Better error array numerics

[xunilrj]

Description

This PR fixes: #6376

The issue is that we use the first element of the array as its element type. This can be problematic as the first element is not "concrete". For example, when it is Numeric.

Now we iterate the array elements, search for the first "good" element, and use its type as the array type. This means:
1 - if the type is concrete, we use it;
2 - if the type is Numeric, we keep searching for something better;

This allows consistent arrays, but where the concrete types are at the end.

let a = [1, 2u8]; <- we will choose u8
let a = [None, Some(1), Some(2u8)]; <- we will choose Option<u8>, instead of Option<T>

snapshot tests

This PR also brings an experiment. Now when an "e2e" test has a file snapshot.toml it will run as cargo insta snapshot tests. These tests can be run as normal: cargo r -p test, and in two new ways:

> cargo t -p test
> cargo insta test

Snapshots can be reviewed using normal "cargo insta" workflow.

The secret lies in https://github.com/LukasKalbertodt/libtest-mimic. I also want to suggest that all our tests migrate to this model, where we could just use "cargo test" normally. Apart from that, I opted to use the forc executable, which means tests will run in parallel. In the future, if we want to poke compiler internals, we can expose them using flags, or create a special binary to fit our needs.

snapshot.toml is still empty because I have no idea to what put there at the moment. My initial idea was to configure what ends up into snapshots (multiple runs, compiler flags, script/contract calls) etc... But for the moment, empty is enough.

Checklist

• I have linked to any relevant issues.
• I have commented my code, particularly in hard-to-understand areas.
• I have updated the documentation where relevant (API docs, the reference, and the Sway book).
  • If my change requires substantial documentation changes, I have requested support from the DevRel team
• I have added tests that prove my fix is effective or that my feature works.
• I have added (or requested a maintainer to add) the necessary Breaking* or New Feature labels where relevant.
• I have done my best to ensure that my PR adheres to the Fuel Labs Code Review Standards.
• I have requested a review from the relevant team or maintainers.

[github-actions]

Benchmark for a0deb0b

Click to view benchmark

Test	Base	PR	%
code_action	5.2±0.08ms	5.2±0.07ms	0.00%
code_lens	292.0±7.96ns	296.8±11.03ns	+1.64%
compile	2.7±0.05s	2.6±0.04s	-3.70%
completion	4.7±0.04ms	4.7±0.05ms	0.00%
did_change_with_caching	2.6±0.03s	2.6±0.02s	0.00%
document_symbol	905.6±28.75µs	889.8±52.15µs	-1.74%
format	83.2±0.98ms	82.5±1.62ms	-0.84%
goto_definition	336.5±6.81µs	338.0±5.37µs	+0.45%
highlight	9.0±0.08ms	9.1±0.15ms	+1.11%
hover	489.1±6.03µs	490.7±8.55µs	+0.33%
idents_at_position	117.5±0.48µs	121.2±1.60µs	+3.15%
inlay_hints	641.3±19.23µs	648.3±19.75µs	+1.09%
on_enter	2.3±0.03µs	2.0±0.06µs	-13.04%
parent_decl_at_position	3.7±0.03ms	3.7±0.03ms	0.00%
prepare_rename	333.9±3.06µs	338.9±7.50µs	+1.50%
rename	9.4±0.10ms	9.3±0.12ms	-1.06%
semantic_tokens	1141.2±37.23µs	1228.2±12.73µs	+7.62%
token_at_position	338.7±2.35µs	337.9±2.68µs	-0.24%
tokens_at_position	3.7±0.06ms	3.7±0.05ms	0.00%
tokens_for_file	463.4±1.99µs	394.8±4.22µs	-14.80%
traverse	37.6±1.24ms	36.6±0.72ms	-2.66%

[github-actions]

