debugger: By default, only save non-tail-recursive calls

By default, the debugger use to save all calls on its simulated
stack. That could facilitate finding errors, but it could also
mean that the Debugger could become very slow while executing
tail-recursive code.

1 files changed, 8 insertions, 10 deletions

diff --git a/lib/debugger/doc/src/debugger_chapter.xml b/lib/debugger/doc/src/debugger_chapter.xml
index 67e95cd083..2d812b0236 100644
--- a/lib/debugger/doc/src/debugger_chapter.xml
+++ b/lib/debugger/doc/src/debugger_chapter.xml
@@ -274,17 +274,15 @@ c_break(Bindings) ->
       <seealso marker="#attach">the Attach Process window</seealso>.
     </p>
 
-    <p>By default, the Debugger saves information about all current
+    <p>By default, the Debugger only saves information about recursive
       function calls, that is, function calls that have not yet returned
-      a value (option 'Stack On, Tail').</p>
-
-    <p>This means, however, that information is saved also for tail
-      recursive calls. For example, repeated calls to the <c>loop</c>
-      function of an Erlang process. This may consume unnecessary
-      amounts of memory for debugged processes with long lifetimes and
-      many tail recursive calls. It is therefore possible to set
-      the option 'Stack On, no tail', in which case information about
-      previous calls are discarded when a tail recursive call is made.
+      a value (option 'Stack On, No Tail').</p>
+
+    <p>Sometimes, however, it can be useful to save all calls, even
+      tail-recursive calls. That can be done with the 'Stack On, Tail'
+      option. Note that this option will consume more memory and slow
+      down execution of interpreted functions when there are many
+      tail-recursive calls.
     </p>
 
     <p>It is also possible to turn off the Debugger stack trace