Split TODO into pre-v1 and post-v1

Split things that need to be done before a v1 release and which can be
done after.

Signed-off-by: Sjoerd Simons <sjoerd.simons@collabora.co.uk>

TODO

@@ -1,18 +1,32 @@
+Before 1.0
+* Rename to debos
+
+* Stabelize the format (Can add actions and keys, but not remove keys/changes
+  behaviour)
+
 * Have all actions return error to handle in main rather then fatal
 ** In case any action fails, stop the full process rather then banging on
 
 * Use phase key in run action rather then postprocess: true
 
+* Make action naming consistent and add a stability guarantee
+
+* Split deploying root dir onto image from setup-image action
+
+* Add description field in baseaction and print it by default
+* Ensure quoted variables get passed to the inner instance properly
+
+* SPlit rootdir-deploy action out of setup-image
+
+Potentially After 1.0
+
 * Make logging more consistent
 * Have a mode to output the final yaml after templating and then exit (dry-run?)
 
-* Rename to debos
 
 * Documentation of all actions and overall concepts!
 * Tests, individual actions are mostly testable..
 
-* Make action naming consistent and add a stability guarantee
-
 * Support having devices as image files (fakemachine should handle it ok)
 
 * Create a standard docker image with debos to push to docker hub for easy
@@ -44,18 +58,16 @@
 * Rewrite pack/unpack in pure go and support more formats
 
 
-* Split deploying root dir onto image from setup-image action
-
-* Add description field in baseaction and print it by default
-
 * Make actions using (host) commands check their existance early
 
 
 * Ensure we copy xattrs?
 
-* Fix race in qemu-helper
+* Fix race in qemu-helper (if qemu-user-static gets installed in the system
+  chroot  things will get confused)
 
 * Do shell compatible parsing of script: argument to run actions and
   environment substitution
 
 * Support gpg signing ostree commits
+