Benchmark for 4c3ff5d

Click to view benchmark

Test	Base	PR	%
code_action	5.2±0.01ms	5.2±0.08ms	0.00%
code_lens	286.7±7.24ns	286.6±7.52ns	-0.03%
compile	2.7±0.04s	2.7±0.04s	0.00%
completion	4.7±0.03ms	4.7±0.06ms	0.00%
did_change_with_caching	2.6±0.04s	2.6±0.04s	0.00%
document_symbol	886.1±39.17µs	888.6±32.54µs	+0.28%
format	72.0±0.99ms	72.1±0.59ms	+0.14%
goto_definition	333.9±7.31µs	333.5±7.01µs	-0.12%
highlight	9.1±0.14ms	9.1±0.04ms	0.00%
hover	485.9±5.54µs	489.4±7.61µs	+0.72%
idents_at_position	118.9±0.38µs	118.5±0.71µs	-0.34%
inlay_hints	636.0±20.85µs	648.2±25.41µs	+1.92%
on_enter	2.1±0.04µs	2.1±0.06µs	0.00%
parent_decl_at_position	3.7±0.03ms	3.7±0.02ms	0.00%
prepare_rename	334.5±4.74µs	340.3±23.05µs	+1.73%
rename	9.4±0.07ms	9.4±0.08ms	0.00%
semantic_tokens	1162.0±18.58µs	1187.3±17.40µs	+2.18%
token_at_position	328.4±4.86µs	335.2±1.56µs	+2.07%
tokens_at_position	3.7±0.03ms	3.7±0.02ms	0.00%
tokens_for_file	399.3±3.25µs	403.8±6.99µs	+1.13%
traverse	34.8±0.81ms	34.5±0.72ms	-0.86%

[github-actions]

Benchmark for 8ee81cc

Click to view benchmark

Test	Base	PR	%
code_action	5.3±0.42ms	5.1±0.11ms	-3.77%
code_lens	286.6±7.88ns	300.2±8.62ns	+4.75%
compile	2.7±0.06s	2.7±0.07s	0.00%
completion	4.7±0.07ms	4.6±0.07ms	-2.13%
did_change_with_caching	2.6±0.04s	2.6±0.04s	0.00%
document_symbol	874.0±29.55µs	938.2±33.53µs	+7.35%
format	74.5±0.95ms	71.8±1.30ms	-3.62%
goto_definition	336.9±10.15µs	346.5±7.58µs	+2.85%
highlight	9.1±0.18ms	8.7±0.09ms	-4.40%
hover	489.8±6.82µs	503.0±6.82µs	+2.69%
idents_at_position	118.7±1.45µs	119.2±0.51µs	+0.42%
inlay_hints	637.7±29.70µs	624.4±31.16µs	-2.09%
on_enter	1960.7±116.19ns	2.1±0.06µs	+7.10%
parent_decl_at_position	3.7±0.04ms	3.6±0.06ms	-2.70%
prepare_rename	346.4±6.37µs	341.3±6.50µs	-1.47%
rename	9.4±0.12ms	9.3±0.24ms	-1.06%
semantic_tokens	1153.7±20.46µs	1246.9±50.43µs	+8.08%
token_at_position	328.9±3.21µs	335.2±4.93µs	+1.92%
tokens_at_position	3.7±0.01ms	3.6±0.05ms	-2.70%
tokens_for_file	398.9±2.70µs	404.5±4.88µs	+1.40%
traverse	34.5±0.65ms	35.6±0.91ms	+3.19%

[Voxelot]

+1 for libtest mimic, we've been using it over in the fuel-core repo already as well.

[ironcev]

Great that we are starting with snapshot tests!! 💯 🥇 Can we just add to the README.md of the e2e_vm_tests short instructions how to use them? Basically this what is written in the PR description.

[xunilrj]

Great that we are starting with snapshot tests!! 💯 🥇 Can we just add to the README.md of the e2e_vm_tests short instructions how to use them? Basically this what is written in the PR description.

Done.

