feat!: support dataProperty with ParallelRequestExecutor

[stuebingerb]

Adds proper handling for dataProperties to ParallelRequestExecutor and gets rid of DataLoaderPreparedRequestExecutor, along with the option to configure the request executor in SchemaConfiguration and related DataLoaderPreparedRequestExecutor-specific options.

This also consolidates some tests that were hardcoded to use one of the request executors, and therefore missed the other, and extends coverage for dataProperties a bit. Further, this reduces visibility of some functions to internal to reduce namespace pollution.

Resolves #58
Resolves #268

BREAKING CHANGE: The option to configure request executors has been removed, and execution will now always use the ParallelRequestExecutor.
BREAKING CHANGE: The option to configure a timeout has been removed, as it was only respected by DataLoaderPreparedRequestExecutor.
BREAKING CHANGE: The possibility to pass ExecutionOptions to schema execution has been removed, as it was only respected by DataLoaderPreparedRequestExecutor.
BREAKING CHANGE: The DataLoaderPreparedRequestExecutor, Executor, and ExecutionOptions classes have been removed.

[stuebingerb]

This feels way too easy but there seems to be only one test that actually validates DataLoader specifics (DataLoaderTest#data loader cache per request only).

Already open for review but I want to have a bit more test coverage.

Edit: basic data loader validates as well but I still feel like I'm missing something.

[codecov]

Codecov Report

❌ Patch coverage is 92.13483% with 7 lines in your changes missing coverage. Please review.
✅ Project coverage is 77.32%. Comparing base (1cb755d) to head (0d499f8).

Files with missing lines	Patch %	Lines
...raphql/schema/execution/ParallelRequestExecutor.kt	93.24%	0 Missing and 5 partials ⚠️
...n/com/apurebase/kgraphql/schema/execution/Merge.kt	81.81%	0 Missing and 2 partials ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #305      +/-   ##
==========================================
+ Coverage   76.20%   77.32%   +1.12%     
==========================================
  Files         155      152       -3     
  Lines        5047     4843     -204     
  Branches      896      843      -53     
==========================================
- Hits         3846     3745     -101     
+ Misses        861      781      -80     
+ Partials      340      317      -23     

☔ 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.

[stuebingerb]

Removing the one request executor that already works with kotlinx.serialization seems bad for #224 but my gut feeling is that this would still be a step in the right direction. Both executors have diverged and I believe it is easier to migrate the existing one than solving existing issues with DataLoaderPreparedRequestExecutor and adding missing functionality there. Feel free to disagree, though, happy to have a discussion.

[stuebingerb]

Back with some more test coverage and an implementation that looks a bit more promising, albeit not very polished.

[stuebingerb]

That might be an issue. Just sharing it here for now as I'm not sure if I am able to analyze further before my absence:

jvm summary:
Benchmark                                  (withNewExecutor)  (withSuspendResolvers)  Mode  Cnt  Score    Error  Units
ParallelExecutionBenchmark.queryBenchmark               true                    true  avgt    5  0,015 ±  0,001   s/op
ParallelExecutionBenchmark.queryBenchmark               true                   false  avgt    5  3,762 ±  0,035   s/op
ParallelExecutionBenchmark.queryBenchmark              false                    true  avgt    5  0,009 ±  0,001   s/op
ParallelExecutionBenchmark.queryBenchmark              false                   false  avgt    5  0,314 ±  0,005   s/op

"withNewExecutor" either selects the ParallelRequestExecutor of this PR or the code from the commit before.

[stuebingerb]

Better. Just needs a bit more polishing than before:

jvm summary:
Benchmark                                  (withNewExecutor)  (withSuspendResolvers)  Mode  Cnt  Score    Error  Units
ParallelExecutionBenchmark.queryBenchmark               true                    true  avgt    5  0,006 ±  0,001   s/op
ParallelExecutionBenchmark.queryBenchmark               true                   false  avgt    5  0,315 ±  0,002   s/op
ParallelExecutionBenchmark.queryBenchmark              false                    true  avgt    5  0,010 ±  0,001   s/op
ParallelExecutionBenchmark.queryBenchmark              false                   false  avgt    5  0,310 ±  0,002   s/op