As we enter the OE-Core universe, I think it is a good time to think
about where the Angstrom Distribution needs to focus its efforts.

Currently, the most user visible feature of Angstrom is the availability
of binary feeds so end users can add packages via the internet easily.

Some ideas:

1) Continue providing package feeds.
2) Improve web services (Narcissus)

One of the issues I have with Angstrom is the config files are hard to
follow, there are files pinning package and toolchain versions. Most of
these exist to control the flow of data from upstream OE, although that
may not be as large a problem with the new pull model for OE-core and
meta-oe.

The other issue is toolchain management. Angstrom uses various gcc
versions for different archs. The idea is we can select toolchains that
work well for each arch, but my concern is this is hard to evaluate
across the range of architectures supported. I'd like to see the
toolchian selection clearly documented somewhere, including detailed
reasons so we can pass this info upstream and try and get the current
toolchains usable by all archs.

Finally, there is the issue of older hardware that is poorly supported.
Narcissus has selections for older hardware, but it does not produce
working images. The problem here is that most people using it post they
have a problem and we have no way to find and fix the problem. Working
and non-working, or supported and un-supported machines should be
clearly documented to avoid false expectations by users.

Philip

_______________________________________________
Angstrom-distro-devel mailing list
Angstrom-distro-devel@linuxtogo.org
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/angstrom-distro-devel

Reply via email to