[New-bugs-announce] [issue6240] API to get source encoding as defined by PEP 263

[Sridhar Ratnakumar]

New submission from Sridhar Ratnakumar <sridharr at activestate.com>:

It'd be nice to get the encoding used by a specific Python file.
Considering that 'print' uses sys.stdout.encoding which is always set to
None when the Python process is run by subprocess, knowing the source
encoding is absolutely necessary in decoding the output generated by
that script.

eg: Run 'python setup.py --author' in the python-wifi-0.3.1 source
package as a subprocess.Popen(...) call.. and print the stdout.read()
string; you'll get encoding error.. unless you do
stdout.read().decode('latin1') .. where latin1 is specified as a coding:
line in setup.py.

The following function tries to detect the coding, but this guess work
not necessary when this is integrated with the standard library whose
implementation maps directly to that of PEP 263.

+def get_python_source_encoding(python_file):
+    """Detect the encoding used in the file ``python_file``
 
+    Detection is done as per http://www.python.org/dev/peps/pep-0263/
+    """
+    first_two_lines = open(python_file).readlines()[:2]
+    coding_line_regexp = ".*coding[:=]\s*([-\w.]+).*"
+
+    for line in first_two_lines:
+        m = re.match(coding_line_regexp, line)
+        if m:
+            return m.group(1)
+
+    # if no encoding is defined, use the default encoding
+    return 'ascii'

ref:
subprocess encoding mess: http://bugs.python.org/issue6135

----------
components: Interpreter Core, Library (Lib)
messages: 89097
nosy: lemburg, loewis, srid
severity: normal
status: open
title: API to get source encoding as defined by PEP 263
type: feature request
versions: Python 2.7, Python 3.1, Python 3.2

_______________________________________
Python tracker <report at bugs.python.org>
<http://bugs.python.org/issue6240>
_______________________________________