Improve handling of remediations with non-ascii data

leapp/reporting/__init__.py

@@ -238,30 +238,61 @@ def path(self):
         return ('detail', 'remediations')
 
 
+def _guarantee_encoded_str(text, encoding='utf-8'):
+    """Guarantees that result is a utf-8 encoded string"""
+    # Apply conversion only in py2 case
+    if six.PY3:
+        return text
+    # Unless data is already encoded -> encode it
+    try:
+        return text.encode(encoding)
+    except (UnicodeDecodeError, AttributeError):
+        return text
+
+
+def _guarantee_decoded_str(text, encoding='utf-8'):
+    """Guarantees that result is 'native' text"""
+    # Apply conversion only in py2 case
+    if six.PY3:
+        return text
+    # Unless data is already decoded -> decode it
+    try:
+        return text.decode(encoding)
+    except (UnicodeEncodeError, AttributeError):
+        return text
+
+
 class RemediationCommand(BaseRemediation):
     def __init__(self, value=None):
         if not isinstance(value, list):
             raise TypeError('Value of "RemediationCommand" must be a list')
-        self._value = {'type': 'command', 'context': value}
+        self._value = {'type': 'command', 'context': [_guarantee_decoded_str(c) for c in value]}
 
     def __repr__(self):
-        return "[{}] {}".format(self._value['type'], ' '.join(self._value['context']))
+        # NOTE(ivasilev) As the message can contain non-ascii characters let's deal with it properly.
+        # As per python practices repr has to return an encoded string
+        return "[{}] {}".format(self._value['type'],
+                                ' '.join([_guarantee_encoded_str(c) for c in self._value['context']]))
 
 
 class RemediationHint(BaseRemediation):
     def __init__(self, value=None):
-        self._value = {'type': 'hint', 'context': value}
+        self._value = {'type': 'hint', 'context': _guarantee_decoded_str(value)}
 
     def __repr__(self):
-        return "[{}] {}".format(self._value['type'], self._value['context'])
+        # NOTE(ivasilev) As the message can contain non-ascii characters let's deal with it properly.
+        # As per python practices repr has to return an encoded string
+        return "[{}] {}".format(self._value['type'], _guarantee_encoded_str(self._value['context']))
 
 
 class RemediationPlaybook(BaseRemediation):
     def __init__(self, value=None):
-        self._value = {'type': 'playbook', 'context': value}
+        self._value = {'type': 'playbook', 'context': _guarantee_decoded_str(value)}
 
     def __repr__(self):
-        return "[{}] {}".format(self._value['type'], self._value['context'])
+        # NOTE(ivasilev) As the message can contain non-ascii characters let's deal with it properly.
+        # As per python practices repr has to return an encoded string
+        return "[{}] {}".format(self._value['type'], _guarantee_encoded_str(self._value['context']))
 
 
 class Remediation(object):