[github-actions]

Benchmark for 30f23e8

Click to view benchmark

Test	Base	PR	%
code_action	5.3±0.05ms	5.1±0.03ms	-3.77%
code_lens	299.6±8.12ns	307.7±13.43ns	+2.70%
compile	2.7±0.06s	2.7±0.04s	0.00%
completion	4.7±0.05ms	4.6±0.08ms	-2.13%
did_change_with_caching	2.6±0.05s	2.7±0.03s	+3.85%
document_symbol	850.7±14.92µs	857.9±13.65µs	+0.85%
format	71.2±1.67ms	72.2±1.04ms	+1.40%
goto_definition	338.2±6.62µs	350.4±11.15µs	+3.61%
highlight	9.1±0.02ms	8.7±0.12ms	-4.40%
hover	493.2±4.39µs	497.5±5.23µs	+0.87%
idents_at_position	121.3±0.26µs	119.6±4.83µs	-1.40%
inlay_hints	632.8±6.43µs	632.7±22.34µs	-0.02%
on_enter	2.1±0.06µs	1886.7±25.11ns	-10.16%
parent_decl_at_position	3.7±0.07ms	3.5±0.04ms	-5.41%
prepare_rename	333.4±4.36µs	343.5±5.92µs	+3.03%
rename	9.4±0.04ms	9.0±0.46ms	-4.26%
semantic_tokens	1246.7±12.76µs	1265.4±12.28µs	+1.50%
token_at_position	337.1±4.26µs	330.9±2.52µs	-1.84%
tokens_at_position	3.7±0.03ms	3.6±0.10ms	-2.70%
tokens_for_file	398.9±12.45µs	401.3±3.16µs	+0.60%
traverse	34.1±0.88ms	35.5±0.97ms	+4.11%

[github-actions]

Benchmark for 80f8904

Click to view benchmark

Test	Base	PR	%
code_action	5.3±0.09ms	5.1±0.10ms	-3.77%
code_lens	289.5±10.78ns	303.1±6.37ns	+4.70%
compile	2.7±0.06s	2.7±0.04s	0.00%
completion	4.7±0.03ms	4.5±0.06ms	-4.26%
did_change_with_caching	2.6±0.05s	2.6±0.03s	0.00%
document_symbol	853.0±24.11µs	934.7±25.83µs	+9.58%
format	72.2±1.40ms	70.8±1.08ms	-1.94%
goto_definition	339.7±7.47µs	337.7±6.59µs	-0.59%
highlight	9.1±0.09ms	8.7±0.10ms	-4.40%
hover	490.4±3.65µs	492.3±7.29µs	+0.39%
idents_at_position	118.8±0.53µs	118.5±0.28µs	-0.25%
inlay_hints	647.5±27.06µs	631.3±24.68µs	-2.50%
on_enter	2.1±0.03µs	2.0±0.08µs	-4.76%
parent_decl_at_position	3.7±0.01ms	3.6±0.04ms	-2.70%
prepare_rename	335.0±5.45µs	336.2±5.87µs	+0.36%
rename	9.4±0.13ms	9.0±0.15ms	-4.26%
semantic_tokens	1205.7±14.91µs	1222.7±11.21µs	+1.41%
token_at_position	330.9±1.67µs	332.2±1.68µs	+0.39%
tokens_at_position	3.7±0.03ms	3.5±0.03ms	-5.41%
tokens_for_file	398.8±2.52µs	404.5±2.52µs	+1.43%
traverse	33.9±0.74ms	34.4±0.80ms	+1.47%

[tritao]

About switching to libtest mimic, what is your idea on how it should be done?

I assume the idea is we would get rid of our test running code test, which would now be rewritten (or replaced by another target) on top of libtest mimic.

Upon an initial look at it, from its docs, looks like output capture and --nocapture are not supported. It redirects for LukasKalbertodt/libtest-mimic#9 for more information, and looks like its not going to be supported.

This means we lose the way to see individual test outputs when running the test suite (LukasKalbertodt/libtest-mimic#9).

Another potential issue is the filtering support we have built in our test runner over time, which we would need to reimplement on top of the new system.

It certainly seems appealing to offload some of this test running code, but my concern is we might not be easily able to extend in the future if we needed.

