note about things still to do with RAND_bytes

diff --git a/CHANGES b/CHANGES
index 48053f22012dce13ad6eae4c5d2868a800460302..326b5cf7a1dfbb3f90df067b5980d6a3ab385942 100644 (file)
--- a/CHANGES
+++ b/CHANGES
@@ -5,8 +5,13 @@
  Changes between 0.9.4 and 0.9.5  [xx XXX 1999]
 
   *) Precautions against using the PRNG uninitialized: RAND_bytes() now
-     has a return value which indicated the quality of the random data
-     (1 = ok, 0 = not seeded).
+     has a return value which indicates the quality of the random data
+     (1 = ok, 0 = not seeded).  Also an error is recorded on the thread's
+     error queue.
+     (TO DO: always check the result of RAND_bytes when it is used in the
+     library, because leaving the error in the error queue but reporting
+     success in a function that uses RAND_bytes could confuse things
+     considerably.)
      [Ulf Möller]
 
   *) Do more iterations of Rabin-Miller probable prime test (specifically,