data source, as that's just meant to mark streams.
 fn collect(data: PipelineData, fallback_span: Span) -> Result<PipelineData, ShellError> {
     let span = data.span().unwrap_or(fallback_span);
-    let metadata = data.metadata();
+    let metadata = match data.metadata() {
+        // Remove the `FilePath` metadata, because after `collect` it's no longer necessary to
+        // check where some input came from.
+        Some(PipelineMetadata {
+            data_source: DataSource::FilePath(_),
+            content_type: None,
+        }) => None,
+        other => other,
+    };
     let value = data.into_value(span)?;
     Ok(PipelineData::Value(value, metadata))
 }

crates/nu-parser/src/flatten.rs

@@ -189,6 +189,9 @@ fn flatten_expression_into(
             ));
             flatten_expression_into(working_set, not, output);
         }
+        Expr::Collect(_, expr) => {
+            flatten_expression_into(working_set, expr, output);
+        }
         Expr::Closure(block_id) => {
             let outer_span = expr.span;
 

crates/nu-parser/src/lite_parser.rs

@@ -2,6 +2,7 @@
 //! can be parsed.
 
 use crate::{Token, TokenContents};
+use itertools::{Either, Itertools};
 use nu_protocol::{ast::RedirectionSource, ParseError, Span};
 use std::mem;
 
@@ -24,6 +25,15 @@ impl LiteRedirectionTarget {
             | LiteRedirectionTarget::Pipe { connector } => *connector,
         }
     }
+
+    pub fn spans(&self) -> impl Iterator<Item = Span> {
+        match *self {
+            LiteRedirectionTarget::File {
+                connector, file, ..
+            } => Either::Left([connector, file].into_iter()),
+            LiteRedirectionTarget::Pipe { connector } => Either::Right(std::iter::once(connector)),
+        }
+    }
 }
 
 #[derive(Debug, Clone)]
@@ -38,6 +48,17 @@ pub enum LiteRedirection {
     },
 }
 
+impl LiteRedirection {
+    pub fn spans(&self) -> impl Iterator<Item = Span> {
+        match self {
+            LiteRedirection::Single { target, .. } => Either::Left(target.spans()),
+            LiteRedirection::Separate { out, err } => {
+                Either::Right(out.spans().chain(err.spans()).sorted())
+            }
+        }
+    }
+}
+
 #[derive(Debug, Clone, Default)]
 pub struct LiteCommand {
     pub pipe: Option<Span>,
@@ -113,6 +134,14 @@ impl LiteCommand {
 
         Ok(())
     }
+
+    pub fn parts_including_redirection(&self) -> impl Iterator<Item = Span> + '_ {
+        self.parts.iter().copied().chain(
+            self.redirection
+                .iter()
+                .flat_map(|redirection| redirection.spans()),
+        )
+    }
 }
 
 #[derive(Debug, Clone, Default)]

crates/nu-parser/src/parser.rs

@@ -5299,6 +5299,7 @@ pub fn parse_expression(working_set: &mut StateWorkingSet, spans: &[Span]) -> Ex
             let mut block = Block::default();
             let ty = output.ty.clone();
             block.pipelines = vec![Pipeline::from_vec(vec![output])];
+            block.span = Some(Span::concat(spans));
 
             compile_block(working_set, &mut block);
 
@@ -5393,9 +5394,19 @@ pub fn parse_builtin_commands(
     match name {
         b"def" => parse_def(working_set, lite_command, None).0,
         b"extern" => parse_extern(working_set, lite_command, None),
-        b"let" => parse_let(working_set, &lite_command.parts),
+        b"let" => parse_let(
+            working_set,
+            &lite_command
+                .parts_including_redirection()
+                .collect::<Vec<Span>>(),
+        ),
         b"const" => parse_const(working_set, &lite_command.parts),
-        b"mut" => parse_mut(working_set, &lite_command.parts),
+        b"mut" => parse_mut(
+            working_set,
+            &lite_command
+                .parts_including_redirection()
+                .collect::<Vec<Span>>(),
+        ),
         b"for" => {
             let expr = parse_for(working_set, lite_command);
             Pipeline::from_vec(vec![expr])
@@ -5647,169 +5658,73 @@ pub(crate) fn redirecting_builtin_error(
     }
 }
 