What are we gaining by introducing libtest mimic? Is it just to make the output as similiar to cargo test? Or is there anything else?

[xunilrj]

This means we lose the way to see individual test outputs when running the test suite (LukasKalbertodt/libtest-mimic#9).

We can still use what we have: gag would still work, but ideally, we would move to a better model, allowing us to run more tests in parallel.

I think that for compilation, we can use the forc executable like I am doing for snapshots. cargo insta would be much better for warnings/errors for both should_pass and should_fail tests.

For tests that need the VM to run, we can keep what we have because we do not capture output, right?

For "IR generation tests", we could change them to start using forc and use cargo insta with --ir initial. But they would work as is now.

"sdk harness" can also benefit from this workflow, but we would need to build the sway code using build.rs, because of all the macros that need access to the ABI jsons. So I would probably keep them in a different binary and only compile them when needed.

Another potential issue is the filtering support we have built in our test runner over time, which we would need to reimplement on top of the new system.

I think we have two options. One is to keep the current filter and only return the filtered tests to libmimic_test. The only difference I would propose is that when no filter is specified, we would expand all tests and all options. For example, no filter runs both debug AND release etc...

A second option would be to always expand every single option, and let cargo t filter do its thing. So for example, instead of

> cargo r -p test -- --abi-only

It could be something like

> cargo r -p test -- "(abi)"

This would work because our "test finder" can append "(abi)" when a test, tests the abi. I think we can think of something similar to other cases.

What are we gaining by introducing libtest mimic? Is it just to make the output as similiar to cargo test? Or is there anything else?

I think the biggest gain would be that our tests would be more "standard rust tests", which means they would work with the rest of the ecosystem. cargo insta, cargo nextest are two examples.

Maybe even things like: https://github.com/swellaby/vscode-rust-test-adapter

But, for me, honestly, would be that cargo t actually tests everything.

[github-actions]

Benchmark for 496f79c

Click to view benchmark

Test	Base	PR	%
code_action	5.0±0.07ms	5.3±0.19ms	+6.00%
code_lens	301.6±11.70ns	302.2±15.23ns	+0.20%
compile	2.7±0.05s	2.8±0.04s	+3.70%
completion	4.5±0.07ms	4.7±0.14ms	+4.44%
did_change_with_caching	2.6±0.06s	2.7±0.04s	+3.85%
document_symbol	927.1±40.68µs	862.5±17.53µs	-6.97%
format	68.1±0.68ms	68.2±0.92ms	+0.15%
goto_definition	341.2±6.84µs	360.9±14.34µs	+5.77%
highlight	8.9±0.32ms	8.9±0.27ms	0.00%
hover	487.6±5.04µs	581.8±7.79µs	+19.32%
idents_at_position	117.9±0.66µs	117.6±0.56µs	-0.25%
inlay_hints	632.4±22.03µs	644.7±31.76µs	+1.94%
on_enter	2.1±0.05µs	2.0±0.09µs	-4.76%
parent_decl_at_position	3.6±0.04ms	3.6±0.04ms	0.00%
prepare_rename	336.8±4.74µs	395.4±8.59µs	+17.40%
rename	9.0±0.14ms	9.2±0.16ms	+2.22%
semantic_tokens	1170.4±11.07µs	1265.0±11.60µs	+8.08%
token_at_position	342.4±4.65µs	348.5±6.53µs	+1.78%
tokens_at_position	3.6±0.06ms	3.7±0.12ms	+2.78%
tokens_for_file	402.8±6.06µs	409.3±4.34µs	+1.61%
traverse	34.5±0.67ms	34.8±0.94ms	+0.87%

[github-actions]

