GH-14912: [Java] Remove usage of PlatformDependent in arrow-vector, arrow-jdbc and arrow-algorithm

[gf2121]

• Closes: [Java] Remove usage of PlatformDependent in arrow-vector #14912

[github-actions]

• Closes: [Java] Remove usage of PlatformDependent in arrow-vector #14912

[github-actions]

⚠️ GitHub issue #14912 has been automatically assigned in GitHub to PR creator.

[github-actions]

⚠️ GitHub issue #14912 has no components, please add labels for components.

[lidavidm]

Since we also maintain a copy of sun.misc.Unsafe in MemoryUtil, can we add our own getInt, getLong, etc. there and replace Netty Unsafe with our Unsafe? That way we can at least keep the same behavior as before.

[gf2121]

Thanks @lidavidm for feedback!

I think you are meaning to delegate methods on Unsafe in MemoryUtil (like what PlatformDependent do, just discard the params around netty). But i'm also seeing that ArrowBuf directly replaced PlatformDependent.xxx with MemoryUtil.UNSAFE.xxx in #6131. So I wonder if the logic in PlatformDependent(like this) is still wanted:

• If yes, we can introduce delegating methods in this PR and replaces the direct operation on MemoryUtil#UNSAFE in a follow up.

• If no, we can just operate on MemoryUtil#UNSAFE without introducing extra methods in MemoryUtil.

[lidavidm]

Using MemoryUtil.UNSAFE directly is OK, IMO.

[gf2121]

PlatformDependent was used in arrow-algorithm, arrow-jdbc, arrow-memory-netty, arrow-vector. I've removed the usages in arrow-algorithm, arrow-jdbc and arrow-vector, and removed the dependency of netty-common in arrow-algorithm and arrow-jdbc. It would take a bit more job to remove dependency of netty-common in arrow-vector as it is using some Netty collection utils llike IntObjectMap. Maybe do that in follow up.

[lidavidm]

Thanks!

CC @lwhite1 you may want to glance through this too

[lidavidm]

I filed #14936 for dropping Netty entirely.

[maosuhan]

How about merging this MR first? We are also blocked by this netty unsafe issue. @lidavidm

[lidavidm]

I rebased and will merge after that

[ursabot]

Benchmark runs are scheduled for baseline = 5765aa2 and contender = 5618a47. 5618a47 is a master commit associated with this PR. Results will be available as each benchmark for each run completes.
Conbench compare runs links:
[Failed ⬇️0.0% ⬆️0.0% ⚠️ Contender and baseline run contexts do not match] ec2-t3-xlarge-us-east-2
[Failed ⬇️0.45% ⬆️0.0%] test-mac-arm
[Failed ⬇️0.0% ⬆️0.0% ⚠️ Contender and baseline run contexts do not match] ursa-i9-9960x
[Finished ⬇️0.0% ⬆️0.0% ⚠️ Contender and baseline run contexts do not match] ursa-thinkcentre-m75q
Buildkite builds:
[Failed] 5618a477 ec2-t3-xlarge-us-east-2
[Failed] 5618a477 test-mac-arm
[Failed] 5618a477 ursa-i9-9960x
[Finished] 5618a477 ursa-thinkcentre-m75q
[Finished] 5765aa29 ec2-t3-xlarge-us-east-2
[Failed] 5765aa29 test-mac-arm
[Failed] 5765aa29 ursa-i9-9960x
[Finished] 5765aa29 ursa-thinkcentre-m75q
Supported benchmarks:
ec2-t3-xlarge-us-east-2: Supported benchmark langs: Python, R. Runs only benchmarks with cloud = True
test-mac-arm: Supported benchmark langs: C++, Python, R
ursa-i9-9960x: Supported benchmark langs: Python, R, JavaScript
ursa-thinkcentre-m75q: Supported benchmark langs: C++, Java