[jdbc.read] Remove value caching from JDBC reading machinery

[alexander-yakushev]

As discussed in #179.

[codecov]

Codecov Report

Attention: Patch coverage is 50.00000% with 4 lines in your changes missing coverage. Please review.

Project coverage is 83.18%. Comparing base (86929ea) to head (3cf911b).
Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
src/toucan2/jdbc/read.clj	33.33%	4 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #183      +/-   ##
==========================================
- Coverage   83.20%   83.18%   -0.03%     
==========================================
  Files          37       37              
  Lines        2524     2503      -21     
  Branches      214      213       -1     
==========================================
- Hits         2100     2082      -18     
+ Misses        210      208       -2     
+ Partials      214      213       -1     

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

[alexander-yakushev]

Weird that code coverage says some lines aren't tested. Any test that triggers fetching values from a row should cover them, right?

[camsaul]

It would be really nice to see some Criterium benchmarks around these changes, like "fetching a million rows with select-reducible used to take a median 10ms but now it takes 9ms" or something like that. Otherwise I'm just approving a PR to delete caching code without a clear sense of how it affects performance

[camsaul]

Ok so after thinking about this a bit you are totally right in that caching stuff at the thunk layer is pointless because TransientRow has its own caching logic as well. So we can go ahead and merge this in