Merged in gauravjuvekar/scons (pull request #427)

string-escape unicode characters while printing --tree

2 files changed, 19 insertions, 3 deletions

diff --git a/src/CHANGES.txt b/src/CHANGES.txt
index cb6ab21..5127c51 100644
--- a/src/CHANGES.txt
+++ b/src/CHANGES.txt
@@ -13,6 +13,10 @@ RELEASE  VERSION/DATE TO BE FILLED IN LATER
 
 RELEASE  VERSION/DATE TO BE FILLED IN LATER
 
+  From Gaurav Juvekar:
+    - Fix issue #2910: Make --tree=all handle Unicode. (PR #427)
+    - Fix issue #2788: Fix typo in documentation example for sconf. (PR #388)
+
   From Manish Vachharajani:
     - Update debian rules, compat, and control to not use features
       deprecated or obsolete in later versions of debhelpers
diff --git a/src/engine/SCons/Util.py b/src/engine/SCons/Util.py
index a8a6990..2d8e75a 100644
--- a/src/engine/SCons/Util.py
+++ b/src/engine/SCons/Util.py
@@ -31,6 +31,7 @@ import sys
 import copy
 import re
 import types
+import codecs
 
 try:
     from UserDict import UserDict
@@ -148,7 +149,7 @@ class NodeList(UserList):
 #             else:
 #                 self.data = [ initlist,]
 
-    
+
     def __nonzero__(self):
         return len(self.data) != 0
 
@@ -170,10 +171,10 @@ class NodeList(UserList):
         return self.__class__(result)
 
     def __getitem__(self, index):
-        """ 
+        """
         This comes for free on py2,
         but py3 slices of NodeList are returning a list
-        breaking slicing nodelist and refering to 
+        breaking slicing nodelist and refering to
         properties and methods on contained object
         """
 #        return self.__class__(self.data[index])
@@ -289,6 +290,17 @@ def print_tree(root, child_func, prune=0, showtags=0, margin=[0], visited=None):
     """
 
     rname = str(root)
+    if sys.version_info.major < 3:
+        # Python 2 UTF-8 encoded str are str. escape_encode is a str to str
+        # encoding
+        rname = codecs.escape_encode(rname)[0]
+    else:
+        # Python 3 UTF-8 encoded str are bytes. escape_encode is a byte to byte
+        # encoding here.
+        rname = rname.encode('utf-8')
+        rname = codecs.escape_encode(rname)[0]
+        # Finally, we need a string again.
+        rname = rname.decode('ascii')
 
     # Initialize 'visited' dict, if required
     if visited is None: