Tests: Minor cleanups to OSS-Fuzz files.

Most of these fixes are small typos and tweaks. A few were caused by bad
advice from me. Here is the summary of what is changed:

- Author line edits

- Small comment changes/additions

- Using the return value in the error messages in the fuzz targets'
  coder initialization code

- Removed fuzz_encode_stream.options. This set a max length, which may
  prevent some worthwhile code paths from being properly exercised.

- Removed the max_len option from fuzz_decode_stream.options for the
  same reason as fuzz_encode_stream. The alone decoder fuzz target still
  has this restriction.

- Altered the dictionary contents for fuzz_lzma.dict. Instead of keeping
  the properties static and varying the dictionary size, the properties
  are varied and the dictionary size is kept small. The dictionary size
  doesn't have much impact on the code paths but the properties do.

Closes: https://github.com/tukaani-project/xz/pull/73

1 files changed, 3 insertions, 0 deletions

diff --git a/tests/ossfuzz/Makefile b/tests/ossfuzz/Makefile
index 008cd7df..a25bd0db 100644
--- a/tests/ossfuzz/Makefile
+++ b/tests/ossfuzz/Makefile
@@ -8,5 +8,8 @@ all: $(FUZZ_TARGET_BINS)
 	$(CXX) $(CXXFLAGS) $(LIB_FUZZING_ENGINE) $(<:.c=.o) -o $(OUT)/$@ \
 		../../src/liblzma/.libs/liblzma.a ;
 
+# The generated binaries are not removed, just the object files. The
+# binaries are created to the $(OUT) directory and must be removed by the
+# fuzzing framework.
 clean:
 	rm -f *.o