Benchmark for 03bb5de

Click to view benchmark

Test	Base	PR	%
code_action	5.1±0.09ms	5.1±0.14ms	0.00%
code_lens	292.2±10.52ns	290.5±9.92ns	-0.58%
compile	2.7±0.06s	2.7±0.05s	0.00%
completion	4.6±0.24ms	4.5±0.05ms	-2.17%
did_change_with_caching	2.6±0.03s	2.5±0.04s	-3.85%
document_symbol	844.3±17.39µs	910.8±38.67µs	+7.88%
format	67.1±0.58ms	68.1±0.56ms	+1.49%
goto_definition	345.5±10.73µs	339.8±5.41µs	-1.65%
highlight	8.8±0.13ms	8.8±0.05ms	0.00%
hover	501.1±6.40µs	497.8±6.70µs	-0.66%
idents_at_position	118.4±0.46µs	118.9±1.01µs	+0.42%
inlay_hints	635.0±14.50µs	639.1±23.07µs	+0.65%
on_enter	1914.3±35.52ns	2.1±0.04µs	+9.70%
parent_decl_at_position	3.6±0.06ms	3.6±0.03ms	0.00%
prepare_rename	341.1±5.01µs	341.0±10.00µs	-0.03%
rename	9.0±0.03ms	9.2±0.39ms	+2.22%
semantic_tokens	1181.1±16.24µs	1211.2±7.64µs	+2.55%
token_at_position	339.3±2.75µs	336.2±4.42µs	-0.91%
tokens_at_position	3.6±0.03ms	3.6±0.04ms	0.00%
tokens_for_file	405.5±2.82µs	404.8±3.35µs	-0.17%
traverse	35.1±0.95ms	35.0±0.68ms	-0.28%

[tritao]

LGTM, left just some minor suggestion.

[github-actions]

Benchmark for 061859d

Click to view benchmark

Test	Base	PR	%
code_action	5.3±0.08ms	5.3±0.12ms	0.00%
code_lens	286.1±7.42ns	284.7±10.30ns	-0.49%
compile	2.7±0.08s	2.7±0.09s	0.00%
completion	4.8±0.13ms	4.7±0.10ms	-2.08%
did_change_with_caching	2.6±0.04s	2.6±0.04s	0.00%
document_symbol	912.8±22.21µs	883.3±50.70µs	-3.23%
format	69.7±1.09ms	70.1±1.15ms	+0.57%
goto_definition	342.7±6.67µs	340.5±7.02µs	-0.64%
highlight	9.1±0.06ms	9.1±0.04ms	0.00%
hover	494.7±5.58µs	495.8±8.60µs	+0.22%
idents_at_position	117.8±0.53µs	117.2±0.67µs	-0.51%
inlay_hints	641.3±6.91µs	641.4±8.07µs	+0.02%
on_enter	1985.6±76.27ns	2.1±0.06µs	+5.76%
parent_decl_at_position	3.7±0.03ms	3.7±0.03ms	0.00%
prepare_rename	338.4±6.87µs	338.7±7.94µs	+0.09%
rename	9.5±0.16ms	9.4±0.12ms	-1.05%
semantic_tokens	1189.3±12.65µs	1223.9±10.79µs	+2.91%
token_at_position	327.1±2.62µs	334.6±2.57µs	+2.29%
tokens_at_position	3.7±0.05ms	3.7±0.05ms	0.00%
tokens_for_file	402.1±2.22µs	405.6±2.64µs	+0.87%
traverse	34.5±0.76ms	34.2±0.98ms	-0.87%

[github-actions]

