Status notes after failed pbuilder-in-Docker experiments

1 files changed, 29 insertions, 0 deletions

diff --git a/README.md b/README.md
index 4f757f2..c84548b 100644
--- a/README.md
+++ b/README.md
@@ -65,3 +65,32 @@ is...kind of weird.  Partly that's because of the size of some of the
 files involved, but mostly it's because the Xvfb/ratpoison hack we're
 using to drive ISE installation doesn't work under `docker build`.
 Don't know why, don't really care (so many windmills, so little time).
+
+Current status
+--------------
+
+* Building the bitstream and the HSM firmware works.
+
+* Building the fimrware tarball doesn't work yet, because
+  `releng/alpha/Makefile` assumes it can use the releng PGP key.  This
+  will require a minor change to `releng/alpha/Makefile` and the
+  supporting scripts, deferred until we figure out what other changes
+  are needed for anyone other than the rleease engineer to use this.
+  
+* Building the host-side source packages (tarball, `.dsc`, `.rb`)
+  should work once firmware tarball problem solved.
+  
+* Running `pbuilder` within Docker does not look practical.  There are
+  several Docker-friendly alternatives, eg `whalebuilder`, but nested
+  Docker is a bad idea, so either we'd have to work out the socket
+  hack for automated Docker-next-to-Docker or we'd leave building
+  binary packages for the user.  Not clear that ordinary user really
+  wants us to spend the time building binary packages for every
+  supported Debian and Ubuntu release anyway.
+  
+So we have some restructuring to do in `releng/alpha/`, probably with
+some automated conditional tests to see whether we have the releng PGP
+key, etc, with things rearranges so to make it easier for an
+ordinary user to run the dockerized stuff and further conditionals and
+make variables to let the user specify which binary packages to build
+(if any).