tewip-ai 0.51.0 → 0.52.0

package/dist/engine/schema/migration-files.js

@@ -0,0 +1,79 @@
+// Which files in a project are migrations at all.
+//
+// Split out of schema-command.ts in Phase 44, when the corpus harness needed to ask the question
+// across a whole checkout rather than across a diff and that file went past the 300-line rule.
+// Keeping it in one place is the point: the alternative was a copy of these patterns in a script,
+// and two definitions of "is this a migration" drift the moment one of them is fixed.
+import { readdirSync, statSync } from 'node:fs';
+import { join, relative } from 'node:path';
+// A directory called `migration/` says what is probably in it, not what every file in it is. This
+// matched *anything* underneath one, so Keycloak's `docs/guides/migration/index.adoc` and its 113
+// `migration/migrators/MigrateTo12_0_0.java` classes - application code that moves data between
+// Keycloak versions, nothing to do with a schema file - were selected as migrations and then listed
+// as ones Tewip could not read. 205 files chosen in order to read 89. A "could not read" list that
+// is mostly documentation teaches the reader to skip the list.
+//
+// So the directory decides *where*, and the extension decides *whether it could hold DDL at all*.
+const MIGRATION_DIR = /(^|\/)(migrations?|migration|db\/migrate|prisma\/migrations|alembic|flyway|changelogs?|sql)\//i;
+const COULD_HOLD_DDL = /\.(sql|py|rb|js|mjs|cjs|ts|php|cs|xml|ya?ml|json|go|exs?)$/i;
+// Flyway's Java migrations are real, and they are named the way Flyway requires: `V2__add_index.java`
+// or `R__rebuild_views.java`. Keycloak's `MigrateTo12_0_0.java` is not one. Taking `.java` on the
+// strength of its directory alone is what filled that list with source files.
+const FLYWAY_JAVA = /(^|\/)[VR][\w.]*__[\w.-]+\.(java|kt|scala)$/;
+export const MIGRATION = {
+    test: (path) => MIGRATION_DIR.test(path) && (COULD_HOLD_DDL.test(path) || FLYWAY_JAVA.test(path)),
+};
+// A changelog is often not in a directory that says so. OpenMRS keeps 44 Liquibase changelogs in
+// `api/src/main/resources/` named `liquibase-*.xml`, and the path filter saw none of them - so a
+// real healthcare system with 44 migration files was told it had none. The name counts too.
+//
+// And a changelog is named for what it is, but hardly ever *only* that. Phase 42 taught this filter
+// to read `liquibase-update-to-latest.xml` and it still missed Keycloak's 83, every one called
+// `jpa-changelog-<version>.xml`: the word had to *begin* the name. Prefixes are the norm once a
+// project has more than one database. Letting a release-notes `CHANGELOG.xml` through costs nothing,
+// because `dialectOf` still has to find `<databaseChangeLog` in it before it is read as a migration.
+export const NAMED_MIGRATION = /(^|\/)[\w.-]*(liquibase|changelog)[\w.-]*\.(xml|ya?ml|json)$/i;
+export const SCHEMA_FILE = /\.(sql)$|(^|\/)(migrations?|migration|db\/migrate)\/.*\.(py|rb|js|ts|php|cs)$/i;
+/**
+ * Every migration in a whole repository, by the same filters a diff is read through.
+ *
+ * `schemaFiles` answers "which of these changed files are migrations", which is the question a pull
+ * request asks. "What migrations does this project have at all" is a different question, and the
+ * corpus harness needs it to re-derive a verified number across a pinned checkout.
+ */
+export function migrationsUnder(root, limit = 20_000) {
+    const skip = new Set(['node_modules', '.git', 'dist', 'build', 'target', 'vendor', '__pycache__', '.venv', 'venv', '.next', '.cache']);
+    const found = [];
+    const walk = (where, depth) => {
+        if (depth > 12 || found.length >= limit)
+            return;
+        let entries;
+        try {
+            entries = readdirSync(where);
+        }
+        catch {
+            return;
+        }
+        for (const entry of entries) {
+            if (skip.has(entry) || entry.startsWith('.'))
+                continue;
+            const full = join(where, entry);
+            let info;
+            try {
+                info = statSync(full);
+            }
+            catch {
+                continue;
+            }
+            if (info.isDirectory()) {
+                walk(full, depth + 1);
+                continue;
+            }
+            const rel = relative(root, full).replaceAll(String.fromCharCode(92), '/');
+            if (SCHEMA_FILE.test(rel) || MIGRATION.test(rel) || NAMED_MIGRATION.test(rel))
+                found.push(rel);
+        }
+    };
+    walk(root, 0);
+    return found.sort();
+}

