ci: run Scalpel shadow comparison for all PRs, add depth=50 step

[gnodet]

Summary

• Run Scalpel for all PRs — not just those changing pom.xml. Scalpel detects source file changes (SOURCE_CHANGE, TEST_CHANGE) and distinguishes test-only changes from main source changes, building confidence for a future switch to Scalpel-driven builds.
• Add depth=50 first step — recent PRs where the merge base is close avoid fetching 200 commits upfront. Progressive depths: 50 → 200 → 1000 → full.
• Fix Scalpel comparison — compare against full current detection (file-path + grep + -amd), not just grep POM. Show modules each approach finds that the other misses.
• Fix reactor list parsing — long module names (e.g. "CSimple Maven Plugin (deprecated)") had no dots in Maven output, leaving SUCCESS/FAILURE attached to the name → duplicates in the "All tested modules" list.

One-line summary format

The Scalpel shadow comparison now shows:

🔬 Scalpel shadow comparison — Scalpel: 10 tested, 29 compile-only — current: 39 all tested

This tells reviewers at a glance: Scalpel would only run tests on 10 modules instead of 39, skipping tests for 29 meta/generated modules while still compiling them.

Changes

File	Change
pr-build-main.yml	Deepening with 50 → 200 → 1000 → full (was 200 → 1000 → full)
incremental-build.sh	Remove pom.xml condition — Scalpel runs for every PR. Fix comparison to use full reactor. Fix reactor name parsing.
CI-ARCHITECTURE.md	Update docs

Test plan

• Java-only PR: verify Scalpel shadow comparison appears, no duplicates in module list
• POM-changing PR: verify Scalpel comparison shows correct set differences

Claude Code on behalf of Guillaume Nodet

🤖 Generated with Claude Code

[github-actions]

🌟 Thank you for your contribution to the Apache Camel project! 🌟
🤖 CI automation will test this PR automatically.

🐫 Apache Camel Committers, please review the following items:

• First-time contributors require MANUAL approval for the GitHub Actions to run
• You can use the command /component-test (camel-)component-name1 (camel-)component-name2.. to request a test from the test bot although they are normally detected and executed by CI.
• You can label PRs using skip-tests and test-dependents to fine-tune the checks executed by this PR.
• Build and test logs are available in the summary page. Only Apache Camel committers have access to the summary.

⚠️ Be careful when sharing logs. Review their contents before sharing them publicly.

[oscerd]

LGTM. Low-risk CI change, and the trust model is sound: on: pull_request (not pull_request_target), permissions: contents: read, persist-credentials: false, and no untrusted PR input is interpolated into run: steps — comment posting is offloaded to a separate workflow_run. CI-ARCHITECTURE.md is updated and matches the code. Scalpel already runs on main; this only widens when it's invoked, and it's defensively coded (guards merge-base reachability, catches mvn validate failure, never fails the build).

Two small non-blocking notes:

• The step-2b comment calls this "shadow mode … does not affect actual test execution," but step 2c unions scalpel_module_ids into dep_module_ids, which drives the real -pl build. The effect is conservative — it only adds modules, never removes, so it can't cause under-testing — but the wording could mislead a future maintainer.
• Dropping the grep … pom.xml gate partially relaxes CAMEL-23565 / #23371 (root-pom-only PRs will now run Scalpel and may enlarge the tested set). This looks intentional given the "build confidence for a future switch" goal — a one-line acknowledgement that it relaxes #23371 would help the next reader.

---

Reviewed with Claude Code on behalf of Andrea Cosentino. This review was generated by an AI agent and may contain inaccuracies; please verify all suggestions before applying.