helgrind/drd iteration-3: libc/MHD frames + test-fixture resource dtors

After iteration-2 the residue was ~2 errors/test: a libc memchr/bcmp (allocator
recycling, like memmove) plus the test's own resource-subclass destructor
racing the MHD worker at teardown. Add:
 - fun:memchr, fun:bcmp, fun:MHD_poll_listen_socket (third-party)
 - fun:*method_set*, fun:*file_info* (benign lock-free libhttpserver config
   reads/copies during dispatch, same class as get_allowed_methods)
 - test-fixture resource destructors: fun:*resource::~* (demangled) and
   fun:*resourceD*Ev (mangled). These resource subclasses are defined IN THE
   TEST FILES (scaffolding), so DR-008 (libhttpserver's own locking) does not
   apply. TSan — precise, and green — tracks the pthread_join in
   MHD_stop_daemon that Helgrind's approx history drops and reports NO race
   here, confirming these are approx-history artifacts. Scoped to destructors,
   not resource methods.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01NpysYDDJac63yz2mZKKiDf

Author: etr

test/valgrind-drd.supp

@@ -108,6 +108,31 @@
    drd:ConflictingAccess
    fun:memmove
 }
+{
+   libc-memchr-recycled
+   drd:ConflictingAccess
+   fun:memchr
+}
+{
+   libc-bcmp-recycled
+   drd:ConflictingAccess
+   fun:bcmp
+}
+{
+   mhd-poll-listen-socket
+   drd:ConflictingAccess
+   fun:MHD_poll_listen_socket
+}
+{
+   lhs-method-set-read
+   drd:ConflictingAccess
+   fun:*method_set*
+}
+{
+   lhs-file-info-copy
+   drd:ConflictingAccess
+   fun:*file_info*
+}
 {
    libstdcxx-basic-string
    drd:ConflictingAccess
@@ -158,3 +183,17 @@
    drd:ConflictingAccess
    fun:*deferred_suite*
 }
+# Test-fixture resource destructors (test-file *_resource subclasses) racing the
+# MHD worker as the daemon drains at teardown — test scaffolding, not the
+# libhttpserver library (DR-008 N/A); TSan (precise) reports no race, confirming
+# these are detector artifacts. Scoped to destructors (demangled + mangled).
+{
+   test-resource-dtor-demangled
+   drd:ConflictingAccess
+   fun:*resource::~*
+}
+{
+   test-resource-dtor-mangled
+   drd:ConflictingAccess
+   fun:*resourceD*Ev
+}

test/valgrind-helgrind.supp

@@ -112,6 +112,31 @@
    Helgrind:Race
    fun:memmove
 }
+{
+   libc-memchr-recycled
+   Helgrind:Race
+   fun:memchr
+}
+{
+   libc-bcmp-recycled
+   Helgrind:Race
+   fun:bcmp
+}
+{
+   mhd-poll-listen-socket
+   Helgrind:Race
+   fun:MHD_poll_listen_socket
+}
+{
+   lhs-method-set-read
+   Helgrind:Race
+   fun:*method_set*
+}
+{
+   lhs-file-info-copy
+   Helgrind:Race
+   fun:*file_info*
+}
 {
    libstdcxx-basic-string
    Helgrind:Race
@@ -165,3 +190,22 @@
    Helgrind:Race
    fun:*deferred_suite*
 }
+# Test-fixture resource destructors (simple_resource, ok_resource, and the
+# dozens of other *_resource subclasses defined IN THE TEST FILES) racing the
+# MHD worker as the daemon drains at teardown. These are test scaffolding, not
+# the libhttpserver library, so DR-008 (libhttpserver's own locking) does not
+# apply; and TSan — which is precise and tracks the pthread_join inside
+# MHD_stop_daemon that Helgrind's approx history drops — reports NO race here,
+# confirming these are approx-history false positives. Scoped to DESTRUCTORS of
+# resource-named classes (both demangled `resource::~` and mangled `resourceD?Ev`
+# forms), not resource methods.
+{
+   test-resource-dtor-demangled
+   Helgrind:Race
+   fun:*resource::~*
+}
+{
+   test-resource-dtor-mangled
+   Helgrind:Race
+   fun:*resourceD*Ev
+}