From: Kostya Serebryany <kcc@google.com>
Date: Tue, 8 Nov 2016 21:57:37 +0000 (+0000)
Subject: [libFuzzer] minor docs update
X-Git-Url: https://granicus.if.org/sourcecode?a=commitdiff_plain;h=9a9bc142bf538f7bc7562f9ac29ccc3fe0779a9a;p=llvm

[libFuzzer] minor docs update

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@286299 91177308-0d34-0410-b5e6-96231b3b80d8
---

diff --git a/docs/LibFuzzer.rst b/docs/LibFuzzer.rst
index cf49c8def01..a496af26630 100644
--- a/docs/LibFuzzer.rst
+++ b/docs/LibFuzzer.rst
@@ -72,7 +72,7 @@ Like this:
   }
 
 Note that this fuzz target does not depend on libFuzzer in any way
-ans so it is possible and even desirable to use it with other fuzzing engines
+and so it is possible and even desirable to use it with other fuzzing engines
 e.g. AFL_ and/or Radamsa_.
 
 Some important things to remember about fuzz targets:
@@ -81,8 +81,8 @@ Some important things to remember about fuzz targets:
 * It must tolerate any kind of input (empty, huge, malformed, etc).
 * It must not `exit()` on any input.
 * It may use threads but ideally all threads should be joined at the end of the function.
-* It must be as deterministic as possible. Non-determinism (e.g. random decisions not based on the input byte) will make fuzzing inefficient.
-* It must be fast. Try avoiding cubic or greater complexity.
+* It must be as deterministic as possible. Non-determinism (e.g. random decisions not based on the input bytes) will make fuzzing inefficient.
+* It must be fast. Try avoiding cubic or greater complexity, logging, or excessive memory consumption.
 * Ideally, it should not modify any global state (although that's not strict).