Introduce ctl::set and ctl::map

We now have a C++ red-black tree implementation that implements standard
template library compatible APIs while compiling 10x faster than libcxx.
It's not as beautiful as the red-black tree implementation in Plinko but
this will get the job done and the test proves it upholds all invariants

This change also restores CheckForMemoryLeaks() support and fixes a real
actual bug I discovered with Doug Lea's dlmalloc_inspect_all() function.

third_party/dlmalloc/README.cosmo

@@ -9,6 +9,7 @@ LICENSE
 
 LOCAL CHANGES
 
+  - Fix bug in dlmalloc_inspect_all()
   - Define dlmalloc_requires_more_vespene_gas()
   - Make dlmalloc scalable using sched_getcpu()
   - Use faster two power roundup for memalign()

third_party/dlmalloc/dlmalloc.c

@@ -25,6 +25,7 @@
 #include "third_party/dlmalloc/vespene.internal.h"
 #include "libc/thread/tls.h"
 #include "libc/intrin/kprintf.h"
+#include "libc/intrin/kprintf.h"
 #include "third_party/nsync/mu.h"
 
 #if !IsTiny()
@@ -1221,7 +1222,8 @@ static void internal_inspect_all(mstate m,
           }
         }
         if (start < (void*)next)  /* skip if all space is bookkeeping */
-          handler(start, next, used, arg);
+          if (start != s) /* [jart] fix phantom alloc bug w/ mspace+mmap */
+            handler(start, next, used, arg);
         if (q == top)
           break;
         q = next;

third_party/dlmalloc/threaded.inc

@@ -25,6 +25,8 @@
 #include "libc/nexgen32e/x86feature.h"
 #include "libc/runtime/runtime.h"
 #include "libc/thread/thread.h"
+#include "libc/runtime/runtime.h"
+#include "libc/intrin/weaken.h"
 #include "third_party/dlmalloc/dlmalloc.h"
 
 #if !FOOTERS || !MSPACES
@@ -64,11 +66,28 @@ size_t dlbulk_free(void *array[], size_t nelem) {
   return 0;
 }
 
+struct ThreadedMallocVisitor {
+  mstate heap;
+  void (*handler)(void *start, void *end,
+                  size_t used_bytes, void *arg);
+  void *arg;
+};
+
+static void threaded_malloc_visitor(void *start, void *end,
+                                    size_t used_bytes, void *arg) {
+  struct ThreadedMallocVisitor *tmv = arg;
+  if (start == tmv->heap)
+    return;
+  tmv->handler(start, end, used_bytes, tmv->arg);
+}
+
 void dlmalloc_inspect_all(void handler(void *start, void *end,
-                                       size_t used_bytes, void *callback_arg),
+                                       size_t used_bytes, void *arg),
                           void *arg) {
-  for (unsigned i = 0; i < g_heapslen; ++i)
-    mspace_inspect_all(g_heaps[i], handler, arg);
+  for (unsigned i = 0; i < g_heapslen; ++i) {
+    struct ThreadedMallocVisitor tmv = {g_heaps[i], handler, arg};
+    mspace_inspect_all(g_heaps[i], threaded_malloc_visitor, &tmv);
+  }
 }
 
 forceinline mstate get_arena(void) {