Re: [ECOS] GNU toolchains for eCos 3.0

[Andrew Lunn <andrew at lunn dot ch>]

> There isn't anything "magic" that ties eCos to any particular toolchain, etc.

I tend to both agree and disagree here.

We have seen some toolchain bugs when moving away from the version
released on the eCos website. Those tools might be old, but they are
well tested and known to work well.

On the other hand i always use the latest Debian unstable toolchain
for the synthetic target. This has found toolchain bugs which i've
reported back and had fixed. It has found eCos things which need work.

Using a new toolchain requires work for eCos. We need to fix all the
new warnings. Some are easy. Some are hard. The network stack is
particularly hard since it is old code and we want to minimise the
changes to allow a more up to date import sometime in the future.

I'm not the release manager, so i can only speculate from what has
happened in the past. What i think will happen is that eCosCentric
will donate a toolchain build for the common architectures. They will
become the recommended toolchain for the next X years. We will fix up
all the compiler warnings for these tools and do all the release
testing with them.

In my experience, the toolchains are pretty static and self contained,
and are not really a support issue. We get more problems with the GUI
configuration tool, which does have library version issues etc and it
is made worse by it being so hard to build, so nobody is really
interested in looking after it.

        Andrew

-- 
Before posting, please read the FAQ: http://ecos.sourceware.org/fom/ecos
and search the list archive: http://ecos.sourceware.org/ml/ecos-discuss