4.7 branch and 4.8 plans

[Joel Sherrill]

Hi,

The Steering Committee has had some
discussions about branching 4.7 and
plans for 4.8.  This is my an edited
version of my email which started the
discussion.

With my week on the edge of Lake Michigan and not being to
answer email, I thought a lot, drank some, and did some
geocaching. My conclusion was that we don't get enough
branches so.........

         !!! The time is now to branch 4.7 !!!

I will be gone again the last week of June (next week) so don't want to
branch before that.  This would make the branching some time
the 1st week of July.

I know there are a few nits in the CVS tree but overall it is
in good enough shape that I think forking a branch is OK.  Branches
get more stable as time goes on.  

4.7 tools would be gcc 4.1.x with Ralf's MIPS patch, binutils 2.17
(hopefully unreleased as of yet), and the current newlib and gdb.   The 
tools
can get updated after we branch

The biggest issues I know of that need to be addressed before branching
are:

  + that the pc386 BSP does not run netdemo on qemu.  Need help!!!

+ a significant SPARC BSP update is missing.  Need help!!!

+ I need to merge a POSIX semaphore timedwait patch.

While I am gone, please try to test, update the Wiki to include test 
reports and
let's get stable.

I have started a Wiki page for 4.8 ideas and hope we can all work 
together to
improve the RTEMS release cycle scheduling.  I for one, need to be 
whipped when
it looks like the code is in good enough shape to branch.  It needs to 
happen more
frequently.

http://www.rtems.com/wiki/index.php/RTEMS48Ideas

Comments, suggestions, etc. are always appreciated.

--joel