Revisiting 1.0 publicity plan

Revisiting 1.0 publicity plan

[Rich Felker]

Here's what I've dug up from the old publicity plan thread and more
recent developments, organized into a few specific areas we can work
on:


1. Reviewing status of musl support in:

- Buildroot (http://buildroot.uclibc.org/)
- Crosstool-ng (http://crosstool-ng.org/)
- EmbToolkit (https://www.embtoolkit.org/)
- OpenWRT (http://www.openwrt.org/)
- Gentoo (http://www.gentoo.org/)
- Alpine Linux (http://alpinelinux.org/)
- Aboriginal Linux (http://landley.net/aboriginal/)
- ...?

For the ones where support is complete or significant process is being
made, we should highlight their support of musl in the publicity work
we do elsewhere; doing so seems mutually beneficial. I think we can
also mail them announcing 1.0, thanking them for supporting musl, etc.
and offering to follow up on any issues they're still having.


2. Introducing musl to projects that could benefit from it but which
may not be aware:

- Cyanogenmod (http://www.cyanogenmod.org/)
- Code Sourcery (http://www.mentor.com/embedded-software/codesourcery)
- Linux-initramfs (http://vger.kernel.org/vger-lists.html#initramfs)
- DD-WRT (http://www.dd-wrt.com/)
- Raspberry Pi (http://www.raspberrypi.org/)
- ... ?

They would be candidates for sending our release announcement/press
release to, with an offer to help if they want to use musl. I'm
especially interested in promoting musl to Cyanogenmod.


3. Announcing to news sites and traditional media:

- LWN
- Linux Journal
- Linux Today
- Slashdot
- Ars Technica
- [Android news sites?]
- Paper magazines (PC Mag, Computer World, ...)
- ... ?

For these we probably need multiple versions of the release
announcement, ranging from more technial to talking points with
mainstream interest (bridging Android and Linux apps by providing a
full-feature libc that's light and license-compatible for use on
Android, etc.).


4. Calling on the musl community to blog/tweet/etc. about the 1.0
release when it happens.

This is a really obvious way to step up publicity without depending on
anyone else. Some ideas: what you like about musl, why you use it or
are involved with the project, success stories about deployment, etc.


5. Collecting patches for software that needs fixes to build on musl
and pushing them all upstream.

Re: Revisiting 1.0 publicity plan

[Timo Teras]

On Fri, 28 Feb 2014 06:24:54 -0500
Rich Felker <dalias@aerifal.cx> wrote:

> Here's what I've dug up from the old publicity plan thread and more
> recent developments, organized into a few specific areas we can work
> on:
> 
> 
> 1. Reviewing status of musl support in:
> 
> - Alpine Linux (http://alpinelinux.org/)
> 
> For the ones where support is complete or significant process is being
> made, we should highlight their support of musl in the publicity work
> we do elsewhere; doing so seems mutually beneficial. I think we can
> also mail them announcing 1.0, thanking them for supporting musl, etc.
> and offering to follow up on any issues they're still having.

As noted on IRC, we are debating whether or not to switch main builds
to musl on next release (spring) or the release after. In any case, the
builds are very usable already. Recent stats of our build is:

                main            testing
musl-x86:       1684/1729       822/1104
musl-x86_64:    1684/1727       835/1099
musl-armhf:     1639/1719       815/1092

legend being: #built-successfully / #relevant-packages

These include non-trivial stuff including gcc/fortran, gcc/java,
gcc/ada, openjdk6, openjdk7, qt, firefox, samba, wine (x86), and more.

Of course not everything is tested well - but I know our 'build master'
runs musl on his new x86_64 laptop. I also run the musl builds on my
Wandboard quad core arm build server, as well as on test Raspberry Pi.

I've done my x86/x86_64 testing in chroot only, but plan to setup first
x86_64 physical boxes with only alpine linux/musl in them.

> 2. Introducing musl to projects that could benefit from it but which
> may not be aware:
> 
> - Raspberry Pi (http://www.raspberrypi.org/)
> - ... ?
> 
> They would be candidates for sending our release announcement/press
> release to, with an offer to help if they want to use musl. I'm
> especially interested in promoting musl to Cyanogenmod.

Our armhf build is currently targetting Raspberry Pi - and I have it
running pretty nicely on it - though graphical side is still lacking Qt
and other stuff. omxplayer works, and vlc "sort of works". Though I
hope to get a few moments to put latest Qt running on it with hw accel.

I think this is known to some degree as someone posted even an early
review about it. Though significant improvements have been done since.

> 5. Collecting patches for software that needs fixes to build on musl
> and pushing them all upstream.

We've pushed some of the upstream; though we carry a lot of patches to
make stuff build also in our 'aports' git tree.

- Timo