sourcecode 1.72.0__tar.gz → 2.0.0__tar.gz

{sourcecode-1.72.0 → sourcecode-2.0.0}/.gitignore

@@ -35,6 +35,10 @@ Thumbs.db
 
 .planning
 
+# ContextGraph baseline — full canonical outputs (multi-MB, from external repos).
+# The light oracle (index.json) and convergence.json ARE committed.
+tests/contextgraph_baseline/baseline/artifacts/
+
 # Supabase local temp
 supabase/.temp/
 # Internal agent playbook (not distributed)

{sourcecode-1.72.0 → sourcecode-2.0.0}/CHANGELOG.md

@@ -1,5 +1,102 @@
 # Changelog
 
+## [2.0.0] — 2026-07-06
+
+**ASK Engine 2.0** — the product identity is now final. The engine is unchanged;
+this major marks the CLI/product rename and is fully backward compatible.
+
+### Changed — canonical command is now `ask`
+
+- **`ask` is the canonical CLI command.** `ask impact`, `ask migrate-check`,
+  `ask review-pr`, `ask mcp init`, `ask activate …` — one short verb that matches
+  the product name (ASK Engine).
+- **`sourcecode` becomes a deprecated compatibility alias.** It keeps working
+  (same single implementation, `prog_name` pinned to `ask`), prints a one-line
+  deprecation notice **only on an interactive terminal**, and forwards to `ask`.
+  Pipes and agents that still call `sourcecode` get byte-identical stdout **and**
+  clean (JSON-only) stderr — the machine contract is unchanged. The alias will be
+  removed in a future major.
+- **New welcome wordmark**: a large `ASK` block with an `E N G I N E` label
+  replaces the old `SOURCE`/`CODE` banner. Help and banners read `CLI: ask`.
+- **Config home moved to `~/.ask`** (canonical writes); the legacy `~/.sourcecode`
+  is still read as a fallback, so **no existing user loses their license or config**.
+  Cache dir (`.sourcecode-cache/`) and `SOURCECODE_*` env vars are unchanged.
+- **MCP**: `ask mcp init` registers the client entry under `ask`; a pre-existing
+  `sourcecode` entry is recognized and migrated (and both are removed on
+  `mcp remove`).
+- **Docs**: README, user guide, manual, tiers, demo and onboarding now say
+  **ASK Engine** and use `ask …` in every example. Package/install/config and
+  compatibility references are preserved and the compatibility strategy is
+  documented in `docs/PRODUCT_IDENTITY.md` (normative).
+
+### Compatibility
+
+- **PyPI/Homebrew distribution stays `sourcecode` this release**; installing it
+  provides both the `ask` and `sourcecode` commands. Renaming the distribution to
+  `ask-engine` is a separate, explicitly-breaking future change.
+- Python package/import name (`sourcecode`), MCP protocol identity, cache format,
+  and all serialized/JSON outputs are unchanged.
+
+### Also in this line (identity + precision, pre-2.0 increments)
+
+- **Impact/`review-pr`**: a changed repository/service now surfaces **every**
+  affected entry point (controller) instead of collapsing an N-controller blast
+  radius to one — bounded by `max_impacts`, no new semantic fact, hash-neutral.
+- **Product identity consolidation**: audited every `sourcecode` occurrence and
+  established `docs/PRODUCT_IDENTITY.md` as the normative product↔CLI mapping.
+
+## [1.73.0] — 2026-07-01
+
+### Fixed — audit reliability on a large non-Boot Spring monolith (Alfresco Community Repo field test)
+
+A real architecture audit of **alfresco-community-repo** (`fe6dd26c3f`; 6,549
+classes, 24 Maven modules, Spring 7 **without Spring Boot**, iBatis persistence,
+Camel/ActiveMQ messaging, Acegi-derived security, WebScript REST surface)
+surfaced seven defects — each capable of leading an automated consultant to a
+false conclusion. All fixed at root cause with regression coverage and
+re-verified on the live repo.
+
+- **`--compact` did not resolve `language_version` from a Maven property (BUG #1).**
