chore(build): bump toolchain to 2022-10-27

[BugenZhao]

Co-Authored-By: xxchan [email protected]
Signed-off-by: Bugen Zhao [email protected]

I hereby agree to the terms of the Singularity Data, Inc. Contributor License Agreement.

What's changed and what's your intention?

This PR is mostly a replication of #5119. We use a workaround for zip_order_key to make it compile.

Checklist

• I have written necessary rustdoc comments
• I have added necessary unit tests and integration tests
• All checks passed in ./risedev check (or alias, ./risedev c)

Refer to a related PR or issue link (optional)

close #5977

[CLAassistant]

All committers have signed the CLA.

[skyzh]

@MrCroxx I've workaround the file cache lifetime problem. You should fix it if you have time later.

[wangrunji0408]

@BugenZhao Any idea why building simulation timed out on CI? 🥵

[BugenZhao]

@BugenZhao Any idea why building simulation timed out on CI? 🥵

Nope. 🥺 Does it mean we should write code really carefully after this PR? 🥵

[wangrunji0408]

Building nexmark_chaos.rs become extremely slow in the new toolchain. Need help from rustc experts 🆘

[skyzh]

Thanks for identifying 🤣

[BugenZhao]

Building nexmark_chaos.rs become extremely slow in the new toolchain. Need help from rustc experts 🆘

There shouldn't be much black magic inside. Perhaps const generic?

[skyzh]

Probably yes. Also why don't we merge hello, nexmark_chaos, nexmark_q4 into one single binary?

[BugenZhao]

That's okay. But I think there's no problem to separate integrated tests into multiple files? 🤣 This should be a best practice of Rust.

[BugenZhao]

Still don't think we should get this merged before the compiler fixes the issue. It's abnormal to put such effort or attention on these hacks while programming. 🤔 The attempts to bump still deserve acknowledgment though.

[skyzh]

Still don't think we should get this merged before the compiler fixes the issue. It's abnormal to put such effort or attention on these hacks while programming. 🤔

I think it's more like we are abusing the type system. Sometimes we just insert random 'a at some random place and we even don't know why it works.

[BugenZhao]

Agree on some storage stuffs, but it makes little sense on such a simple integrated test. 🤣

[skyzh]

Agree on some storage stuffs, but it makes little sense on such a simple integrated test. 🤣

The potential problems lying in the integrated test are something I've been fixing before. All interfaces exposed by lib and not needed for inlining should have BoxFuture. Otherwise it will be recompiled in every binary. Let's see if it improves compile speed.

[skyzh]

I feel like it's also a bug of lifetime inference. FnMut seems to be the root cause for slow integration test compile. Will reproduce later. Let's workaround it first.

[skyzh]

Removing impl<ToString> and impl<IntoIterator> seem to fix everything.

[codecov]

Codecov Report

Merging #6025 (70665fa) into main (fe021b0) will increase coverage by 0.06%.
The diff coverage is 83.33%.

@@            Coverage Diff             @@
##             main    #6025      +/-   ##
==========================================
+ Coverage   74.64%   74.70%   +0.06%     
==========================================
  Files         931      933       +2     
  Lines      148816   149009     +193     
==========================================
+ Hits       111078   111312     +234     
+ Misses      37738    37697      -41     

Flag	Coverage Δ
rust	74.70% <83.33%> (+0.06%)	⬆️

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

Impacted Files	Coverage Δ
src/batch/src/execution/grpc_exchange.rs	55.55% <ø> (ø)
src/batch/src/execution/local_exchange.rs	56.84% <ø> (ø)
src/batch/src/executor/test_utils.rs	67.15% <ø> (ø)
src/batch/src/lib.rs	100.00% <ø> (ø)
src/batch/src/task/broadcast_channel.rs	0.00% <ø> (ø)
src/batch/src/task/env.rs	31.81% <0.00%> (ø)
src/batch/src/task/fifo_channel.rs	86.11% <ø> (ø)
src/batch/src/task/hash_shuffle_channel.rs	0.00% <ø> (ø)
src/cmd_all/src/bin/risingwave.rs	2.63% <ø> (ø)
src/common/src/lib.rs	100.00% <ø> (ø)
... and 100 more

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

[BugenZhao]

Removing impl<ToString> and impl<IntoIterator> seem to fix everything.

This is unreasonable. 🤣🤣 There should be tons of occurrences in well-known library crates of them. Why do they still behave well? 🤔