core: meta/veloxpart — resume sidecar, reader first + fuzzed (stage 5)

util/crc32.hpp — header-only CRC-32 (zlib polynomial, reflected), used to
integrity-check the sidecar.

meta/veloxpart — the <name>.veloxpart.meta resume file (docs/04 §5).
Little-endian, versioned, CRC-32 over the whole record. Layout: magic,
version, flags, total_size, downloaded, url set (original/effective/
mirrors), etag/last-modified/content-type, segment records (start, end
INCLUSIVE, completed), optional sha256 streaming-hash blob.

parse_veloxpart() is the attacker-facing surface (the file sits in a
world-writable-ish download dir) and is total on any byte string: CRC
checked before any field is interpreted; magic, a version it understands,
every count and length bounded by a hard cap AND checked against the
remaining buffer; ByteReader latches on overrun; trailing bytes rejected.
Every malformation is meta_corrupt / meta_version_unsupported, never a
crash or an unbounded allocation. serialize_veloxpart() is deterministic
(unchanged sidecar isn't rewritten). File helpers write atomically
(temp + rename) and fdatasync the file and its directory.

Tests: crc32 known vector; full + minimal round-trips; deterministic
serialize; file round-trip; and a truncation/corruption table — bad
magic, CRC mismatch (payload and CRC-field flips), future version,
truncation at every stage, hostile url_count / segment_count / lp_string
length (the case the brief singles out), trailing bytes, impossible
segment.completed.

tools/fuzz/fuzz_veloxpart — feeds raw bytes and bytes-with-valid-CRC
(so the field parser and ByteReader bounds checks are actually reached),
and round-trip-stability-checks anything accepted. Ran 1.1M execs clean
under ASan+UBSan+libFuzzer (clang++-21); fuzz_content_disposition and
fuzz_url likewise re-run to 1.1M. tools/fuzz gains a -runs=0 seed-replay
CTest smoke per target (regression tripwire; the campaign stays manual).

Fuzz-found and fixed: parse_content_disposition could emit a filename
containing NUL / control bytes from a mangled filename* ext-value —
strip_path only removed path separators. Now sanitize_leaf() also drops
C0 controls and DEL (rules/ still owns the authoritative sanitize; `..`
and printable-unsafe content pass through as before).

Co-Authored-By: Claude Sonnet 5 <[email protected]>
Claude-Session: https://claude.ai/code/session_01HPPSGhiArbvQgwC2DNiURS

tools/fuzz/CMakeLists.txt

@@ -19,24 +19,38 @@ endif()
 set(_core ${CMAKE_SOURCE_DIR}/core)
 set(_fuzz_flags -g -O1 -fsanitize=fuzzer,address,undefined -fno-omit-frame-pointer)
 
-function(vdm_add_fuzzer name)
+# corpus_dir is the seed set (checked in, small, hand-named). It is passed read-only:
+# the CTest smoke run uses -runs so libFuzzer exits, and a real campaign is
+# `bin/<name> corpus/<x>/` by hand.
+function(vdm_add_fuzzer name corpus_dir)
     add_executable(${name} ${ARGN})
     target_include_directories(${name} PRIVATE ${_core}/include ${_core}/src)
     target_compile_features(${name} PRIVATE cxx_std_23)
     target_compile_options(${name} PRIVATE ${_fuzz_flags})
     target_link_options(${name} PRIVATE ${_fuzz_flags})
+
+    if(VELOX_BUILD_TESTS)
+        # Regression tripwire only: replay the checked-in seeds once (-runs=0, no
+        # mutation, no corpus writes) so a parser change that breaks a known-good or
+        # known-hostile input fails CI in a fraction of a second. The 1M-exec bar
+        # (AGENT-CORE M1 DoD) is a separate campaign job: `bin/<name> corpus/<x>/`.
+        add_test(NAME ${name}_smoke
+                 COMMAND ${name} -runs=0 ${CMAKE_CURRENT_SOURCE_DIR}/${corpus_dir})
+        set_tests_properties(${name}_smoke PROPERTIES LABELS "fuzz" TIMEOUT 60)
+    endif()
 endfunction()
 
-vdm_add_fuzzer(fuzz_content_disposition
+vdm_add_fuzzer(fuzz_content_disposition corpus/content_disposition
     content_disposition_fuzz.cpp
     ${_core}/src/net/content_disposition.cpp
     ${_core}/src/net/text_codec.cpp)
 
-vdm_add_fuzzer(fuzz_url
+vdm_add_fuzzer(fuzz_url corpus/url
     url_fuzz.cpp
     ${_core}/src/net/url.cpp
     ${_core}/src/net/text_codec.cpp)
 
-# Seed corpora live next to the harnesses.
-file(GLOB _cd_seeds ${CMAKE_CURRENT_SOURCE_DIR}/corpus/content_disposition/*)
-file(GLOB _url_seeds ${CMAKE_CURRENT_SOURCE_DIR}/corpus/url/*)
+vdm_add_fuzzer(fuzz_veloxpart corpus/veloxpart
+    veloxpart_fuzz.cpp
+    ${_core}/src/meta/veloxpart.cpp
+    ${_core}/src/util/error.cpp)