leapp/utils/report.py

@@ -1,7 +1,10 @@
 import hashlib
+import io
 import json
 import os
 
+import six
+
 from leapp.reporting import (
     _DEPRECATION_FLAGS,
     Groups,
@@ -110,21 +113,29 @@ def generate_report_file(messages_to_report, context, path, report_schema='1.1.0
     # NOTE(ivasilev) Int conversion should not break as only specific formats of report_schema versions are allowed
     report_schema_tuple = tuple(int(x) for x in report_schema.split('.'))
     if path.endswith(".txt"):
-        with open(path, 'w') as f:
+        with io.open(path, 'w', encoding='utf-8') as f:
             for message in sorted(messages_to_report, key=importance):
-                f.write('Risk Factor: {}{}\n'.format(message['severity'],
-                                                     ' (inhibitor)' if is_inhibitor(message) else ''))
-                f.write('Title: {}\n'.format(message['title']))
-                f.write('Summary: {}\n'.format(message['summary']))
+                f.write(u'Risk Factor: {}{}\n'.format(message['severity'],
+                                                      ' (inhibitor)' if is_inhibitor(message) else ''))
+                f.write(u'Title: {}\n'.format(message['title']))
+                f.write(u'Summary: {}\n'.format(message['summary']))
                 remediation = Remediation.from_dict(message.get('detail', {}))
                 if remediation:
-                    f.write('Remediation: {}\n'.format(remediation))
+                    # NOTE(ivasilev) Decoding is necessary in case of python2 as remediation is an encoded string,
+                    # while io.open expects "true text" input. For python3 repr will return proper py3 str, no
+                    # decoding will be needed.
+                    # This hassle and clumsiness makes me sad, so suggestions are welcome.
+                    remediation_text = 'Remediation: {}\n'.format(remediation)
+                    if isinstance(remediation_text, six.binary_type):
+                        # This will be true for py2 where repr returns an encoded string
+                        remediation_text = remediation_text.decode('utf-8')
+                    f.write(remediation_text)
                 if report_schema_tuple > (1, 0, 0):
                     # report-schema 1.0.0 doesn't have a stable report key
-                    f.write('Key: {}\n'.format(message['key']))
-                f.write('-' * 40 + '\n')
+                    f.write(u'Key: {}\n'.format(message['key']))
+                f.write(u'-' * 40 + '\n')
     elif path.endswith(".json"):
-        with open(path, 'w') as f:
+        with io.open(path, 'w', encoding='utf-8') as f:
             # Here all possible convertions will take place
             if report_schema_tuple < (1, 1, 0):
                 # report-schema 1.0.0 doesn't have a stable report key
@@ -150,4 +161,8 @@ def generate_report_file(messages_to_report, context, path, report_schema='1.1.0
                     # NOTE(ivasilev) As flags field is created only if there is an inhibitor
                     # a default value to pop has to be specified not to end up with a KeyError
                     msg.pop('flags', None)
-            json.dump({'entries': messages_to_report, 'leapp_run_id': context}, f, indent=2)
+            # NOTE(ivasilev) Has to be done in this way (dumps + py2 conversion + write to file instead of json.dump)
+            # because of a py2 bug in the json module that can produce a mix of unicode and str objects that will be
+            # incompatible with write. https://bugs.python.org/issue13769
+            data = json.dumps({'entries': messages_to_report, 'leapp_run_id': context}, indent=2, ensure_ascii=False)
+            f.write(data)

tests/scripts/test_reporting.py

@@ -1,3 +1,4 @@
+# -*- coding: utf-8 -*-
 from collections import namedtuple
 import datetime
 import json
@@ -13,7 +14,7 @@
     create_report_from_deprecation,
     create_report_from_error,
     _create_report_object,
-    Audience, Flags, Groups, Key, RelatedResource, Summary, Severity, Tags, Title
+    Audience, Flags, Groups, Key, RelatedResource, Remediation, Summary, Severity, Tags, Title
 )
 from leapp.utils.report import generate_report_file
 
@@ -102,6 +103,21 @@ def test_report_tags_and_flags():
     assert Flags(["This is a new flag", Groups.INHIBITOR]).value == ["This is a new flag", "inhibitor"]
 
 
+@pytest.mark.parametrize("report_suffix", ('.json', '.txt'))
+def test_remediation_with_non_ascii_value(report_suffix):
+    report_entries = [Title('Some report title'), Summary('Some summary not used for dynamical key generation'),
+                      Audience('sysadmin')]
+    # Partly related to leapp-repository PR1052
+    rem_command = ["ln", "-snf", "root/břeťa", "/bobošík"]
+    rem_hint = "Don't forget to check /bobošík directory!"
+    rem_playbook = "bobošík.yml"
+    rem = Remediation(commands=[rem_command], hint=rem_hint, playbook=rem_playbook)
+    report_entries.append(rem)
+    report_message = _create_report_object(report_entries).report
+    with tempfile.NamedTemporaryFile(suffix=report_suffix) as reportfile:
+        generate_report_file([report_message], 'leapp-run-id', reportfile.name, '1.1.0')
+
+
 def test_convert_from_error_to_report():
     error_dict_no_details = {
         'message': 'The system is not registered or subscribed.',