set default encoding of written files to UTF-8, added ability to pass

custom file enconding

4 files changed, 23 insertions, 3 deletions

diff --git a/CHANGES.txt b/CHANGES.txt
index 5a3651ad8..27be9f83d 100644
--- a/CHANGES.txt
+++ b/CHANGES.txt
@@ -134,6 +134,11 @@ RELEASE  VERSION/DATE TO BE FILLED IN LATER
       or not if it is already in the $LIBS construction var in the
       configure context. (issue #2768).
 
+  From Nickolai Korshunov
+    - Added explicit file encoding in env.Textfile. All written files by default 
+      will have utf-8 encoding. User can specify encoding manually with 
+      $FILE_ENCODING env var.
+
 
 RELEASE 4.4.0 -  Sat, 30 Jul 2022 14:08:29 -0700
 
diff --git a/SCons/Tool/textfile.py b/SCons/Tool/textfile.py
index 7fdc8b7be..c960e67cf 100644
--- a/SCons/Tool/textfile.py
+++ b/SCons/Tool/textfile.py
@@ -117,9 +117,14 @@ def _action(target, source, env):
                 value = str(value)
             subs.append((k, value))
 
+    if 'FILE_ENCODING' not in env:
+        file_encoding = 'utf-8'
+    else:
+        file_encoding = env['FILE_ENCODING']
+
     # write the file
     try:
-        target_file = open(target[0].get_path(), TEXTFILE_FILE_WRITE_MODE, newline='')
+        target_file = open(target[0].get_path(), TEXTFILE_FILE_WRITE_MODE, newline='', encoding=file_encoding)
     except (OSError, IOError) as e:
         raise SCons.Errors.UserError("Can't write target file %s [%s]" % (target[0],e))
 
diff --git a/SCons/Tool/textfile.xml b/SCons/Tool/textfile.xml
index f2e8bb89e..a40fef295 100644
--- a/SCons/Tool/textfile.xml
+++ b/SCons/Tool/textfile.xml
@@ -56,7 +56,7 @@ Nested lists of source strings
 are flattened.
 Source strings need not literally be Python strings:
 they can be Nodes or Python objects that convert cleanly
-to &f-link-Value; nodes
+to &f-link-Value; nodes.
 </para>
 
 <para>
@@ -64,6 +64,7 @@ The prefix and suffix specified by the &cv-link-TEXTFILEPREFIX;
 and &cv-link-TEXTFILESUFFIX; &consvars;
 (by default an empty string and <filename>.txt</filename>, respectively)
 are automatically added to the target if they are not already present.
+By default file encoding is "utf-8" and can be changed by &cv-link-FILE_ENCODING;
 Examples:
 </para>
 
@@ -259,4 +260,12 @@ The suffix used for &b-link-Textfile; file names;
 </summary>
 </cvar>
 
+<cvar name="FILE_ENCODING">
+<summary>
+<para>
+File encoding. "utf-8" by default.
+</para>
+</summary>
+</cvar>
+
 </sconsdoc>
diff --git a/test/textfile/textfile.py b/test/textfile/textfile.py
index a2d005cfb..f614dfc94 100644
--- a/test/textfile/textfile.py
+++ b/test/textfile/textfile.py
@@ -44,7 +44,8 @@ linesep = '\n'
 
 textparts = ['lalala', '42',
              'Goethe', 'Schiller',
-             'tanteratei']
+             'tanteratei',
+             '×']  # <-- this is unicode /xd7 symbol   
 foo1Text = linesep.join(textparts)
 foo2Text = '|*'.join(textparts)
 foo1aText = foo1Text + linesep