reuse some standard functions; eliminate multiple passes in a few places

[philnguyen]

This commit attempts to clean up and micro-optimize some by:

• reusing library functions where possible (e.g. dropf-right, hash-update!, hash-ref!)
• avoiding repeatedly calling linear-time functions (e.g. length, list-ref)

[rfindler]

Are these changes covered by the existing test suites?

[philnguyen]

Not that I'm aware of :D. I did a search for search/next and it didn't seem to be directly tested. Are functions in this module indirectly tested by tests on judgment-form-based random generation? I can try adding some tests if there are guidelines about how these functions with random behaviors are tested.

Otherwise it's not really important. I was just interested in how this thing is tuned in practice and made some edits along the way to help myself understand.

[rfindler]

Thanks, @philnguyen. I'll look into testing the changes more carefully later on; they look good to me, but I wouldn't want to merge them without being able to rely on just our eyes looking at the code.

[rfindler]

Indeed, the existing test suites get to all of the functions that are changed.

Thanks!!