Division by zero in libtheora-1.0

Hello everybody,

while developing an application for the company I work for, I noticed 
that very often libtheora reports division by zero errors. I fixed those 
errors by adding a few additional checks, but I'm not 100% sure that 
these checks are correct.

Could you please check out the attached patch file and tell me what you 
think?

Best regards,
Marco Giovannini
-- 
/*
  *========================================================
  * Marco Giovannini
  * Software Engineer
  * ------------------------------------------------------
  * TRADERLINK S.r.l.
  * Via IV Giugno, 84
  * Republic of San Marino
  * Tel. (+378) 0549 900757
  * ------------------------------------------------------
  * http://www.traderlink.it
  *========================================================
  */

_______________________________________________
theora-dev mailing list
[email protected]
http://lists.xiph.org/mailman/listinfo/theora-dev

Index: lib/enc/encoder_toplevel.c
===================================================================
--- lib/enc/encoder_toplevel.c	(revision 15558)
+++ lib/enc/encoder_toplevel.c	(working copy)
@@ -705,7 +705,7 @@
   /* Calculate the Q Modifier step size required to cause a step down
      from full target bandwidth to 40% of target between max Q and
      best Q */
-  QModStep = 0.5 / (double)((Q_TABLE_SIZE - 1) -
+  QModStep = 0.5 / (double)((Q_TABLE_SIZE) -
                             cpi->Configuration.ActiveMaxQ);

   /* Set up the cpi->QTargetModifier[] table. */
@@ -755,6 +755,9 @@
                                 Configuration.TargetBandwidth)) * 0.0175) ));
   }

+  if (cpi->ThisFrameTargetBytes <= 0)
+    cpi->ThisFrameTargetBytes = 1;
+
   if ( !DropFrame ) {
     /*  pick all the macroblock modes and motion vectors */
     ogg_uint32_t InterError;