musl 0.9.15 released

musl 0.9.15 released

[Rich Felker]

Hi everyone,

I'm pleased to annonce the (much-belated) release of musl 0.9.15. I
was originally hoping to have it ready a couple months ago, but
drawing out this release cycle did allow us to find and fix a number
of bugs and incompatibilities and build confidence in the quality of
what will become the 1.0 release. Here's the quick release blurb for
0.9.15:

    Major bug fixes include a buffer overflow in mbsrtowcs, various
    group file handling errors, failure of execle to pass on the new
    environment, and timezone-parsing crashes on 64-bit systems. Also
    fixed are several file descriptor leak (close-on-exec) issues,
    handling of invalid IP address strings, several fnmatch corner
    cases possibly leading to out-of-bound access, and failure of
    faccessat with the AT_EACCESS flag. This release also adds support
    for mixing IPv4 and v6 nameservers in resolv.conf, expanded shadow
    password API, IPv6 multicast structures, and the ability for
    libc.so to report the version installed.

    http://www.musl-libc.org/releases/musl-0.9.15.tar.gz

Due to the number of changes, the above blurb is highly condensed, so
see the in-tree release notes for further details:

    http://git.musl-libc.org/cgit/musl/tree/WHATSNEW?h=v0.9.15

I'm hoping we can move ahead towards 1.0 quickly now. I updated the
Roadmap on the wiki to reflect the remaining goals for 1.0:

Key targets:

- Polished documentation, and determining whether to include
  documentation with source or as a separate repository/download.
- Finish setting up automated regression test procedures for releases.
- Organized and coordinated publicity plan.
- Softfloat subarch for mips.

Secondary/optional targets:

- Integrating x32 (32-bit x86_64 ABI) port (contingent on limiting
  invasiveness).
- One or more other new ports labelled experimental (possibly aarch64
  and/or sh).
- Any other non-invasive but noteworthy features that can be added
  quickly.

Hardly any of these goals deal with my personal active work on the
codebase. My role between now and 1.0 should probably be focused on
reviewing and integrating any patches that we receive and writing
documentation, and of course the publicity plan.

After 1.0 I have a lot more exciting ideas for actual coding tasks;
these are mentioned on the wiki Roadmap.

Rich