Fixed coloring conflicts in rbyd tree renderers

A bit of a hack, but rather than handling conditional alt branches, our
dbg rbyd tree renderers just represent single-pointer alts as an alt
with both branches pointing to the place.

Unfortunately, the two branches technically have different colors. This
resulted in a bit of contention when chosing how to color the tree.
Basically Python's dict ordering would determine which color won.

Which was a bit confusing when dbgrbyd.py displayed different tree
colorings for the same rbyd. dbgrbyd.py should be idempotent!

This is solved by adding another hack to check explicitly for
same-destination branches.

scripts/dbgbtree.py

@@ -548,6 +548,7 @@ class Rbyd:
                 d=t_depth-1 - alt['h'],
                 c=alt['c'],
             ))
+            if alt['ft'] != alt['nft']:
                 tree.add(TBranch(
                     a=alt['nft'],
                     b=alt['ft'],

scripts/dbglfs.py

@@ -579,6 +579,7 @@ class Rbyd:
                 d=t_depth-1 - alt['h'],
                 c=alt['c'],
             ))
+            if alt['ft'] != alt['nft']:
                 tree.add(TBranch(
                     a=alt['nft'],
                     b=alt['ft'],

scripts/dbgmtree.py

@@ -563,6 +563,7 @@ class Rbyd:
                 d=t_depth-1 - alt['h'],
                 c=alt['c'],
             ))
+            if alt['ft'] != alt['nft']:
                 tree.add(TBranch(
                     a=alt['nft'],
                     b=alt['ft'],

scripts/dbgrbyd.py

@@ -768,6 +768,7 @@ def dbg_tree(data, block_size, rev, trunk, weight, *,
                 d=t_depth-1 - alt['h'],
                 c=alt['c'],
             ))
+            if alt['ft'] != alt['nft']:
                 tree.add(TBranch(
                     a=alt['nft'],
                     b=alt['ft'],