refactor(error): reduce error type size

[wangrunji0408]

I hereby agree to the terms of the RisingWave Labs, Inc. Contributor License Agreement.

What's changed and what's your intention?

This PR reduces size of the following error types:

Error	Before Size	After Size
StreamError	240	8
StreamExecutorError	240	8
ExprError	184	40
UdfError	184	32
RpcError	176	16

by boxing the error body and offloading them to the heap.

The common pattern of this technique is:

pub struct Error {
    inner: Box<Inner>,
}
struct Inner {
    kind: ErrorKind,
    backtrace: Backtrace,
}
enum ErrorKind {
    Storage(StorageError),
    ...
}

This partially resolves #11263.

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• I have added fuzzing tests or opened an issue to track them. (Optional, recommended for new SQL features Sqlsmith: Sql feature generation #7934).
• My PR contains breaking changes. (If it deprecates some features, please create a tracking issue to remove them in the future).
• All checks passed in ./risedev check (or alias, ./risedev c)
• My PR changes performance-critical code. (Please run macro/micro-benchmarks and show the results.)

• My PR contains critical fixes that are necessary to be merged into the latest release. (Please check out the details)

Documentation

• My PR contains user-facing changes.

[wangrunji0408]

I'm considering whether the ExprError is also worth reducing to 8 bytes. Because it's used in scalar functions which are also performance sensitive.

[xxchan]

🔥🥵

• merge StreamExecutorError and StreamError: not sure. What’s the reason? I generally prefer more fine-grained error types, if no much trouble introduced, because it made clear what’s possible errors for a function.
• XxError -> Error. Not sure. Personally slightly prefer the former, but no strong opinions.

Others lgtm

[wangrunji0408]

merge StreamExecutorError and StreamError: not sure. What’s the reason? I generally prefer more fine-grained error types, if no much trouble introduced, because it made clear what’s possible errors for a function.

I found the main body of stream crate is the executors. Having two error types cause double memory usage in most cases. And I prefer one unified error type in one crate.

XxError -> Error. Not sure. Personally slightly prefer the former, but no strong opinions.

In most 3rd party libraries and std library, their error types are named Error. 🤔

[BugenZhao]

And I prefer one unified error type in one crate.

Personally, I prefer to keep the granularity as fine as possible. 😰 Currently the difference between StreamError and StreamExecutorError is kind of clear. The former is for controllers and managers, and the latter is for executor computations.

Similar ideas are applied to expression as well. -> #10764

[TennyZhuang]

I'd prefer split Error instead of merge Error.

[wangrunji0408]

Reverted merging and renaming. PTAL again. 🤡

[codecov]

Codecov Report

Merging #11275 (f4d371e) into main (a84d8e1) will decrease coverage by 0.01%.
Report is 1 commits behind head on main.
The diff coverage is 19.19%.

@@            Coverage Diff             @@
##             main   #11275      +/-   ##
==========================================
- Coverage   69.79%   69.78%   -0.01%     
==========================================
  Files        1359     1360       +1     
  Lines      225851   225880      +29     
==========================================
+ Hits       157630   157635       +5     
- Misses      68221    68245      +24     

Flag	Coverage Δ
rust	69.78% <19.19%> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files Changed	Coverage Δ
src/batch/src/executor/hop_window.rs	75.24% <0.00%> (-0.25%)	⬇️
src/expr/src/error.rs	21.05% <0.00%> (ø)
src/expr/src/expr/test_utils.rs	84.51% <0.00%> (-2.24%)	⬇️
src/expr/src/vector_op/date_trunc.rs	6.25% <0.00%> (ø)
src/expr/src/vector_op/encdec.rs	72.72% <0.00%> (ø)
src/expr/src/vector_op/jsonb_info.rs	15.38% <0.00%> (ø)
src/expr/src/vector_op/overlay.rs	94.02% <0.00%> (ø)
src/expr/src/vector_op/trim_array.rs	4.54% <0.00%> (ø)
...tend/src/optimizer/plan_node/generic/hop_window.rs	91.32% <0.00%> (-1.15%)	⬇️
src/rpc_client/src/error.rs	8.33% <0.00%> (-8.34%)	⬇️
... and 12 more

... and 1 file with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[BugenZhao]

Rest LGTM

[BugenZhao]

LGTM! Really appreciate.

[MrCroxx]

Nice! Looking forward to seeing the performance!