fix(fqn): cap derived project-name length to avoid ENAMETOOLONG

cbm_project_name_from_path hex-encodes non-ASCII path bytes (#571), but had no
length cap -- a deep CJK / non-ASCII path triples in length past the
filesystem's 255-byte filename-component limit, so <cache>/<name>.db becomes
un-openable (ENAMETOOLONG).

Cap the derived name at FQN_MAX_NAME_LEN (200): names within the cap are
returned UNCHANGED (no drift for already-indexed projects), while a longer name
keeps its first (CAP-9) bytes plus an FNV-1a hash of the full name as a "-%08x"
suffix, so distinct long paths stay distinct. The hex encoding scheme is
unchanged.

Distilled from #624 by Matt Van Horn (the ENAMETOOLONG catch); the rest of that
PR was redundant with the #571 fix already on main and would have swapped the
encoding scheme, causing project-name drift.

Reproduce-first: project_name_length_capped_issue624 builds a deep CJK path
(~376-byte raw name) and asserts the result is <= 200, validator-safe, and that
two long paths differing only in the trailing segment map to different names; a
short CJK path stays byte-identical (no drift). RED (376 > 200) -> GREEN. Full
suite 5738/0.

Signed-off-by: Martin Vogel <martin.vogel.tech@gmail.com>
Co-Authored-By: Matt Van Horn <455140+mvanhorn@users.noreply.github.com>

Author: DeusData

src/pipeline/fqn.c

@@ -10,6 +10,7 @@
 
 #include <stdbool.h>
 #include <stddef.h> // NULL
+#include <stdint.h> // uint32_t
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h> // strdup
@@ -18,6 +19,13 @@
 #define FQN_MAX_PATH_SEGS 254
 #define FQN_MAX_DIR_SEGS 255
 
+/* Max bytes for a derived project name. The name becomes a filename component
+ * ("<cache>/<name>.db" and sidecars ".db-wal"/".db.corrupt"), so it must stay
+ * under the filesystem's 255-byte component limit. 200 leaves headroom for the
+ * longest sidecar suffix. #571 hex-encodes each non-ASCII byte to 2 chars, so a
+ * deep CJK path can triple past 255 and make the DB file un-openable (#624). */
+#define FQN_MAX_NAME_LEN 200
+
 /* ── Internal helpers ─────────────────────────────────────────────── */
 
 /* Build a dot-joined string from segments. Returns heap-allocated string. */
@@ -351,6 +359,31 @@ char *cbm_pipeline_fqn_folder(const char *project, const char *rel_dir) {
     return result;
 }
 
+/* Bound a derived project name to FQN_MAX_NAME_LEN bytes so "<cache>/<name>.db"
+ * stays within the filesystem's 255-byte filename-component limit (#624). Names
+ * within the cap are returned UNCHANGED (no drift). Longer names keep their first
+ * (CAP-9) bytes and get a "-XXXXXXXX" FNV-1a hash of the FULL name appended, so
+ * two long paths that share a prefix but differ later still map to distinct
+ * names. The suffix ends in a hex digit, so the result stays validator-safe. */
+static char *fqn_bound_name_len(char *name) {
+    if (!name) {
+        return name;
+    }
+    size_t n = strlen(name);
+    if (n <= FQN_MAX_NAME_LEN) {
+        return name; /* within cap → unchanged, no drift */
+    }
+    uint32_t h = 2166136261u; /* FNV-1a offset basis over the FULL name */
+    for (size_t i = 0; i < n; i++) {
+        h ^= (unsigned char)name[i];
+        h *= 16777619u;
+    }
+    /* Keep first (CAP-9) bytes + "-" + 8 hex = CAP total. The buffer holds n+1
+     * bytes and n > CAP, so writing 9 chars + NUL at offset (CAP-9) fits. */
+    snprintf(name + (FQN_MAX_NAME_LEN - 9), 10, "-%08x", h);
+    return name;
+}
+
 char *cbm_project_name_from_path(const char *abs_path) {
     if (!abs_path || !abs_path[0]) {
         return strdup("root");
@@ -433,5 +466,8 @@ char *cbm_project_name_from_path(const char *abs_path) {
 
     char *result = strdup(start);
     free(path);
+    if (result) {
+        result = fqn_bound_name_len(result); /* #624: cap filename-component length */
+    }
     return result;
 }

tests/test_fqn.c

@@ -491,6 +491,63 @@ TEST(project_name_encodes_unicode_segments_issue571) {
     PASS();
 }
 