Benchmark for e322267

Click to view benchmark

Test	Base	PR	%
code_action	5.2±0.20ms	5.4±0.24ms	+3.85%
code_lens	283.4±7.48ns	288.8±17.63ns	+1.91%
compile	2.9±0.03s	2.8±0.04s	-3.45%
completion	4.7±0.22ms	4.6±0.21ms	-2.13%
did_change_with_caching	2.7±0.07s	2.7±0.05s	0.00%
document_symbol	900.2±46.20µs	944.5±43.40µs	+4.92%
format	71.5±1.09ms	70.5±1.01ms	-1.40%
goto_definition	339.5±6.84µs	350.1±7.28µs	+3.12%
highlight	8.8±0.21ms	8.9±0.18ms	+1.14%
hover	491.9±8.04µs	504.9±10.02µs	+2.64%
idents_at_position	118.2±0.77µs	118.7±1.63µs	+0.42%
inlay_hints	635.4±25.20µs	642.3±25.70µs	+1.09%
on_enter	2.0±0.08µs	2.0±0.08µs	0.00%
parent_decl_at_position	3.7±0.13ms	3.6±0.07ms	-2.70%
prepare_rename	341.5±5.50µs	347.0±4.73µs	+1.61%
rename	9.1±0.20ms	9.2±0.33ms	+1.10%
semantic_tokens	1187.0±60.11µs	1227.4±23.32µs	+3.40%
token_at_position	339.7±3.83µs	341.2±5.09µs	+0.44%
tokens_at_position	3.6±0.09ms	3.6±0.07ms	0.00%
tokens_for_file	404.2±5.22µs	406.5±5.86µs	+0.57%
traverse	36.4±1.06ms	36.1±0.79ms	-0.82%

[github-actions]

Benchmark for 08445df

Click to view benchmark

Test	Base	PR	%
code_action	5.1±0.10ms	5.1±0.01ms	0.00%
code_lens	283.7±5.44ns	284.3±5.52ns	+0.21%
compile	2.7±0.06s	2.7±0.05s	0.00%
completion	4.5±0.05ms	4.6±0.01ms	+2.22%
did_change_with_caching	2.6±0.02s	2.6±0.05s	0.00%
document_symbol	886.5±16.34µs	915.6±25.10µs	+3.28%
format	69.4±0.85ms	69.7±0.92ms	+0.43%
goto_definition	339.4±3.42µs	348.5±3.40µs	+2.68%
highlight	8.7±0.10ms	8.7±0.01ms	0.00%
hover	495.4±6.32µs	506.9±5.61µs	+2.32%
idents_at_position	121.9±0.22µs	118.5±0.75µs	-2.79%
inlay_hints	634.1±27.16µs	629.7±7.57µs	-0.69%
on_enter	1879.0±23.41ns	2.0±0.06µs	+6.44%
parent_decl_at_position	3.6±0.03ms	3.6±0.02ms	0.00%
prepare_rename	340.7±9.21µs	344.2±6.79µs	+1.03%
rename	9.0±0.12ms	9.1±0.28ms	+1.11%
semantic_tokens	1203.4±12.31µs	1260.6±18.20µs	+4.75%
token_at_position	334.5±2.57µs	339.2±2.50µs	+1.41%
tokens_at_position	3.6±0.06ms	3.6±0.03ms	0.00%
tokens_for_file	401.1±5.95µs	403.4±2.20µs	+0.57%
traverse	34.3±1.11ms	34.3±0.74ms	0.00%

[github-actions]

Benchmark for d00b206

Click to view benchmark

Test	Base	PR	%
code_action	5.2±0.13ms	5.3±0.13ms	+1.92%
code_lens	288.0±5.73ns	289.2±9.21ns	+0.42%
compile	2.8±0.06s	2.8±0.08s	0.00%
completion	4.6±0.08ms	4.8±0.15ms	+4.35%
did_change_with_caching	2.7±0.05s	2.7±0.08s	0.00%
document_symbol	935.2±21.35µs	910.9±37.42µs	-2.60%
format	72.5±0.91ms	72.1±0.68ms	-0.55%
goto_definition	337.6±6.43µs	334.4±7.63µs	-0.95%
highlight	8.7±0.09ms	9.2±0.67ms	+5.75%
hover	489.0±5.03µs	486.4±4.03µs	-0.53%
idents_at_position	117.6±0.35µs	118.0±0.41µs	+0.34%
inlay_hints	620.9±22.54µs	637.4±33.91µs	+2.66%
on_enter	2.0±0.03µs	2.0±0.02µs	0.00%
parent_decl_at_position	3.6±0.03ms	3.7±0.09ms	+2.78%
prepare_rename	338.8±6.69µs	337.8±6.93µs	-0.30%
rename	9.6±0.11ms	9.4±0.19ms	-2.08%
semantic_tokens	1152.9±13.79µs	1268.4±16.32µs	+10.02%
token_at_position	332.2±3.41µs	335.5±3.02µs	+0.99%
tokens_at_position	3.6±0.02ms	3.7±0.07ms	+2.78%
tokens_for_file	398.9±4.33µs	403.3±3.10µs	+1.10%
traverse	35.6±0.77ms	35.8±0.77ms	+0.56%

