Fix some possible encoding issues for non-utf8 systems.

anntzer · anntzer · commit 54c539a44512 · 2022-05-09T20:34:40.000+02:00
read_text/write_text default to the locale encoding, which may or may
not be utf8.  Fix these by making the encoding explicit or by using
bytes.
diff --git a/doc/sphinxext/redirect_from.py b/doc/sphinxext/redirect_from.py
@@ -15,6 +15,7 @@
 This creates in the build directory a file ``build/html/topic/old-page.html``
 that contains a relative refresh::
 
+    <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
     <html>
       <head>
         <meta http-equiv="refresh" content="0; url=new-page.html">
@@ -38,7 +39,9 @@
 logger = logging.getLogger(__name__)
 
 
-HTML_TEMPLATE = """<html>
+HTML_TEMPLATE = """\
+<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
+<html>
   <head>
     <meta http-equiv="refresh" content="0; url={v}">
   </head>
@@ -115,4 +118,4 @@ def _generate_redirects(app, exception):
         else:
             logger.info(f'making refresh html file: {k} redirect to {v}')
             p.parent.mkdir(parents=True, exist_ok=True)
-            p.write_text(html)
+            p.write_text(html, encoding='utf-8')
diff --git a/lib/matplotlib/testing/__init__.py b/lib/matplotlib/testing/__init__.py
@@ -108,7 +108,7 @@ def _check_for_pgf(texsystem):
             \typeout{pgfversion=\pgfversion}
             \makeatletter
             \@@end
-        """)
+        """, encoding="utf-8")
         try:
             subprocess.check_call(
                 [texsystem, "-halt-on-error", str(tex_path)], cwd=tmpdir,
diff --git a/lib/matplotlib/tests/test_animation.py b/lib/matplotlib/tests/test_animation.py
@@ -288,9 +288,8 @@ def test_failing_ffmpeg(tmpdir, monkeypatch, anim):
     with tmpdir.as_cwd():
         monkeypatch.setenv("PATH", ".:" + os.environ["PATH"])
         exe_path = Path(str(tmpdir), "ffmpeg")
-        exe_path.write_text("#!/bin/sh\n"
-                            "[[ $@ -eq 0 ]]\n")
-        os.chmod(str(exe_path), 0o755)
+        exe_path.write_bytes(b"#!/bin/sh\n[[ $@ -eq 0 ]]\n")
+        os.chmod(exe_path, 0o755)
         with pytest.raises(subprocess.CalledProcessError):
             anim.save("test.mpeg")
 
diff --git a/lib/matplotlib/tests/test_sphinxext.py b/lib/matplotlib/tests/test_sphinxext.py
@@ -75,9 +75,9 @@ def plot_directive_file(num):
     assert filecmp.cmp(range_6, plot_file(17))
 
     # Modify the included plot
-    contents = (source_dir / 'included_plot_21.rst').read_text()
-    contents = contents.replace('plt.plot(range(6))', 'plt.plot(range(4))')
-    (source_dir / 'included_plot_21.rst').write_text(contents)
+    contents = (source_dir / 'included_plot_21.rst').read_bytes()
+    contents = contents.replace(b'plt.plot(range(6))', b'plt.plot(range(4))')
+    (source_dir / 'included_plot_21.rst').write_bytes(contents)
     # Build the pages again and check that the modified file was updated
     modification_times = [plot_directive_file(i).stat().st_mtime
                           for i in (1, 2, 3, 5)]
diff --git a/setupext.py b/setupext.py
@@ -668,6 +668,7 @@ def do_custom_build(self, env):
             sln_path = base_path / vc / "freetype.sln"
             # https://developercommunity.visualstudio.com/comments/190992/view.html
             (sln_path.parent / "Directory.Build.props").write_text(
+                "<?xml version='1.0' encoding='utf-8'?>"
                 "<Project>"
                 "<PropertyGroup>"
                 # WindowsTargetPlatformVersion must be given on a single line.
@@ -676,8 +677,8 @@ def do_custom_build(self, env):
                 "::GetLatestSDKTargetPlatformVersion('Windows', '10.0')"
                 ")</WindowsTargetPlatformVersion>"
                 "</PropertyGroup>"
-                "</Project>"
-            )
+                "</Project>",
+                encoding="utf-8")
             # It is not a trivial task to determine PlatformToolset to plug it
             # into msbuild command, and Directory.Build.props will not override
             # the value in the project file.
diff --git a/tools/run_examples.py b/tools/run_examples.py
@@ -10,6 +10,7 @@
 import sys
 from tempfile import TemporaryDirectory
 import time
+import tokenize
 
 
 _preamble = """\
@@ -73,8 +74,9 @@ def main():
                 cwd.mkdir(parents=True)
             else:
                 cwd = stack.enter_context(TemporaryDirectory())
-            Path(cwd, relpath.name).write_text(
-                _preamble + (root / relpath).read_text())
+            with tokenize.open(root / relpath) as src:
+                Path(cwd, relpath.name).write_text(
+                    _preamble + src.read(), encoding="utf-8")
             for backend in args.backend or [None]:
                 env = {**os.environ}
                 if backend is not None:
