Re: Drafting 1.0 announcements

[Rich Felker <dalias@aerifal.cx>]

On Mon, Mar 10, 2014 at 10:23:01AM +0000, Justin Cormack wrote:
> On 10 Mar 2014 06:23, "Rich Felker" <dalias@aerifal.cx> wrote:
> >
> > The below are DRAFTS, not actual announcements of a release. I'm
> > posting them now in search of suggestions for improving them.
> >
> > Rich
> 
> They look pretty good some comments inline.

Thanks!

> > Short release announcement for freecode and anyone already familiar
> > with musl just needing to know about the new release:
> >
> >     This release adds support for a soft-float ABI variant on MIPS as
> >     well as new experimental ports to SuperH and x32 (the new 32-bit
> >     ABI for x86_64). Two floating point printf bugs have been fixed
> >     including a rounding error and off-by-one buffer overflow that
> >     could occur only when printing certain denormal values with
> >     thousands of places of precision. A second overflow issue was
> >     fixed in wcsxfrm where a buffer length of zero was misinterpreted.
> >     Several other minor bug fixes and compatibility improvements have
> >     also been made.
> 
> Something about what 1.0 means for these users? Or even just the contains
> all mandatory POSIX stuff?

I was going to say it's not really news since we've had this level for
a while now, but actually the addition of the stupid legacy DES
setkey/encrypt functions also brought us up to the level of supporting
all mandatory XSI interfaces. So I think this is announcement-worthy.
Let's add after the first sentence:

    The legacy setkey/encrypt API has been added, completing musl's
    coverage of the POSIX XSI option interfaces.

> > Blurb for news sites that accept moderate-length submissions:
> >
> >     The musl libc project has released version 1.0, the result of
> 
> Add URL here?

Yes, of course a link to musl is needed, and everything else probably
needs to be linked too.

> >     three years of development and testing. Musl is a lightweight,
> >     fast, simple, MIT-licensed, correctness-oriented alternative to
> 
> Standards compliant?

Really this list is too long already... If it's important to add,
maybe we could make it "standards/correctness-oriented"?

> >     the GNU C library (glibc), uClibc, or Android's Bionic. At this
> 
> Mention klibc too? Might persuade them one day.

Probably most people don't even know about klibc. My thought was to
raise musl's ability to replace klibc on the initramfs mailing list
first rather than doing it in release announcements.

> >     point musl provides all mandatory C99 and POSIX interfaces (plus a
> >     lot of widely-used extensions), and well over 5000 packages are
> >     known to build successfully against musl.
> >
> >     Several options are available for trying musl. Compiler toolchains
> 
> Mention that you can just use it out of the box you don't need a new distro.

Yes, the compiler toolchains part was supposed to convey that but I
agree it's not clear. Could mention the gcc wrapper too.

Rich