[or-cvs] r9394: More documentation. (Again, trying to see whether I have bui (in tor/trunk: . src/or)

[nickm@xxxxxxxx]

Author: nickm
Date: 2007-01-23 19:43:52 -0500 (Tue, 23 Jan 2007)
New Revision: 9394

Modified:
   tor/trunk/
   tor/trunk/src/or/circuitbuild.c
Log:
 r11285@catbus:  nickm | 2007-01-23 19:43:50 -0500
 More documentation. (Again, trying to see whether I have buildbot working).



Property changes on: tor/trunk
___________________________________________________________________
 svk:merge ticket from /tor/trunk [r11285] on 8246c3cf-6607-4228-993b-4d95d33730f1

Modified: tor/trunk/src/or/circuitbuild.c
===================================================================
--- tor/trunk/src/or/circuitbuild.c	2007-01-24 00:20:49 UTC (rev 9393)
+++ tor/trunk/src/or/circuitbuild.c	2007-01-24 00:43:52 UTC (rev 9394)
@@ -1921,10 +1921,18 @@
   tor_free(s);
 }
 
-/** DOCDOC */
+/** Called when one or more guards that we would previously have used for some
+ * purpose are no longer in use because a higher-priority guard has become
+ * useable again. */
 static void
 control_event_guard_deferred(void)
 {
+  /* XXXX We don't actually have a good way to figure out _how many_ entries
+   * are live for some purpose.  We need an entry_is_even_slightly_live()
+   * function for this to work right.  NumEntryGuards isn't reliable: if we
+   * need guards with weird properties, we can have more than that number
+   * live.
+   **/
 #if 0
   int n = 0;
   or_options_t *options = get_options();