-pub fn parse_pipeline(
+pub fn parse_pipeline(working_set: &mut StateWorkingSet, pipeline: &LitePipeline) -> Pipeline {
-    working_set: &mut StateWorkingSet,
+    let first_command = pipeline.commands.first();
-    pipeline: &LitePipeline,
+    let first_command_name = first_command
-    is_subexpression: bool,
+        .and_then(|command| command.parts.first())
-    pipeline_index: usize,
+        .map(|span| working_set.get_span_contents(*span));
-) -> Pipeline {
+
     if pipeline.commands.len() > 1 {
-        // Special case: allow `let` and `mut` to consume the whole pipeline, eg) `let abc = "foo" | str length`
+        // Special case: allow "let" or "mut" to consume the whole pipeline, if this is a pipeline
-        if let Some(&first) = pipeline.commands[0].parts.first() {
+        // with multiple commands
-            let first = working_set.get_span_contents(first);
+        if matches!(first_command_name, Some(b"let" | b"mut")) {
-            if first == b"let" || first == b"mut" {
+            // Merge the pipeline into one command
-                let name = if first == b"let" { "let" } else { "mut" };
+            let first_command = first_command.expect("must be Some");
-                let mut new_command = LiteCommand {
-                    comments: vec![],
-                    parts: pipeline.commands[0].parts.clone(),
-                    pipe: None,
-                    redirection: None,
-                };
 
-                if let Some(redirection) = pipeline.commands[0].redirection.as_ref() {
+            let remainder_span = first_command
-                    working_set.error(redirecting_builtin_error(name, redirection));
+                .parts_including_redirection()
-                }
+                .skip(3)
+                .chain(
+                    pipeline.commands[1..]
+                        .iter()
+                        .flat_map(|command| command.parts_including_redirection()),
+                )
+                .reduce(Span::append);
 
-                for element in &pipeline.commands[1..] {
+            let parts = first_command
-                    if let Some(redirection) = pipeline.commands[0].redirection.as_ref() {
+                .parts
-                        working_set.error(redirecting_builtin_error(name, redirection));
+                .iter()
-                    } else {
+                .take(3) // the let/mut start itself
-                        new_command.parts.push(element.pipe.expect("pipe span"));
+                .copied()
-                        new_command.comments.extend_from_slice(&element.comments);
+                .chain(remainder_span) // everything else
-                        new_command.parts.extend_from_slice(&element.parts);
+                .collect();
-                    }
-                }
 
-                // if the 'let' is complete enough, use it, if not, fall through for now
+            let comments = pipeline
-                if new_command.parts.len() > 3 {
+                .commands
-                    let rhs_span = Span::concat(&new_command.parts[3..]);
+                .iter()
+                .flat_map(|command| command.comments.iter())
+                .copied()
+                .collect();
 
-                    new_command.parts.truncate(3);
+            let new_command = LiteCommand {
-                    new_command.parts.push(rhs_span);
+                pipe: None,
-
+                comments,
-                    let mut pipeline = parse_builtin_commands(working_set, &new_command);
+                parts,
-
+                redirection: None,
-                    if pipeline_index == 0 {
+            };
-                        let let_decl_id = working_set.find_decl(b"let");
+            parse_builtin_commands(working_set, &new_command)
-                        let mut_decl_id = working_set.find_decl(b"mut");
-                        for element in pipeline.elements.iter_mut() {
-                            if let Expr::Call(call) = &element.expr.expr {
-                                if Some(call.decl_id) == let_decl_id
-                                    || Some(call.decl_id) == mut_decl_id
-                                {
-                                    // Do an expansion
-                                    if let Some(Expression {
-                                        expr: Expr::Block(block_id),
-                                        ..
-                                    }) = call.positional_iter().nth(1)
-                                    {
-                                        let block = working_set.get_block(*block_id);
-
-                                        if let Some(element) = block
-                                            .pipelines
-                                            .first()
-                                            .and_then(|p| p.elements.first())
-                                            .cloned()
-                                        {
-                                            if element.has_in_variable(working_set) {
-                                                let element = wrap_element_with_collect(
-                                                    working_set,
-                                                    &element,
-                                                );
-                                                let block = working_set.get_block_mut(*block_id);
-                                                block.pipelines[0].elements[0] = element;
-                                            }
-                                        }
-                                    }
-                                    continue;
-                                } else if element.has_in_variable(working_set) && !is_subexpression
-                                {
-                                    *element = wrap_element_with_collect(working_set, element);
-                                }
-                            } else if element.has_in_variable(working_set) && !is_subexpression {
-                                *element = wrap_element_with_collect(working_set, element);
-                            }
-                        }
-                    }
-
-                    return pipeline;
-                }
-            }
-        }
-
-        let mut elements = pipeline
-            .commands
-            .iter()
-            .map(|element| parse_pipeline_element(working_set, element))
-            .collect::<Vec<_>>();
-
-        if is_subexpression {
-            for element in elements.iter_mut().skip(1) {
-                if element.has_in_variable(working_set) {
-                    *element = wrap_element_with_collect(working_set, element);
-                }
-            }
         } else {
-            for element in elements.iter_mut() {
+            // Parse a normal multi command pipeline
-                if element.has_in_variable(working_set) {
+            let elements: Vec<_> = pipeline
-                    *element = wrap_element_with_collect(working_set, element);
+                .commands
-                }
+                .iter()
-            }
+                .enumerate()
-        }
+                .map(|(index, element)| {
-
+                    let element = parse_pipeline_element(working_set, element);
-        Pipeline { elements }
+                    // Handle $in for pipeline elements beyond the first one
-    } else {
+                    if index > 0 && element.has_in_variable(working_set) {
-        if let Some(&first) = pipeline.commands[0].parts.first() {
+                        wrap_element_with_collect(working_set, element.clone())
-            let first = working_set.get_span_contents(first);
+                    } else {
-            if first == b"let" || first == b"mut" {
+                        element
-                if let Some(redirection) = pipeline.commands[0].redirection.as_ref() {
-                    let name = if first == b"let" { "let" } else { "mut" };
-                    working_set.error(redirecting_builtin_error(name, redirection));
-                }
-            }
-        }
-
-        let mut pipeline = parse_builtin_commands(working_set, &pipeline.commands[0]);
-
-        let let_decl_id = working_set.find_decl(b"let");
-        let mut_decl_id = working_set.find_decl(b"mut");
-
-        if pipeline_index == 0 {
-            for element in pipeline.elements.iter_mut() {
-                if let Expr::Call(call) = &element.expr.expr {
-                    if Some(call.decl_id) == let_decl_id || Some(call.decl_id) == mut_decl_id {
-                        // Do an expansion
-                        if let Some(Expression {
-                            expr: Expr::Block(block_id),
-                            ..
-                        }) = call.positional_iter().nth(1)
-                        {
-                            let block = working_set.get_block(*block_id);
-
-                            if let Some(element) = block
-                                .pipelines
-                                .first()
-                                .and_then(|p| p.elements.first())
-                                .cloned()
-                            {
-                                if element.has_in_variable(working_set) {
-                                    let element = wrap_element_with_collect(working_set, &element);
-                                    let block = working_set.get_block_mut(*block_id);
-                                    block.pipelines[0].elements[0] = element;
-                                }
-                            }
-                        }
-                        continue;
-                    } else if element.has_in_variable(working_set) && !is_subexpression {
-                        *element = wrap_element_with_collect(working_set, element);
                     }
-                } else if element.has_in_variable(working_set) && !is_subexpression {
+                })
-                    *element = wrap_element_with_collect(working_set, element);
+                .collect();
-                }
-            }
-        }
 
