This version fixes some incompatibilities with recent glibc versions and with libasan.

• new -X flag for hexadecimal dumps
• new “null” operating mode that does not mutate files, for debugging purposes
• many other bugfixes and robustness improvements

• new “copy” operating mode which copies input to temporary files instead of preloading libzzuf into the process
• new -X flag that dumps stdout to hexadecimal
• multiple platform-specific fixes on OS X, Gentoo, platforms with fortified glibc…
• lots of documentation improvements
• rename zzcat to zzat because of namespace conflicts with zziplib
• heavy work on the Win32 port; it’s starting to be usable but still incomplete

• various fixes for Linux, FreeBSD, OS X and OpenBSD libc calls.
• network host filtering support.
• support for memory limits on OS X.
• massive testsuite improvements.
• support most cat options in zzcat, which is now fully programmable.

• no longer crash when the fuzzed application uses realloc() before all libraries have been properly initialised.
• minor bug fixes and changes.
• deprecate -F in favour of -j, to mimic make.

• non-locking stdio functions and dup()/dup2() reimplementations
• improved testsuite with RNG statistical analysis
• better fuzzing ratio range and random number generation

• network port selection for targeted fuzzing

• support for cherry-picking fuzzed descriptors
• support for CPU time limits
• do not interfere with LD_PRELOAD mechanisms

• remove spurious debug message
• fix an unsaved errno issue

• getopt_long() reimplementation
• fix for 64bit issues
• -M works on OpenBSD
• mmap reimplementation does not use the heap

• added new fuzzing modes
• fixed crash in regex handlers
• improved debug messages