Do not report totalFrames from DAP stackTrace request

Currently, gdb will unwind the entire stack in response to the
stackTrace request.  I had erroneously thought that the totalFrames
attribute was required in the response.  However, the spec says:

    If omitted or if `totalFrames` is larger than the available
    frames, a client is expected to request frames until a request
    returns less frames than requested (which indicates the end of the
    stack).

This patch removes this from the response in order to improve
performance when the stack trace is very long.

gdb/python/lib/gdb/dap/bt.py

@@ -50,13 +50,9 @@ def _backtrace(thread_id, levels, startFrame):
         current_frame = gdb.newest_frame()
     except gdb.error:
         current_frame = None
-    # Note that we always iterate over all frames, which is lame, but
-    # seemingly necessary to support the totalFrames response.
-    # FIXME maybe the mildly mysterious note about "monotonically
-    # increasing totalFrames values" would let us fix this.
-    while current_frame is not None:
+    while current_frame is not None and (levels == 0 or len(frames) < levels):
         # This condition handles the startFrame==0 case as well.
-        if current_number >= startFrame and (levels == 0 or len(frames) < levels):
+        if current_number >= startFrame:
             newframe = {
                 "id": frame_id(current_frame),
                 "name": _frame_name(current_frame),
@@ -80,9 +76,11 @@ def _backtrace(thread_id, levels, startFrame):
             frames.append(newframe)
         current_number = current_number + 1
         current_frame = current_frame.older()
+    # Note that we do not calculate totalFrames here.  Its absence
+    # tells the client that it may simply ask for frames until a
+    # response yields fewer frames than requested.
     return {
         "stackFrames": frames,
-        "totalFrames": current_number,
     }