-        pipeline
+            Pipeline { elements }
+        }
+    } else {
+        // If there's only one command in the pipeline, this could be a builtin command
+        parse_builtin_commands(working_set, &pipeline.commands[0])
     }
 }
 
@@ -5840,18 +5755,45 @@ pub fn parse_block(
     }
 
     let mut block = Block::new_with_capacity(lite_block.block.len());
+    block.span = Some(span);
 
-    for (idx, lite_pipeline) in lite_block.block.iter().enumerate() {
+    for lite_pipeline in &lite_block.block {
-        let pipeline = parse_pipeline(working_set, lite_pipeline, is_subexpression, idx);
+        let pipeline = parse_pipeline(working_set, lite_pipeline);
         block.pipelines.push(pipeline);
     }
 
+    // If this is not a subexpression and there are any pipelines where the first element has $in,
+    // we can wrap the whole block in collect so that they all reference the same $in
+    if !is_subexpression
+        && block
+            .pipelines
+            .iter()
+            .flat_map(|pipeline| pipeline.elements.first())
+            .any(|element| element.has_in_variable(working_set))
+    {
+        // Move the block out to prepare it to become a subexpression
+        let inner_block = std::mem::take(&mut block);
+        block.span = inner_block.span;
+        let ty = inner_block.output_type();
+        let block_id = working_set.add_block(Arc::new(inner_block));
+
+        // Now wrap it in a Collect expression, and put it in the block as the only pipeline
+        let subexpression = Expression::new(working_set, Expr::Subexpression(block_id), span, ty);
+        let collect = wrap_expr_with_collect(working_set, subexpression);
+
+        block.pipelines.push(Pipeline {
+            elements: vec![PipelineElement {
+                pipe: None,
+                expr: collect,
+                redirection: None,
+            }],
+        });
+    }
+
     if scoped {
         working_set.exit_scope();
     }
 
-    block.span = Some(span);
-
     let errors = type_check::check_block_input_output(working_set, &block);
     if !errors.is_empty() {
         working_set.parse_errors.extend_from_slice(&errors);
@@ -6220,6 +6162,10 @@ pub fn discover_captures_in_expr(
                 discover_captures_in_expr(working_set, &match_.1, seen, seen_blocks, output)?;
             }
         }
+        Expr::Collect(var_id, expr) => {
+            seen.push(*var_id);
+            discover_captures_in_expr(working_set, expr, seen, seen_blocks, output)?
+        }
         Expr::RowCondition(block_id) | Expr::Subexpression(block_id) => {
             let block = working_set.get_block(*block_id);
 