package/dist/engine/schema/schema-command.js

@@ -13,18 +13,8 @@ import { dirname, resolve } from 'node:path';
 import { changeSummary } from '../quality/grounding.js';
 import { dialectOf, forwardOnly, irreversible, readFramework } from './frameworks.js';
 import { readSql } from './statements.js';
-const MIGRATION = /(^|\/)(migrations?|migration|db\/migrate|prisma\/migrations|alembic|flyway|changelogs?|sql)\//i;
-// A changelog is often not in a directory that says so. OpenMRS keeps 44 Liquibase changelogs in
-// `api/src/main/resources/` named `liquibase-*.xml`, and the path filter saw none of them - so a
-// real healthcare system with 44 migration files was told it had none. The name counts too.
-// A changelog is named for what it is, but hardly ever *only* that. Phase 42 taught this filter to
-// read `liquibase-update-to-latest.xml` and it still missed Keycloak's 83 files, every one of them
-// called `jpa-changelog-<version>.xml`: the name was required to *begin* with the word. Prefixes are
-// the norm - a project with more than one database names each changelog after its module. So the
-// word may sit anywhere in the name. Letting a release-notes `CHANGELOG.xml` through costs nothing,
-// because `dialectOf` still has to find `<databaseChangeLog` in it before it is read as a migration.
-const NAMED_MIGRATION = /(^|\/)[\w.-]*(liquibase|changelog)[\w.-]*\.(xml|ya?ml|json)$/i;
-const SCHEMA_FILE = /\.(sql)$|(^|\/)(migrations?|migration|db\/migrate)\/.*\.(py|rb|js|ts|php|cs)$/i;
+import { MIGRATION, NAMED_MIGRATION, SCHEMA_FILE, migrationsUnder } from './migration-files.js';
+export { migrationsUnder };
 /** The schema files in this change. */
 export function schemaFiles(cmd) {
     const named = cmd.files !== undefined && cmd.files.length > 0;

package/dist/engine/vouch/assertions.js

@@ -30,6 +30,13 @@ const CHECKS = [
     /\bpytest\s*\.\s*(raises|warns|approx|fail)\b/,
     /\bnp\s*\.\s*testing\s*\.\s*assert/,
     /\bassert_(frame|series|index)_equal\b/,
+    // A bare `fail(...)`, which is how JUnit, TestNG and Go spell "this must not be reached" once the
+    // assertion is statically imported. `Assert.fail(...)` was already read by the pattern below;
+    // `fail(...)` on its own was not, so OpenMRS's RefByUuidTest - whose only real check is a `fail()`
+    // in the catch, beside a deliberate `assertTrue(true)` in the try - was called a tautology that
+    // "no change to the code can falsify". Make `getRefByUuid` throw and it falsifies immediately.
+    // The leading class keeps `shouldFail(` and `canFail(` out, and the paren must come next.
+    /(^|[^\w.])(?:Assert(?:ions)?\s*\.\s*)?fail\s*\(/,
     // Java, Kotlin, C#, Scala.
     /\bAssert(ions)?\s*\./,
     /\bassertThat\s*\(/,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "tewip-ai",
-  "version": "0.51.0",
+  "version": "0.52.0",
   "mcpName": "io.github.hmamut39/tewip",
   "description": "Tewip finds why each failing test failed - real bug, flaky test, environment, test data or changed selector - for any stack, from the command line, CI, or your editor.",
   "keywords": [
@@ -63,7 +63,9 @@
     "typecheck": "tsc --noEmit",
     "build": "tsc -p tsconfig.build.json",
     "prepack": "npm run build",
-    "tewip": "node cli/tewip.ts"
+    "tewip": "node cli/tewip.ts",
+    "version": "node scripts/sync-version.ts && git add server.json",
+    "corpus": "node scripts/corpus.ts"
   },
   "license": "Apache-2.0",
   "devDependencies": {