019eff8ea1
5 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
Devyn Cairns
|
aa7d7d0cc3
|
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)
|
||
|
Devyn Cairns
|
a2758e6c40
|
Add IR support to the debugger (#13345)
# Description This adds tracing for each individual instruction to the `Debugger` trait. Register contents can be inspected both when entering and leaving an instruction, and if an instruction produced an error, a reference to the error is also available. It's not the full `EvalContext` but it's most of the important parts for getting an idea of what's going on. Added support for all of this to the `Profiler` / `debug profile` as well, and the output is quite incredible - super verbose, but you can see every instruction that's executed and also what the result was if it's an instruction that has a clearly defined output (many do). # User-Facing Changes - Added `--instructions` to `debug profile`, which adds the `pc` and `instruction` columns to the output. - `--expr` only works in AST mode, and `--instructions` only works in IR mode. In the wrong mode, the output for those columns is just blank. # Tests + Formatting All passing. # After Submitting - [ ] release notes |
||
|
Devyn Cairns
|
ccd0160c32
|
Make the store-env IR instruction also update config (#13351)
# Description Follow up fix to #13332, so that changes to config when running under IR actually happen as well. Since I merged them around the same time, I forgot about this. |
||
Darren Schroeder
|
ac561b1b0e
|
quick fix up for ir pr as_refs (#13340)
# Description Was having an issue compiling main after the IR pr. Talked to devyn and he led me to change a couple things real quick and we're compiling once again. |
||
|
Devyn Cairns
|
d7392f1f3b
|
Internal representation (IR) compiler and evaluator (#13330)
# Description This PR adds an internal representation language to Nushell, offering an alternative evaluator based on simple instructions, stream-containing registers, and indexed control flow. The number of registers required is determined statically at compile-time, and the fixed size required is allocated upon entering the block. Each instruction is associated with a span, which makes going backwards from IR instructions to source code very easy. Motivations for IR: 1. **Performance.** By simplifying the evaluation path and making it more cache-friendly and branch predictor-friendly, code that does a lot of computation in Nushell itself can be sped up a decent bit. Because the IR is fairly easy to reason about, we can also implement optimization passes in the future to eliminate and simplify code. 2. **Correctness.** The instructions mostly have very simple and easily-specified behavior, so hopefully engine changes are a little bit easier to reason about, and they can be specified in a more formal way at some point. I have made an effort to document each of the instructions in the docs for the enum itself in a reasonably specific way. Some of the errors that would have happened during evaluation before are now moved to the compilation step instead, because they don't make sense to check during evaluation. 3. **As an intermediate target.** This is a good step for us to bring the [`new-nu-parser`](https://github.com/nushell/new-nu-parser) in at some point, as code generated from new AST can be directly compared to code generated from old AST. If the IR code is functionally equivalent, it will behave the exact same way. 4. **Debugging.** With a little bit more work, we can probably give control over advancing the virtual machine that `IrBlock`s run on to some sort of external driver, making things like breakpoints and single stepping possible. Tools like `view ir` and [`explore ir`](https://github.com/devyn/nu_plugin_explore_ir) make it easier than before to see what exactly is going on with your Nushell code. The goal is to eventually replace the AST evaluator entirely, once we're sure it's working just as well. You can help dogfood this by running Nushell with `$env.NU_USE_IR` set to some value. The environment variable is checked when Nushell starts, so config runs with IR, or it can also be set on a line at the REPL to change it dynamically. It is also checked when running `do` in case within a script you want to just run a specific piece of code with or without IR. # Example ```nushell view ir { |data| mut sum = 0 for n in $data { $sum += $n } $sum } ``` ```gas # 3 registers, 19 instructions, 0 bytes of data 0: load-literal %0, int(0) 1: store-variable var 904, %0 # let 2: drain %0 3: drop %0 4: load-variable %1, var 903 5: iterate %0, %1, end 15 # for, label(1), from(14:) 6: store-variable var 905, %0 7: load-variable %0, var 904 8: load-variable %2, var 905 9: binary-op %0, Math(Plus), %2 10: span %0 11: store-variable var 904, %0 12: load-literal %0, nothing 13: drain %0 14: jump 5 15: drop %0 # label(0), from(5:) 16: drain %0 17: load-variable %0, var 904 18: return %0 ``` # Benchmarks All benchmarks run on a base model Mac Mini M1. ## Iterative Fibonacci sequence This is about as best case as possible, making use of the much faster control flow. Most code will not experience a speed improvement nearly this large. ```nushell def fib [n: int] { mut a = 0 mut b = 1 for _ in 2..=$n { let c = $a + $b $a = $b $b = $c } $b } use std bench bench { 0..50 | each { |n| fib $n } } ``` IR disabled: ``` ╭───────┬─────────────────╮ │ mean │ 1ms 924µs 665ns │ │ min │ 1ms 700µs 83ns │ │ max │ 3ms 450µs 125ns │ │ std │ 395µs 759ns │ │ times │ [list 50 items] │ ╰───────┴─────────────────╯ ``` IR enabled: ``` ╭───────┬─────────────────╮ │ mean │ 452µs 820ns │ │ min │ 427µs 417ns │ │ max │ 540µs 167ns │ │ std │ 17µs 158ns │ │ times │ [list 50 items] │ ╰───────┴─────────────────╯ ```  ## [gradient_benchmark_no_check.nu](https://github.com/nushell/nu_scripts/blob/main/benchmarks/gradient_benchmark_no_check.nu) IR disabled: ``` ╭───┬──────────────────╮ │ 0 │ 27ms 929µs 958ns │ │ 1 │ 21ms 153µs 459ns │ │ 2 │ 18ms 639µs 666ns │ │ 3 │ 19ms 554µs 583ns │ │ 4 │ 13ms 383µs 375ns │ │ 5 │ 11ms 328µs 208ns │ │ 6 │ 5ms 659µs 542ns │ ╰───┴──────────────────╯ ``` IR enabled: ``` ╭───┬──────────────────╮ │ 0 │ 22ms 662µs │ │ 1 │ 17ms 221µs 792ns │ │ 2 │ 14ms 786µs 708ns │ │ 3 │ 13ms 876µs 834ns │ │ 4 │ 13ms 52µs 875ns │ │ 5 │ 11ms 269µs 666ns │ │ 6 │ 6ms 942µs 500ns │ ╰───┴──────────────────╯ ``` ## [random-bytes.nu](https://github.com/nushell/nu_scripts/blob/main/benchmarks/random-bytes.nu) I got pretty random results out of this benchmark so I decided not to include it. Not clear why. # User-Facing Changes - IR compilation errors may appear even if the user isn't evaluating with IR. - IR evaluation can be enabled by setting the `NU_USE_IR` environment variable to any value. - New command `view ir` pretty-prints the IR for a block, and `view ir --json` can be piped into an external tool like [`explore ir`](https://github.com/devyn/nu_plugin_explore_ir). # Tests + Formatting All tests are passing with `NU_USE_IR=1`, and I've added some more eval tests to compare the results for some very core operations. I will probably want to add some more so we don't have to always check `NU_USE_IR=1 toolkit test --workspace` on a regular basis. # After Submitting - [ ] release notes - [ ] further documentation of instructions? - [ ] post-release: publish `nu_plugin_explore_ir` |