@@ -6270,28 +6216,28 @@ pub fn discover_captures_in_expr(
 
 fn wrap_redirection_with_collect(
     working_set: &mut StateWorkingSet,
-    target: &RedirectionTarget,
+    target: RedirectionTarget,
 ) -> RedirectionTarget {
     match target {
         RedirectionTarget::File { expr, append, span } => RedirectionTarget::File {
             expr: wrap_expr_with_collect(working_set, expr),
-            span: *span,
+            span,
-            append: *append,
+            append,
         },
-        RedirectionTarget::Pipe { span } => RedirectionTarget::Pipe { span: *span },
+        RedirectionTarget::Pipe { span } => RedirectionTarget::Pipe { span },
     }
 }
 
 fn wrap_element_with_collect(
     working_set: &mut StateWorkingSet,
-    element: &PipelineElement,
+    element: PipelineElement,
 ) -> PipelineElement {
     PipelineElement {
         pipe: element.pipe,
-        expr: wrap_expr_with_collect(working_set, &element.expr),
+        expr: wrap_expr_with_collect(working_set, element.expr),
-        redirection: element.redirection.as_ref().map(|r| match r {
+        redirection: element.redirection.map(|r| match r {
             PipelineRedirection::Single { source, target } => PipelineRedirection::Single {
-                source: *source,
+                source,
                 target: wrap_redirection_with_collect(working_set, target),
             },
             PipelineRedirection::Separate { out, err } => PipelineRedirection::Separate {
@@ -6302,65 +6248,24 @@ fn wrap_element_with_collect(
     }
 }
 
-fn wrap_expr_with_collect(working_set: &mut StateWorkingSet, expr: &Expression) -> Expression {
+fn wrap_expr_with_collect(working_set: &mut StateWorkingSet, expr: Expression) -> Expression {
     let span = expr.span;
 
-    if let Some(decl_id) = working_set.find_decl(b"collect") {
+    // IN_VARIABLE_ID should get replaced with a unique variable, so that we don't have to
-        let mut output = vec![];
+    // execute as a closure
+    let var_id = working_set.add_variable(b"$in".into(), expr.span, Type::Any, false);
+    let mut expr = expr.clone();
+    expr.replace_in_variable(working_set, var_id);
 
-        let var_id = IN_VARIABLE_ID;
+    // Bind the custom `$in` variable for that particular expression
-        let mut signature = Signature::new("");
+    let ty = expr.ty.clone();
-        signature.required_positional.push(PositionalArg {
+    Expression::new(
-            var_id: Some(var_id),
+        working_set,
-            name: "$in".into(),
+        Expr::Collect(var_id, Box::new(expr)),
-            desc: String::new(),
+        span,
-            shape: SyntaxShape::Any,
+        // We can expect it to have the same result type
-            default_value: None,
+        ty,
-        });
+    )
-
-        let mut block = Block {
-            pipelines: vec![Pipeline::from_vec(vec![expr.clone()])],
-            signature: Box::new(signature),
-            ..Default::default()
-        };
-
-        compile_block(working_set, &mut block);
-
-        let block_id = working_set.add_block(Arc::new(block));
-
-        output.push(Argument::Positional(Expression::new(
-            working_set,
-            Expr::Closure(block_id),
-            span,
-            Type::Any,
-        )));
-
-        output.push(Argument::Named((
-            Spanned {
-                item: "keep-env".to_string(),
-                span: Span::new(0, 0),
-            },
-            None,
-            None,
-        )));
-
-        // The containing, synthetic call to `collect`.
-        // We don't want to have a real span as it will confuse flattening
-        // The args are where we'll get the real info
-        Expression::new(
-            working_set,
-            Expr::Call(Box::new(Call {
-                head: Span::new(0, 0),
-                arguments: output,
-                decl_id,
-                parser_info: HashMap::new(),
-            })),
-            span,
-            Type::Any,
-        )
-    } else {
-        Expression::garbage(working_set, span)
-    }
 }
 
