CI: revert helgrind --history-level=full (ineffective)

--history-level=full did not reduce the helgrind race reports (still 84
FAIL, identical frames). Root cause: MHD signals its worker thread via an
internal ITC pipe, not a pthread primitive, so NO helgrind history level
can observe that happens-before — the reports are inherent to MHD's design,
not an approx-history artifact. Revert to the ax_valgrind_check default
(approx, faster). The valgrind lanes will be addressed with third-party
suppressions instead (tracked separately).

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

Author: etr

.github/workflows/verify-build.yml

@@ -1395,19 +1395,7 @@ jobs:
       run: |
         cd build ;
         TOOL="${BUILD_TYPE#valgrind-}" ;
-        # Override ax_valgrind_check's --history-level=approx default for
-        # helgrind. approx drops the happens-before history, so helgrind can't
-        # see the synchronisation MHD_start_daemon()'s pthread_create
-        # establishes between the main thread's pre-start setup and the worker
-        # threads reading that immutable-after-start state during dispatch,
-        # producing a flood of benign "data race" reports on libhttpserver
-        # frames whose conflicting access was "the start of the thread". full
-        # tracks the complete graph and proves those reads ordered, removing
-        # the false positives at the source (no libhttpserver-frame suppression,
-        # per DR-008). A make command-line assignment overrides the Makefile's
-        # ?= default and propagates to the recursive check-TESTS make; it is
-        # inert for the memcheck/drd tools that don't consume it.
-        make VALGRIND_helgrind_FLAGS=--history-level=full "check-valgrind-${TOOL}" ;
+        make "check-valgrind-${TOOL}" ;
       if: ${{ startsWith(matrix.build-type, 'valgrind-') }}
 
     - name: Print Valgrind results