Use delegate-attr to replace delegate

[upsuper]

Issue #, if available:

Description of changes: use delegate-attr macro to replace delegate, which should have a nicer syntax.

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.

[zslayton]

Hi, thanks for the PR! At a glance, it seems like these libraries are a wash for our use case; the bulk of this PR removes a level of indentation. Are there other benefits to using the delegate-attr crate over delegate that we should consider?

[codecov]

Codecov Report

Merging #275 (dd73a7c) into main (d1768e8) will increase coverage by 0.05%.
The diff coverage is 100.00%.

❗ Current head dd73a7c differs from pull request most recent head 3ada6a2. Consider uploading reports for the commit 3ada6a2 to get more accurate results

@@            Coverage Diff             @@
##             main     #275      +/-   ##
==========================================
+ Coverage   91.44%   91.49%   +0.05%     
==========================================
  Files          56       55       -1     
  Lines        8241     8093     -148     
==========================================
- Hits         7536     7405     -131     
+ Misses        705      688      -17     

Impacted Files	Coverage Δ
src/binary/cursor.rs	91.98% <ø> (ø)
src/reader.rs	94.81% <100.00%> (-0.04%)	⬇️
src/binary/type_code.rs	81.35% <0.00%> (-6.78%)	⬇️
ion-hash/src/lib.rs	96.87% <0.00%> (-3.13%)	⬇️
ion-c-sys/src/string.rs	82.92% <0.00%> (-1.97%)	⬇️
ion-c-sys/src/writer.rs	90.84% <0.00%> (-1.31%)	⬇️
src/binary/timestamp.rs
src/binary/writer.rs	94.66% <0.00%> (+0.16%)	⬆️
ion-hash/src/representation.rs	88.23% <0.00%> (+0.60%)	⬆️
ion-hash/tests/ion_hash_tests.rs	87.93% <0.00%> (+3.08%)	⬆️
... and 1 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d1768e8...3ada6a2. Read the comment docs.

[zslayton]

Our cargo fmt check is failing for the current commit.

[upsuper]

Are there other benefits to using the delegate-attr crate over delegate that we should consider?

I think this way the code looks cleaner, and more friendlier to things like code formatter, as it uses attribute macro rather than function-style macro, for which the formatter can't really know what to do.

Our cargo fmt check is failing for the current commit.

Hmmm, it's very interesting. I think this is a bug of rustfmt, and I'm raising a PR to fix it. In the mean time I think replacing ; with {} would work around this bug. I'll push a new commit to fix it.

[zslayton]

As this achieves the same thing as our current code but with slightly different syntax, I'm inclined to leave our code as-is. If delegate-attr eventually offers other more objective benefits (for example: better performance), I'd be happy to revisit it.

[upsuper]

I don't think there will be much performance difference as it's just a proc macro... I would note that I think delegate-attr should have a better test coverage, though.