Speed up escaped byte buf parsing

[juntyr]

The fuzzer times out when parsing a 100k long unicode string because checking for escaped strings is slow. This PR makes two small changes to speed this case up by ~ x3

• when skipping byte strings in the struct type test, don't recheck for base64-encoded strings
• use single-char matchers instead of closure based ones for finding '"' and '\\'

* [ ] I've included my change in CHANGELOG.md

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 99.39%. Comparing base (e8ca8c5) to head (1769934).

:exclamation: Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #534 +/- ## =========================================== - Coverage 100.00% 99.39% -0.61% =========================================== Files 85 85 Lines 14658 14271 -387 =========================================== - Hits 14658 14184 -474 - Misses 0 87 +87 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[juntyr]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.39%. Comparing base (e8ca8c5) to head (1769934).

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files ☔ View full report in Codecov by Sentry. 📢 Have feedback on the report? Share it here.

The codecov regression seems unrelated and mostly occurs in derives and test macros - it should be handled separately