perf(schema): optimize convertType with fast paths for compatible types

[baagod]

Implement fast paths in convertType to avoid expensive slice traversal when types are identical or have compatible underlying types (e.g. json.RawMessage to []byte).

[codecov]

Codecov Report

❌ Patch coverage is 87.50000% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 92.58%. Comparing base (88c6cda) to head (62bdeaa).
⚠️ Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
schema.go	87.50%	2 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #931      +/-   ##
==========================================
- Coverage   92.62%   92.58%   -0.04%     
==========================================
  Files          23       23              
  Lines        4499     4503       +4     
==========================================
+ Hits         4167     4169       +2     
- Misses        274      275       +1     
- Partials       58       59       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[wolveix]

Thanks so much for this optimization! I spent some time putting tests and benchmarks together (with Claude's help) to compare this versus the original implementation. If you'd like to run the benchmarks yourself, extract this zip and run ./run.sh.
bench.zip

---

Unfortunately I noticed that your optimization caused a +32.36% increase for same type vars, but that was easy to remedy :)

Here were my results with your original implementation:

goos: darwin
goarch: arm64
pkg: tmp/bench-pr
cpu: Apple M3 Pro
                                 │  v1_norm.txt  │             v2_norm.txt             │
                                 │    sec/op     │   sec/op     vs base                │
ConvertType_Nil-12                   1.902n ± 4%   1.399n ± 3%  -26.44% (p=0.000 n=10)
ConvertType_SameType-12              3.228n ± 4%   4.272n ± 0%  +32.36% (p=0.000 n=10)
ConvertType_IntToInt64-12            24.29n ± 1%   19.43n ± 1%  -20.01% (p=0.000 n=10)
ConvertType_IntToPtr-12              52.59n ± 3%   57.52n ± 1%   +9.35% (p=0.000 n=10)
ConvertType_RawMessageToBytes-12   1635.00n ± 1%   94.26n ± 1%  -94.23% (p=0.000 n=10)
ConvertType_SmallSlice-12            582.2n ± 1%   588.7n ± 0%   +1.10% (p=0.000 n=10)
ConvertType_LargeSlice-12            51.30µ ± 1%   50.57µ ± 0%   -1.43% (p=0.000 n=10)
geomean                              121.2n        78.73n       -35.02%

                                 │  v1_norm.txt   │              v2_norm.txt               │
                                 │      B/op      │     B/op      vs base                  │
ConvertType_Nil-12                   0.000 ± 0%       0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_SameType-12              0.000 ± 0%       0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToInt64-12            8.000 ± 0%       8.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToPtr-12              16.00 ± 0%       16.00 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_RawMessageToBytes-12    855.00 ± 0%       24.00 ± 0%  -97.19% (p=0.000 n=10)
ConvertType_SmallSlice-12            448.0 ± 0%       448.0 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_LargeSlice-12          39.30Ki ± 0%     39.30Ki ± 0%        ~ (p=1.000 n=10) ¹
geomean                                         ²                 -39.98%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

                                 │  v1_norm.txt  │              v2_norm.txt              │
                                 │   allocs/op   │  allocs/op   vs base                  │
ConvertType_Nil-12                  0.000 ± 0%      0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_SameType-12             0.000 ± 0%      0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToInt64-12           1.000 ± 0%      1.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToPtr-12             2.000 ± 0%      2.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_RawMessageToBytes-12   65.000 ± 0%      1.000 ± 0%  -98.46% (p=0.000 n=10)
ConvertType_SmallSlice-12           23.00 ± 0%      23.00 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_LargeSlice-12          2.003k ± 0%     2.003k ± 0%        ~ (p=1.000 n=10) ¹
geomean                                        ²                -44.92%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

And here's the slight improvement with my further optimization:

goos: darwin
goarch: arm64
pkg: tmp/bench-pr
cpu: Apple M3 Pro
                                 │  v1_norm.txt  │             v2_norm.txt             │
                                 │    sec/op     │   sec/op     vs base                │
ConvertType_Nil-12                   1.866n ± 2%   1.412n ± 3%  -24.30% (p=0.000 n=10)
ConvertType_SameType-12              3.213n ± 2%   2.667n ± 1%  -17.00% (p=0.000 n=10)
ConvertType_IntToInt64-12            24.40n ± 2%   19.06n ± 1%  -21.89% (p=0.000 n=10)
ConvertType_IntToPtr-12              51.65n ± 0%   56.73n ± 2%   +9.84% (p=0.000 n=10)
ConvertType_RawMessageToBytes-12   1573.00n ± 0%   92.69n ± 1%  -94.11% (p=0.000 n=10)
ConvertType_SmallSlice-12            576.2n ± 1%   573.4n ± 1%        ~ (p=0.128 n=10)
ConvertType_LargeSlice-12            51.09µ ± 0%   51.11µ ± 3%        ~ (p=0.781 n=10)
geomean                              119.6n        73.01n       -38.95%

                                 │  v1_norm.txt   │              v2_norm.txt               │
                                 │      B/op      │     B/op      vs base                  │
ConvertType_Nil-12                   0.000 ± 0%       0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_SameType-12              0.000 ± 0%       0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToInt64-12            8.000 ± 0%       8.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToPtr-12              16.00 ± 0%       16.00 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_RawMessageToBytes-12    855.00 ± 0%       24.00 ± 0%  -97.19% (p=0.000 n=10)
ConvertType_SmallSlice-12            448.0 ± 0%       448.0 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_LargeSlice-12          39.30Ki ± 0%     39.30Ki ± 0%        ~ (p=1.000 n=10) ¹
geomean                                         ²                 -39.98%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

                                 │  v1_norm.txt  │              v2_norm.txt              │
                                 │   allocs/op   │  allocs/op   vs base                  │
ConvertType_Nil-12                  0.000 ± 0%      0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_SameType-12             0.000 ± 0%      0.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToInt64-12           1.000 ± 0%      1.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_IntToPtr-12             2.000 ± 0%      2.000 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_RawMessageToBytes-12   65.000 ± 0%      1.000 ± 0%  -98.46% (p=0.000 n=10)
ConvertType_SmallSlice-12           23.00 ± 0%      23.00 ± 0%        ~ (p=1.000 n=10) ¹
ConvertType_LargeSlice-12          2.003k ± 0%     2.003k ± 0%        ~ (p=1.000 n=10) ¹
geomean                                        ²                -44.92%                ²
¹ all samples are equal
² summaries must be >0 to compute geomean

---

Great work! Thanks so much!