Half-support for using file checksums instead of stamps.

A new redo-stamp program takes whatever you give it as stdin and uses it to
calculate a checksum for the current target.  If that checksum is the same
as last time, then we consider the target to be unchanged, and we set
checked_runid and stamp, but leave changed_runid alone.  That will make
future callers of redo-ifchange see this target as unmodified.

However, this is only "half" support because by the time we run the .do
script that calls redo-stamp, it's too late; the caller is a dependant of
the stamped program, which is already being rebuilt, even if redo-stamp
turns out to say that this target is unchanged.

The other half is coming up.

builder.py

@@ -1,6 +1,6 @@
 import sys, os, errno, stat
 import vars, jwack, state
-from helpers import log, log_, debug2, err, warn, unlink, close_on_exec
+from helpers import log, log_, debug, debug2, err, warn, unlink, close_on_exec
 
 
 def _possible_do_files(t):
@@ -187,9 +187,17 @@ class BuildJob:
                 unlink(self.tmpname1)
                 unlink(t)
             sf = self.sf
+            sf.refresh()
             sf.is_generated = True
-            sf.update_stamp()
-            sf.set_changed()
+            sf.is_override = False
+            if sf.is_checked() or sf.is_changed():
+                # it got checked during the run; someone ran redo-stamp.
+                # update_stamp would call set_changed(); we don't want that
+                sf.stamp = sf.read_stamp()
+            else:
+                sf.csum = None
+                sf.update_stamp()
+                sf.set_changed()
             sf.save()
         else:
             unlink(self.tmpname1)