online: Disable parallel sorting to prevent memleak from libTBB

algo/base/BuildInfo.h

@@ -6,6 +6,10 @@
 
 #include <string>
 
+#if __has_include(<execution>) && !defined(__CLING__)
+#include <execution>  // for feature test macro __cpp_lib_parallel_algorithm
+#endif
+
 #if defined(HAVE_TBB) && defined(__cpp_lib_parallel_algorithm)
 #define HAVE_PARALLEL_ALGORITHM
 #endif

algo/base/compat/Algorithm.h

@@ -19,8 +19,7 @@
 
 #include <algorithm>
 
-#if defined(__cpp_lib_execution) && defined(HAVE_PARALLEL_ALGORITHM)
-#define WITH_EXECUTION
+#ifdef HAVE_PARALLEL_ALGORITHM
 #include <execution>
 #endif
 
@@ -36,7 +35,11 @@ namespace cbm::algo
   template<typename It, typename Compare>
   void Sort(It first, It last, Compare comp)
   {
-#ifdef WITH_EXECUTION
+// Disable parallel sorting for the moment
+// The underlying implementation in libTBB has a massive memory leak:
+// https://community.intel.com/t5/Intel-oneAPI-Threading-Building/std-sort-std-execution-par-unseq-has-a-memory-leak-on-Linux/m-p/1580910
+#if 0
+// #ifdef HAVE_PARALLEL_ALGORITHM
     std::sort(std::execution::par_unseq, first, last, comp);
 #else
     std::sort(first, last, comp);