 // Parses a vector of u8 to create an AST Block. If a file name is given, then

crates/nu-parser/tests/test_parser.rs

@@ -41,6 +41,41 @@ impl Command for Let {
     }
 }
 
+#[cfg(test)]
+#[derive(Clone)]
+pub struct Mut;
+
+#[cfg(test)]
+impl Command for Mut {
+    fn name(&self) -> &str {
+        "mut"
+    }
+
+    fn usage(&self) -> &str {
+        "Mock mut command."
+    }
+
+    fn signature(&self) -> nu_protocol::Signature {
+        Signature::build("mut")
+            .required("var_name", SyntaxShape::VarWithOptType, "variable name")
+            .required(
+                "initial_value",
+                SyntaxShape::Keyword(b"=".to_vec(), Box::new(SyntaxShape::MathExpression)),
+                "equals sign followed by value",
+            )
+    }
+
+    fn run(
+        &self,
+        _engine_state: &EngineState,
+        _stack: &mut Stack,
+        _call: &Call,
+        _input: PipelineData,
+    ) -> Result<PipelineData, ShellError> {
+        todo!()
+    }
+}
+
 fn test_int(
     test_tag: &str,     // name of sub-test
     test: &[u8],        // input expression
@@ -1149,11 +1184,29 @@ fn test_nothing_comparison_eq() {
 fn test_redirection_with_letmut(#[case] phase: &[u8]) {
     let engine_state = EngineState::new();
     let mut working_set = StateWorkingSet::new(&engine_state);
-    let _block = parse(&mut working_set, None, phase, true);
+    working_set.add_decl(Box::new(Let));
-    assert!(matches!(
+    working_set.add_decl(Box::new(Mut));
-        working_set.parse_errors.first(),
+
-        Some(ParseError::RedirectingBuiltinCommand(_, _, _))
+    let block = parse(&mut working_set, None, phase, true);
-    ));
+    assert!(
+        working_set.parse_errors.is_empty(),
+        "parse errors: {:?}",
+        working_set.parse_errors
+    );
+    assert_eq!(1, block.pipelines[0].elements.len());
+
+    let element = &block.pipelines[0].elements[0];
+    assert!(element.redirection.is_none()); // it should be in the let block, not here
+
+    if let Expr::Call(call) = &element.expr.expr {
+        let arg = call.positional_nth(1).expect("no positional args");
+        let block_id = arg.as_block().expect("arg 1 is not a block");
+        let block = working_set.get_block(block_id);
+        let inner_element = &block.pipelines[0].elements[0];
+        assert!(inner_element.redirection.is_some());
+    } else {
+        panic!("expected Call: {:?}", block.pipelines[0].elements[0])
+    }
 }
 
 #[rstest]

crates/nu-protocol/src/ast/block.rs

@@ -78,6 +78,19 @@ impl Block {
             Type::Nothing
         }
     }
+
+    /// Replace any `$in` variables in the initial element of pipelines within the block
+    pub fn replace_in_variable(
+        &mut self,
+        working_set: &mut StateWorkingSet<'_>,
+        new_var_id: VarId,
+    ) {
+        for pipeline in self.pipelines.iter_mut() {
+            if let Some(element) = pipeline.elements.first_mut() {
+                element.replace_in_variable(working_set, new_var_id);
+            }
+        }
+    }
 }
 
 impl<T> From<T> for Block

crates/nu-protocol/src/ast/expr.rs

@@ -25,6 +25,7 @@ pub enum Expr {
     RowCondition(BlockId),
     UnaryNot(Box<Expression>),
     BinaryOp(Box<Expression>, Box<Expression>, Box<Expression>), //lhs, op, rhs
+    Collect(VarId, Box<Expression>),
     Subexpression(BlockId),
     Block(BlockId),
     Closure(BlockId),
@@ -65,11 +66,13 @@ impl Expr {
         &self,
         working_set: &StateWorkingSet,
     ) -> (Option<OutDest>, Option<OutDest>) {
-        // Usages of `$in` will be wrapped by a `collect` call by the parser,
-        // so we do not have to worry about that when considering
-        // which of the expressions below may consume pipeline output.
         match self {
             Expr::Call(call) => working_set.get_decl(call.decl_id).pipe_redirection(),
+            Expr::Collect(_, _) => {
+                // A collect expression always has default redirection, it's just going to collect
+                // stdout unless another type of redirection is specified
+                (None, None)
+            },
             Expr::Subexpression(block_id) | Expr::Block(block_id) => working_set
                 .get_block(*block_id)
                 .pipe_redirection(working_set),

crates/nu-protocol/src/ast/expression.rs

@@ -6,6 +6,8 @@ use crate::{
 use serde::{Deserialize, Serialize};
 use std::sync::Arc;
 
+use super::ListItem;
+
 /// Wrapper around [`Expr`]
 #[derive(Debug, Clone, PartialEq, Serialize, Deserialize)]
 pub struct Expression {
@@ -106,37 +108,14 @@ impl Expression {
                 left.has_in_variable(working_set) || right.has_in_variable(working_set)
             }
             Expr::UnaryNot(expr) => expr.has_in_variable(working_set),
-            Expr::Block(block_id) => {
+            Expr::Block(block_id) | Expr::Closure(block_id) => {
                 let block = working_set.get_block(*block_id);
-
+                block.captures.contains(&IN_VARIABLE_ID)
-                if block.captures.contains(&IN_VARIABLE_ID) {
+                    || block
-                    return true;
+                        .pipelines
-                }
+                        .iter()
-
+                        .flat_map(|pipeline| pipeline.elements.first())
-                if let Some(pipeline) = block.pipelines.first() {
+                        .any(|element| element.has_in_variable(working_set))
-                    match pipeline.elements.first() {
-                        Some(element) => element.has_in_variable(working_set),
-                        None => false,
-                    }
-                } else {
-                    false
-                }
-            }
-            Expr::Closure(block_id) => {
-                let block = working_set.get_block(*block_id);
-
-                if block.captures.contains(&IN_VARIABLE_ID) {
-                    return true;
-                }
-
-                if let Some(pipeline) = block.pipelines.first() {
-                    match pipeline.elements.first() {
-                        Some(element) => element.has_in_variable(working_set),
-                        None => false,
-                    }
-                } else {
-                    false
-                }
             }
             Expr::Binary(_) => false,
             Expr::Bool(_) => false,
@@ -251,6 +230,9 @@ impl Expression {
             Expr::Signature(_) => false,
             Expr::String(_) => false,
             Expr::RawString(_) => false,
+            // A `$in` variable found within a `Collect` is local, as it's already been wrapped
+            // This is probably unlikely to happen anyway - the expressions are wrapped depth-first
+            Expr::Collect(_, _) => false,
             Expr::RowCondition(block_id) | Expr::Subexpression(block_id) => {
                 let block = working_set.get_block(*block_id);
 