+/* issue #624: #571 preserves non-ASCII path segments by hex-encoding each byte
+ * (1 byte -> 2 hex chars), so a DEEP non-ASCII path triples in length and can
+ * blow past the filesystem's 255-byte filename-component limit. Then
+ * "<cache>/<name>.db" is un-openable (ENAMETOOLONG). The derived name must be
+ * length-capped (with a hash suffix that disambiguates otherwise-identical
+ * prefixes) while staying validator-safe — and SHORT names must NOT drift. */
+#define FQN_CAP_UNDER_TEST 200
+TEST(project_name_length_capped_issue624) {
+    /* 开 = U+5F00, UTF-8 "\xe5\xbc\x80" (3 bytes -> 6 hex chars). 60 copies makes
+     * the raw hex-encoded name ~360 chars, well past the 200-byte cap. */
+    static const char KAI[] = "\xe5\xbc\x80";
+    char deepA[512];
+    char deepB[512];
+    const char *prefix = "/Users/dev/";
+    size_t p = strlen(prefix);
+    memcpy(deepA, prefix, p);
+    for (int i = 0; i < 60; i++) {
+        memcpy(deepA + p, KAI, 3);
+        p += 3;
+    }
+    /* deepB shares the entire deep prefix and differs ONLY in the trailing seg */
+    memcpy(deepB, deepA, p);
+    memcpy(deepA + p, "/alpha", 7); /* includes NUL */
+    memcpy(deepB + p, "/omega", 7);
+
+    char *nameA = cbm_project_name_from_path(deepA);
+    char *nameB = cbm_project_name_from_path(deepB);
+    ASSERT_NOT_NULL(nameA);
+    ASSERT_NOT_NULL(nameB);
+
+    /* (a) capped: name must fit within the filename-component budget */
+    ASSERT_LTE(strlen(nameA), FQN_CAP_UNDER_TEST);
+    ASSERT_LTE(strlen(nameB), FQN_CAP_UNDER_TEST);
+    /* (b) still a valid project name (resolve_store must accept it) */
+    ASSERT_TRUE(cbm_validate_project_name(nameA));
+    ASSERT_TRUE(cbm_validate_project_name(nameB));
+    /* (c) two deep paths differing only in the trailing segment must NOT
+     * collide after capping — the hash suffix disambiguates them. */
+    ASSERT_STR_NEQ(nameA, nameB);
+
+    free(nameA);
+    free(nameB);
+
+    /* Short CJK path (the issue #571 case) must be UNCHANGED — no drift. */
+    char *shortName = cbm_project_name_from_path(
+        "/Users/yunxin/Desktop/\xe5\xbc\x80\xe5\x8f\x91/"
+        "\xe5\x90\x8e\xe7\xab\xaf/"
+        "\xe4\xbf\xa1\xe7\xa7\x9f\xe9\xa3\x8e\xe6\x8e\xa7\xe9\x80\x9a\xe5\x90\x8e\xe7\xab\xaf");
+    ASSERT_NOT_NULL(shortName);
+    ASSERT_LTE(strlen(shortName), FQN_CAP_UNDER_TEST);
+    ASSERT_STR_EQ(shortName, "Users-yunxin-Desktop-e5bc80e58f91-e5908ee7abaf-"
+                             "e4bfa1e7a79fe9a38ee68ea7e9809ae5908ee7abaf");
+    free(shortName);
+
+    PASS();
+}
+
 /* ================================================================
  * Suite
  * ================================================================ */
@@ -593,6 +650,7 @@ SUITE(fqn) {
     RUN_TEST(project_name_deep_path);
     RUN_TEST(project_name_always_validator_safe_issue349);
     RUN_TEST(project_name_encodes_unicode_segments_issue571);
+    RUN_TEST(project_name_length_capped_issue624);
     RUN_TEST(project_name_colon_only);
     RUN_TEST(project_name_backslash_only);
     RUN_TEST(project_name_consecutive_colons);