Read jinja template with UTF-8 encoding set explicitly

Today we started seeing an UnicodeDecodeError error, coming from the
changed line: 'ascii' codec can't decode byte 0xc3 in position 2980:
ordinal not in range(128). This appeared after submitting the form.

The Path.read_text() function uses the locale's encoding by default.

AFAIK there have been some server configuration changes recently. To
my surprise, after ssh'ing into the server, I discovered that the
encoding returned by locale.getencoding() was ANSI_X3.4-1968. We then
changed the encoding to UTF-8 and bounced the server. When run by a
user (me), the locale.getencoding() now returns UTF-8. Yet, when
executed inside store_data.py, it still returns ANSI_X3.4-1968. See
commit f3c82f4.

This commit changes the from_string() call to explicitly use UTF-8. It
should prevent locale settings from interfering with loading the
template (which contains some umlauts).

Author: mslw

server/store_data.py

@@ -372,7 +372,9 @@ def create_result_page(commit_hash: str,
                        templates_directory: Path):
 
     jinja_template_path = templates_directory / "success.html.jinja2"
-    jinja_template = Environment(autoescape=select_autoescape()).from_string(jinja_template_path.read_text())
+    jinja_template = Environment(autoescape=select_autoescape()).from_string(
+        jinja_template_path.read_text(encoding="utf-8")
+    )
     return jinja_template.render(
         sub_project="Z03",
         reference=f"{time_stamp}-{commit_hash}",