Replace substring BufferBuilders with Vec

[cakeni]

Which issue does this PR close?

Part of #10245.

Rationale for this change

Vec provides the same capacity preallocation and append behavior needed by substring_by_char_impl without the additional BufferBuilder abstraction.

What changes are included in this PR?

• Replace the byte-value BufferBuilder with Vec<u8>.
• Replace the offset BufferBuilder with Vec<OffsetSize>.
• Preserve the existing capacity estimates, null handling, and offset construction.

Are there any user-facing changes?

No.

How was this change tested?

• cargo +stable-x86_64-pc-windows-gnu fmt --all -- --check
• cargo +stable-x86_64-pc-windows-gnu clippy -p arrow-string --all-targets --all-features --no-deps -- -D warnings
• cargo +stable-x86_64-pc-windows-gnu test -p arrow-string --all-features (182 unit tests and 10 doc tests passed)

[Jefffrey]

@cakeni i noticed you raised quite a lot of PRs in a short span of time; especially these PRs have the same timestamps:

• [Variant] Support list paths in ShreddedSchemaBuilder #10635
• [Arrow] Move parquet-variant-compute StructArrayBuilder to arrow-array #10636
• [Variant] Add nested object and array interoperability tests #10637

could you explain this? are they truly PRs raised by you or was there an AI agent involved with this?

[cakeni]

Yes, AI assistance was involved in preparing some of these PRs. I reviewed the changes, diffs, and test results before submitting them.

The similar timestamps are because I was working on several independent issues in parallel and submitted the PRs around the same time.

I understand that opening several PRs at once can make review harder for maintainers, so I’ll slow down and wait for feedback on the existing PRs before opening more.