Cygwin: Treat api_fatal() similarly to a core-dumping signal

Provide the same debugging opportunities for api_fatal() as we do for a
core-dumping signal:

1) Break into any attached debugger
2) Start JIT debugger (if configured) (keeping these under DEBUGGING doesn't seem helpful)
3) Write a coredump (if rlim_core > 1MB)
4) Write a stackdump (if that failed, or 0 < rlim_core <= 1MB)

winsup/cygwin/dcrt0.cc

@@ -1250,11 +1250,7 @@ vapi_fatal (const char *fmt, va_list ap)
   __small_vsprintf (buf + n, fmt, ap);
   va_end (ap);
   strace.prntf (_STRACE_SYSTEM, NULL, "%s", buf);
-
-#ifdef DEBUGGING
-  try_to_debug ();
-#endif
-  cygwin_stackdump ();
+  api_fatal_debug();
   myself.exit (__api_fatal_exit_val);
 }
 

winsup/cygwin/exceptions.cc

@@ -1396,6 +1396,24 @@ signal_exit (int sig, siginfo_t *si, void *)
 }
 } /* extern "C" */
 
+/* As above, but before exiting due to api_fatal */
+extern "C"
+void
+api_fatal_debug ()
+{
+  if (try_to_debug ())
+    return;
+
+  if (cygheap->rlim_core == 0Ul)
+    return;
+
+  if (cygheap->rlim_core > 1024*1024)
+    if (exec_prepared_command (dumper_command))
+      return;
+
+  cygwin_stackdump();
+}
+
 /* Attempt to carefully handle SIGCONT when we are stopped. */
 void
 _cygtls::handle_SIGCONT ()

winsup/cygwin/local_includes/winsup.h

@@ -181,6 +181,7 @@ void close_all_files (bool = false);
 extern "C" void error_start_init (const char*);
 extern "C" void dumper_init (void);
 extern "C" int try_to_debug ();
+extern "C" void api_fatal_debug ();
 
 void ld_preload ();
 void fixup_hooks_after_fork ();