@@ -414,6 +396,7 @@ impl Expression {
                     i.replace_span(working_set, replaced, new_span)
                 }
             }
+            Expr::Collect(_, expr) => expr.replace_span(working_set, replaced, new_span),
             Expr::RowCondition(block_id) | Expr::Subexpression(block_id) => {
                 let mut block = (**working_set.get_block(*block_id)).clone();
 
@@ -443,6 +426,129 @@ impl Expression {
         }
     }
 
+    pub fn replace_in_variable(&mut self, working_set: &mut StateWorkingSet, new_var_id: VarId) {
+        match &mut self.expr {
+            Expr::Bool(_) => {}
+            Expr::Int(_) => {}
+            Expr::Float(_) => {}
+            Expr::Binary(_) => {}
+            Expr::Range(_) => {}
+            Expr::Var(var_id) | Expr::VarDecl(var_id) => {
+                if *var_id == IN_VARIABLE_ID {
+                    *var_id = new_var_id;
+                }
+            }
+            Expr::Call(call) => {
+                for arg in call.arguments.iter_mut() {
+                    match arg {
+                        Argument::Positional(expr)
+                        | Argument::Unknown(expr)
+                        | Argument::Named((_, _, Some(expr)))
+                        | Argument::Spread(expr) => {
+                            expr.replace_in_variable(working_set, new_var_id)
+                        }
+                        Argument::Named((_, _, None)) => {}
+                    }
+                }
+                for expr in call.parser_info.values_mut() {
+                    expr.replace_in_variable(working_set, new_var_id)
+                }
+            }
+            Expr::ExternalCall(head, args) => {
+                head.replace_in_variable(working_set, new_var_id);
+                for arg in args.iter_mut() {
+                    match arg {
+                        ExternalArgument::Regular(expr) | ExternalArgument::Spread(expr) => {
+                            expr.replace_in_variable(working_set, new_var_id)
+                        }
+                    }
+                }
+            }
+            Expr::Operator(_) => {}
+            // `$in` in `Collect` has already been handled, so we don't need to check further
+            Expr::Collect(_, _) => {}
+            Expr::Block(block_id)
+            | Expr::Closure(block_id)
+            | Expr::RowCondition(block_id)
+            | Expr::Subexpression(block_id) => {
+                let mut block = Block::clone(working_set.get_block(*block_id));
+                block.replace_in_variable(working_set, new_var_id);
+                *working_set.get_block_mut(*block_id) = block;
+            }
+            Expr::UnaryNot(expr) => {
+                expr.replace_in_variable(working_set, new_var_id);
+            }
+            Expr::BinaryOp(lhs, op, rhs) => {
+                for expr in [lhs, op, rhs] {
+                    expr.replace_in_variable(working_set, new_var_id);
+                }
+            }
+            Expr::MatchBlock(match_patterns) => {
+                for (_, expr) in match_patterns.iter_mut() {
+                    expr.replace_in_variable(working_set, new_var_id);
+                }
+            }
+            Expr::List(items) => {
+                for item in items.iter_mut() {
+                    match item {
+                        ListItem::Item(expr) | ListItem::Spread(_, expr) => {
+                            expr.replace_in_variable(working_set, new_var_id)
+                        }
+                    }
+                }
+            }
+            Expr::Table(table) => {
+                for col_expr in table.columns.iter_mut() {
+                    col_expr.replace_in_variable(working_set, new_var_id);
+                }
+                for row in table.rows.iter_mut() {
+                    for row_expr in row.iter_mut() {
+                        row_expr.replace_in_variable(working_set, new_var_id);
+                    }
+                }
+            }
+            Expr::Record(items) => {
+                for item in items.iter_mut() {
+                    match item {
+                        RecordItem::Pair(key, val) => {
+                            key.replace_in_variable(working_set, new_var_id);
+                            val.replace_in_variable(working_set, new_var_id);
+                        }
+                        RecordItem::Spread(_, expr) => {
+                            expr.replace_in_variable(working_set, new_var_id)
+                        }
+                    }
+                }
+            }
+            Expr::Keyword(kw) => kw.expr.replace_in_variable(working_set, new_var_id),
+            Expr::ValueWithUnit(value_with_unit) => value_with_unit
+                .expr
+                .replace_in_variable(working_set, new_var_id),
+            Expr::DateTime(_) => {}
+            Expr::Filepath(_, _) => {}
+            Expr::Directory(_, _) => {}
+            Expr::GlobPattern(_, _) => {}
+            Expr::String(_) => {}
+            Expr::RawString(_) => {}
+            Expr::CellPath(_) => {}
+            Expr::FullCellPath(full_cell_path) => {
+                full_cell_path
+                    .head
+                    .replace_in_variable(working_set, new_var_id);
+            }
+            Expr::ImportPattern(_) => {}
+            Expr::Overlay(_) => {}
+            Expr::Signature(_) => {}
+            Expr::StringInterpolation(exprs) | Expr::GlobInterpolation(exprs, _) => {
+                for expr in exprs.iter_mut() {
+                    expr.replace_in_variable(working_set, new_var_id);
+                }
+            }
+            Expr::Nothing => {}
+            Expr::Garbage => {}
+        }
+    }
+
     pub fn new(working_set: &mut StateWorkingSet, expr: Expr, span: Span, ty: Type) -> Expression {
         let span_id = working_set.add_span(span);
         Expression {