+  Alfresco declares `<maven.compiler.source>${java.version}</...>` chained to
+  `<java.version>21</java.version>`; the Java-stack detector picked the first key
+  and emitted the literal `${java.version}` while `migrate-check` resolved it. A
+  shared `substitute_maven_properties` resolver (in `detectors/parsers.py`) now
+  backs BOTH paths — `--compact` reports `21`.
+- **`explain --output` wrote Markdown to a `.json` file with no warning (BUG #2).**
+  Broke any consumer assuming JSON parity across subcommands. `explain` now infers
+  JSON from a `.json` output extension (an explicit `--format` always wins), and
+  `--help` documents that its default output is human-readable text/Markdown.
+- **`entry_points.bootstrap` false positives on `*Application.java` by name (BUG #3).**
+  Alfresco's XSD/JAXB-generated `Application.java` model classes (no `main()`, no
+  bootstrap annotation, in a WAR) were listed as application entry points. A file
+  now qualifies only with a real `main(String[])` or a bootstrap annotation
+  (`@SpringBootApplication` / WAR `SpringBootServletInitializer` / `@QuarkusMain`
+  / …); the serializer's name-based bootstrap fallback was removed.
+- **"MyBatis" framework declared against the tool's own zero evidence (BUG #4).**
+  A `org.mybatis:mybatis` pom coordinate alone produced a "MyBatis" framework label
+  while the `mybatis` block reported `mapper_interfaces: 0, xml_files: 0` (Alfresco's
+  SQL layer is iBatis-style `*-SqlMap.xml`, surfaced separately). MyBatis is now
+  emitted only with real usage evidence — an `@Mapper` interface or a `*Mapper.xml`.
+- **`export --integrations` missed ActiveMQ/Camel messaging (BUG #5).**
+  JMS detection recognized little beyond a direct `JmsTemplate`, reporting 1
+  integration on a repo with an ActiveMQ connection-factory config and Camel routes.
+  Added `ActiveMQSslConnectionFactory` / pooled / caching connection-factory tokens
+  and an Apache Camel route pass (literal `activemq:`/`jms:` endpoint URIs, plus
+  variable-URI `from()/to()` routes inside a `RouteBuilder` as honest low-confidence
+  `camel-route` messaging records). Alfresco JMS/messaging surface: 1 → 5.
+- **`readiness_score` Hibernate false-positive — regression test (BUG #6).**
+  v1.63.0 once scored this Hibernate-free repo 42/100 with a phantom
+  `hibernate_rewrite` headline (~46 days). Already fixed by v1.72.0; added an
+  explicit regression test (`no Hibernate/JPA imports → hibernate.detected False,
+  headline_blocker never hibernate_rewrite`) so a future detector refactor cannot
+  reintroduce it.
+- **`endpoints` / C4 `api_surface` returned 0 with no explanation (BUG #7).**
+  On an obvious API server (WebScripts: XML descriptor + handler class, not
+  `@RestController`), a bare `total: 0` reads as "no public API". `endpoints` now
+  emits a first-class `zero_result_reason` naming the routing patterns searched and
+  any non-Spring surface detected (Alfresco: 482 WebScript descriptors / 901 handler
+  classes), and the C4 export's `api_surface` carries the same explanation.
+
 ## [1.72.0] — 2026-07-01
 
 ### Fixed — correctness on a large real-world Spring Boot 3 monolith (Broadleaf Commerce CE field test)

{sourcecode-1.72.0 → sourcecode-2.0.0}/CONTRIBUTING.md

@@ -30,8 +30,8 @@ This installs `sourcecode` in editable mode with all development dependencies (`
 ### Run the CLI locally
 
 ```bash
-sourcecode . --agent
-sourcecode . prepare-context explain
+ask . --agent
+ask . prepare-context explain
 ```
 
 ### Run tests
@@ -150,7 +150,7 @@ Keep commits atomic. One logical change per commit.
 ## Reporting bugs
 
 Open an issue with:
-- `sourcecode --version` output
+- `ask --version` output
 - Command you ran
 - Expected behavior
 - Actual behavior / error output