[github-actions]

Benchmark for 5021abe

Click to view benchmark

Test	Base	PR	%
code_action	5.3±0.14ms	5.1±0.15ms	-3.77%
code_lens	284.7±7.89ns	287.9±16.80ns	+1.12%
compile	2.9±0.04s	2.8±0.05s	-3.45%
completion	4.7±0.14ms	4.6±0.11ms	-2.13%
did_change_with_caching	2.8±0.03s	2.6±0.03s	-7.14%
document_symbol	877.4±29.14µs	854.7±23.29µs	-2.59%
format	75.4±1.42ms	71.6±1.27ms	-5.04%
goto_definition	337.5±9.10µs	339.9±8.05µs	+0.71%
highlight	9.0±0.16ms	8.8±0.13ms	-2.22%
hover	496.2±8.56µs	495.0±6.07µs	-0.24%
idents_at_position	117.3±0.29µs	116.9±0.74µs	-0.34%
inlay_hints	629.7±25.20µs	629.7±17.11µs	0.00%
on_enter	2.0±0.10µs	1915.8±44.39ns	-4.21%
parent_decl_at_position	3.7±0.06ms	3.6±0.04ms	-2.70%
prepare_rename	336.0±5.61µs	338.4±6.12µs	+0.71%
rename	9.6±0.23ms	9.1±0.14ms	-5.21%
semantic_tokens	1189.5±26.60µs	1244.4±16.75µs	+4.62%
token_at_position	332.5±3.20µs	411.2±2.72µs	+23.67%
tokens_at_position	3.7±0.05ms	3.6±0.05ms	-2.70%
tokens_for_file	401.5±3.08µs	402.8±3.71µs	+0.32%
traverse	36.6±0.97ms	36.8±0.72ms	+0.55%

[github-actions]

Benchmark for 0b582b3

Click to view benchmark

Test	Base	PR	%
code_action	5.2±0.03ms	5.1±0.04ms	-1.92%
code_lens	287.6±9.07ns	284.9±9.42ns	-0.94%
compile	2.7±0.05s	2.8±0.03s	+3.70%
completion	4.7±0.06ms	4.6±0.06ms	-2.13%
did_change_with_caching	2.6±0.03s	2.7±0.03s	+3.85%
document_symbol	924.8±33.12µs	868.9±34.88µs	-6.04%
format	71.4±0.94ms	71.6±2.59ms	+0.28%
goto_definition	336.8±6.89µs	337.3±5.36µs	+0.15%
highlight	9.0±0.01ms	8.7±0.16ms	-3.33%
hover	492.5±7.47µs	493.3±4.43µs	+0.16%
idents_at_position	118.7±0.37µs	118.2±0.71µs	-0.42%
inlay_hints	639.8±21.74µs	635.8±28.04µs	-0.63%
on_enter	2.0±0.05µs	2.0±0.09µs	0.00%
parent_decl_at_position	3.7±0.02ms	3.6±0.03ms	-2.70%
prepare_rename	337.9±6.76µs	342.4±6.63µs	+1.33%
rename	9.6±1.05ms	9.0±0.11ms	-6.25%
semantic_tokens	1233.1±12.78µs	1211.0±10.01µs	-1.79%
token_at_position	330.3±1.70µs	344.4±2.07µs	+4.27%
tokens_at_position	3.7±0.04ms	3.5±0.03ms	-5.41%
tokens_for_file	400.3±2.63µs	397.2±5.61µs	-0.77%
traverse	34.9±0.65ms	35.2±0.51ms	+0.86%