crates/nu-protocol/src/ast/pipeline.rs

@@ -1,4 +1,4 @@
-use crate::{ast::Expression, engine::StateWorkingSet, OutDest, Span};
+use crate::{ast::Expression, engine::StateWorkingSet, OutDest, Span, VarId};
 use serde::{Deserialize, Serialize};
 use std::fmt::Display;
 
@@ -62,6 +62,19 @@ impl RedirectionTarget {
             RedirectionTarget::Pipe { .. } => {}
         }
     }
+
+    pub fn replace_in_variable(
+        &mut self,
+        working_set: &mut StateWorkingSet<'_>,
+        new_var_id: VarId,
+    ) {
+        match self {
+            RedirectionTarget::File { expr, .. } => {
+                expr.replace_in_variable(working_set, new_var_id)
+            }
+            RedirectionTarget::Pipe { .. } => {}
+        }
+    }
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize, PartialEq)]
@@ -75,6 +88,23 @@ pub enum PipelineRedirection {
         err: RedirectionTarget,
     },
 }
+impl PipelineRedirection {
+    pub fn replace_in_variable(
+        &mut self,
+        working_set: &mut StateWorkingSet<'_>,
+        new_var_id: VarId,
+    ) {
+        match self {
+            PipelineRedirection::Single { source: _, target } => {
+                target.replace_in_variable(working_set, new_var_id)
+            }
+            PipelineRedirection::Separate { out, err } => {
+                out.replace_in_variable(working_set, new_var_id);
+                err.replace_in_variable(working_set, new_var_id);
+            }
+        }
+    }
+}
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct PipelineElement {
@@ -120,6 +150,17 @@ impl PipelineElement {
     ) -> (Option<OutDest>, Option<OutDest>) {
         self.expr.expr.pipe_redirection(working_set)
     }
+
+    pub fn replace_in_variable(
+        &mut self,
+        working_set: &mut StateWorkingSet<'_>,
+        new_var_id: VarId,
+    ) {
+        self.expr.replace_in_variable(working_set, new_var_id);
+        if let Some(redirection) = &mut self.redirection {
+            redirection.replace_in_variable(working_set, new_var_id);
+        }
+    }
 }
 
 #[derive(Debug, Clone, Serialize, Deserialize)]

crates/nu-protocol/src/debugger/profiler.rs

@@ -343,6 +343,7 @@ fn expr_to_string(engine_state: &EngineState, expr: &Expr) -> String {
         Expr::String(_) | Expr::RawString(_) => "string".to_string(),
         Expr::StringInterpolation(_) => "string interpolation".to_string(),
         Expr::GlobInterpolation(_, _) => "glob interpolation".to_string(),
+        Expr::Collect(_, _) => "collect".to_string(),
         Expr::Subexpression(_) => "subexpression".to_string(),
         Expr::Table(_) => "table".to_string(),
         Expr::UnaryNot(_) => "unary not".to_string(),

crates/nu-protocol/src/eval_base.rs

@@ -159,6 +159,9 @@ pub trait Eval {
             Expr::ExternalCall(head, args) => {
                 Self::eval_external_call(state, mut_state, head, args, expr_span)
             }
+            Expr::Collect(var_id, expr) => {
+                Self::eval_collect::<D>(state, mut_state, *var_id, expr)
+            }
             Expr::Subexpression(block_id) => {
                 Self::eval_subexpression::<D>(state, mut_state, *block_id, expr_span)
             }
@@ -356,6 +359,13 @@ pub trait Eval {
         span: Span,
     ) -> Result<Value, ShellError>;
 
