working on it

05/musl-final/WHATSNEW

@@ -0,0 +1,40 @@
+0.5.0 - initial release
+
+0.5.9 - signal ABI bugfix, various cleanup and fixes:
+
+sigset_t was wrongly defined as 1024 bytes instead of 1024 bits,
+breaking the intended ABI compatibility with the LSB/glibc sigaction
+structure. users should upgrade immediately and rebuild any libraries
+or object files that might be using the incorrect definitions.
+
+improved security against DoS with tcb shadow passwords by checking
+that the file opened was really an ordinary file.
+
+fixed a bug in the implementation of atomic ops that could have
+allowed the compiler to incorrectly reorder them (in practice, gcc
+with the default settings on i386 was not reordering them).
+
+greatly improved conformance to the C and POSIX standards regarding
+what the standard header files make visible. _POSIX_C_SOURCE is now
+needed to get POSIX functions in standard C headers, and _XOPEN_SOURCE
+or _GNU_SOURCE are required to get XSI interfaces or GNU extensions,
+respectively.
+
+many internal improvements have been made to the syscall-related code
+in preparation for porting to x86_64 and other archs.
+
+0.6.0 - x86_64 port, various important bugs fixed
+
+new x86_64 (amd64) architecture port, contributed by Nicholas J. Kain,
+along with PORTING guide. source tree layout and build system have
+been improved to accommodate further ports.
+
+various bugs that were introduced while making the headers respect C
+and POSIX namespace standards have been fixed. conformance to the
+standards has been improved.
+
+fixed an inefficiency in qsort that triggered a bug (occasionaly
+internal compiler error) in some versions of gcc.
+
+fixed a major bug in the printf %n specifier that prevented it from
+working and caused memory corruption.