[reference] rxjava-3.0 async blind regeneration for #11927

[jordan-wong]

[reference] rxjava-3.0 blind regeneration for #11927

Reference output from the toolkit's automated instrumentation-authoring
workflow, run against the draft skill in #11927. Not intended to be merged
as-is.

Research context

This PR is part of the async-category research cycle: verifying whether
the current skill draft in #11927 generalizes to Category B
(context-propagation) libraries after the HTTP-cycle skill updates. rxjava-3.0
was chosen because it has a mature existing implementation on master,
making regen output directly comparable.

Changes applied on top of eval output

Original eval branch inherited a stash-merge revert of #11852 during eval
branch setup, dropping DD_TRACE_ROBOLECTRIC_ENABLED from metadata plus 172
unrelated files. Root cause was a stale local checkout, not a toolkit or
skill issue. Rebased onto current master by cherry-picking only in-scope
paths under dd-java-agent/instrumentation/rxjava/rxjava-3.0/.

Original: bd134f34bd → rebased: b46346b45b

Where to focus review

The reference output should be compared against the master implementation
to surface how well the current skill draft preserves regen-critical
details. Known divergences:

1. super(...) alias dropped. Master has super("rxjava", "rxjava-3");
   eval produced super("rxjava"). This silently breaks any customer
   using DD_TRACE_RXJAVA_3_ENABLED as an opt-out toggle. This is the
   canonical Rule First PR - Implements the Opentracing API #2 (regen-preservation) violation the skill should
   prevent.

2. muzzle { fail { ... } } block missing. Master has an explicit
   fail block asserting rxjava3 advice must not match against the
   rxjava2 artifact (rxjava3 and rxjava2 use different Java namespaces
   but the same Maven module name, so this is a real hazard). Eval
   dropped the block entirely.

3. reactive-streams compileOnly version regressed from 1.0.3 to
   1.0.0. Silent narrowing of the tested surface.

Local CI-equivalent (check + muzzle + latestDepTest) is fully
green. The one remaining red check on CI (muzzle: [3/8]) does not
reproduce locally and appears to be a CI-side sharding flake.

Base: master @ 05671ce3b0

[dd-octo-sts]

🟢 Java Benchmark SLOs — All performance SLOs passed

Suite	Status
Startup	🟢 pass

SLO thresholds are defined here based on automatically generated metrics. A warning is raised when results are within 5% of the threshold.

PR vs. master results

Scenario	Candidate	master	Δ (95% CI of mean)
startup:insecure-bank:iast:Agent	13.98 s	14.02 s	[-0.9%; +0.4%] (no difference)
startup:insecure-bank:tracing:Agent	12.87 s	13.08 s	[-2.1%; -1.0%] (maybe better)
startup:petclinic:appsec:Agent	16.43 s	16.71 s	[-5.9%; +2.5%] (no difference)
startup:petclinic:iast:Agent	16.41 s	16.83 s	[-6.7%; +1.7%] (no difference)
startup:petclinic:profiling:Agent	16.25 s	16.80 s	[-7.5%; +0.9%] (no difference)
startup:petclinic:sca:Agent	16.78 s	16.66 s	[-0.3%; +1.7%] (no difference)
startup:petclinic:tracing:Agent	15.60 s	15.64 s	[-6.1%; +5.7%] (unstable)

Commit: b46346b4 · CI Pipeline · Benchmarking Platform UI

---

Load and DaCapo benchmarks can be triggered manually in the GitLab pipeline. Results will appear in the Benchmarking Platform UI after completion.

[jordan-wong]

Rebased 2026-07-14 — branch force-pushed onto current master to fix the shared config-inversion-linter failure.

Root cause (not a toolkit bug): the original eval branch inherited a stash-merge revert of #11852 during branch setup. The merge commit 8481176619 ("pre-rxjava3-clean-branch") combined two parents that individually both had DD_TRACE_ROBOLECTRIC_ENABLED, but the git-plumbing merge (stash-based) silently dropped it along with 172 other unrelated files' worth of #11852 changes. The toolkit workflow then generated its output on that already-broken tree, so the terminal commit inherited the deletion.

Fix: cherry-picked only the in-scope dd-java-agent/instrumentation/rxjava/rxjava-3.0/ paths from the original terminal SHA (bd134f34bd) onto fresh master (05671ce3b0), following the ready-for-review skill Phase 3 recipe. All rxjava3 metadata entries were already present on master, so no metadata change is needed.

Result: 21 files changed (all under rxjava/rxjava-3.0/), no sibling contamination, no unrelated file deletions.

Original commit: bd134f34bd → rebased: b46346b45b

[mcculls]

This is remarkably close, if we ignore some of the additional local knowledge / testing