+    fn eval_collect<D: DebugContext>(
+        state: Self::State<'_>,
+        mut_state: &mut Self::MutState,
+        var_id: VarId,
+        expr: &Expression,
+    ) -> Result<Value, ShellError>;
+
     fn eval_subexpression<D: DebugContext>(
         state: Self::State<'_>,
         mut_state: &mut Self::MutState,

crates/nu-protocol/src/eval_const.rs

@@ -422,6 +422,15 @@ impl Eval for EvalConst {
         Err(ShellError::NotAConstant { span })
     }
 
+    fn eval_collect<D: DebugContext>(
+        _: &StateWorkingSet,
+        _: &mut (),
+        _var_id: VarId,
+        expr: &Expression,
+    ) -> Result<Value, ShellError> {
+        Err(ShellError::NotAConstant { span: expr.span })
+    }
+
     fn eval_subexpression<D: DebugContext>(
         working_set: &StateWorkingSet,
         _: &mut (),

crates/nu-protocol/src/ir/display.rs

@@ -102,6 +102,10 @@ impl<'a> fmt::Display for FmtInstruction<'a> {
                 let var = FmtVar::new(self.engine_state, *var_id);
                 write!(f, "{:WIDTH$} {var}, {src}", "store-variable")
             }
+            Instruction::DropVariable { var_id } => {
+                let var = FmtVar::new(self.engine_state, *var_id);
+                write!(f, "{:WIDTH$} {var}", "drop-variable")
+            }
             Instruction::LoadEnv { dst, key } => {
                 let key = FmtData(self.data, *key);
                 write!(f, "{:WIDTH$} {dst}, {key}", "load-env")

crates/nu-protocol/src/ir/mod.rs

@@ -127,6 +127,8 @@ pub enum Instruction {
     LoadVariable { dst: RegId, var_id: VarId },
     /// Store the value of a variable from the `src` register
     StoreVariable { var_id: VarId, src: RegId },
+    /// Remove a variable from the stack, freeing up whatever resources were associated with it
+    DropVariable { var_id: VarId },
     /// Load the value of an environment variable into the `dst` register
     LoadEnv { dst: RegId, key: DataSlice },
     /// Load the value of an environment variable into the `dst` register, or `Nothing` if it
@@ -290,6 +292,7 @@ impl Instruction {
             Instruction::Drain { .. } => None,
             Instruction::LoadVariable { dst, .. } => Some(dst),
             Instruction::StoreVariable { .. } => None,
+            Instruction::DropVariable { .. } => None,
             Instruction::LoadEnv { dst, .. } => Some(dst),
             Instruction::LoadEnvOpt { dst, .. } => Some(dst),
             Instruction::StoreEnv { .. } => None,

crates/nuon/src/from.rs

@@ -329,6 +329,12 @@ fn convert_to_value(
             msg: "glob interpolation not supported in nuon".into(),
             span: expr.span,
         }),
+        Expr::Collect(..) => Err(ShellError::OutsideSpannedLabeledError {
+            src: original_text.to_string(),
+            error: "Error when loading".into(),
+            msg: "`$in` not supported in nuon".into(),
+            span: expr.span,
+        }),
         Expr::Subexpression(..) => Err(ShellError::OutsideSpannedLabeledError {
             src: original_text.to_string(),
             error: "Error when loading".into(),

tests/repl/test_engine.rs

@@ -52,6 +52,29 @@ fn in_and_if_else() -> TestResult {
     )
 }
 
+#[test]
+fn in_with_closure() -> TestResult {
+    // Can use $in twice
+    run_test(r#"3 | do { let x = $in; let y = $in; $x + $y }"#, "6")
+}
+
+#[test]
+fn in_with_custom_command() -> TestResult {
+    // Can use $in twice
+    run_test(
+        r#"def foo [] { let x = $in; let y = $in; $x + $y }; 3 | foo"#,
+        "6",
+    )
+}
+
+#[test]
+fn in_used_twice_and_also_in_pipeline() -> TestResult {
+    run_test(
+        r#"3 | do { let x = $in; let y = $in; $x + $y | $in * 4 }"#,
+        "24",
+    )
+}
+
 #[test]
 fn help_works_with_missing_requirements() -> TestResult {
     run_test(r#"each --help | lines | length"#, "72")

tests/repl/test_type_check.rs

@@ -129,3 +129,16 @@ fn transpose_into_load_env() -> TestResult {
         "10",
     )
 }
+
+#[test]
+fn in_variable_expression_correct_output_type() -> TestResult {
+    run_test(r#"def foo []: nothing -> string { 'foo' | $"($in)" }"#, "")
+}
+
+#[test]
+fn in_variable_expression_wrong_output_type() -> TestResult {
+    fail_test(
+        r#"def foo []: nothing -> int { 'foo' | $"($in)" }"#,
+        "expected int",
+    )
+}