Issue 104 - added checktranslations function

localeTools.py

Index: localeTools.py
diff --git a/localeTools.py b/localeTools.py
index b6554e0aa52461a7347c843a3b251d035ec34652..43d3642b6054886c345ce0721612c89b3903fd35 100644
--- a/localeTools.py
+++ b/localeTools.py
@@ -300,3 +300,36 @@ def getTranslations(localeConfig, projectName, key):
             valid_extension = file.endswith('.json')
             if os.path.isfile(path) and valid_extension and not file in files:
                 os.remove(path)
+
+
+def check_translations(dir):

[tlucas, 2017/10/25 15:09:23]

The parameter is not used anywhere in this function.

[erick, 2017/10/29 03:45:56]

Acknowledged.

[erick, 2017/10/31 04:30:00]

Done.

[erick, 2017/10/31 04:30:01]

Done.

+    for locale in os.listdir('_locales/'):

[tlucas, 2017/10/25 15:09:23]

"_locales" is only valid for adblockpluschrome.

in adblockplusui it's "locale".

The proper solution would be to parse the target's metadata and determine the
locale-path by reading "base_path" from the "locales" section.

Hint: this is already done in "readLocaleConfig" in build.py
Additional Hint: "readLocaleConfig" also provides you with a list of available
locales.

Please use it.

[erick, 2017/10/29 03:45:55]

Acknowledged.

[erick, 2017/10/31 04:29:59]

Done.

+        jsonpath = os.path.join('_locales/', locale, 'messages.json')
+
+        if not os.path.exists(jsonpath):
+            continue
+
+        with codecs.open(jsonpath, 'rb', encoding='utf-8') as f:
+            try:
+                data = json.load(f)
+            except ValueError, e:
+                print >>sys.stderr, '%s -> File is not valid. %s' % e

[tlucas, 2017/10/25 15:09:23]

nit: https://adblockplus.org/en/coding-style (section "Python"):

"Use the + operator when concatenating exactly two strings, use the format()
method for more complex string formatting, use the join() method when
concatenating pre-existing sequences."

[erick, 2017/10/29 03:45:55]

Acknowledged.

[tlucas, 2017/10/29 23:10:03]

This is NOT done.

[erick, 2017/10/31 04:30:01]

Done.

+
+        for key, value in data.iteritems():
+            variables = set()
+            max_length = value.get('maxLength')
+            msg_length = len(value['message'])
+            if (max_length is not None

[tlucas, 2017/10/25 15:09:23]

nit: I still suggest using a shorter variable name for "msg_length", so that the
if conditions can fit in one line.

[erick, 2017/10/29 03:45:56]

Acknowledged.

[erick, 2017/10/31 04:30:00]

Done.

[erick, 2017/10/31 04:30:01]

Done.

+                and msg_length > max_length

[tlucas, 2017/10/25 15:09:23]

Tox fails on this line:

./localeTools.py:323:17: W503 line break before binary operator

[erick, 2017/10/29 03:45:55]

Acknowledged.

[erick, 2017/10/31 04:29:59]

Done.

[erick, 2017/10/31 04:29:59]

Done.

+                or msg_length > 160):

[tlucas, 2017/10/25 15:09:23]

Tox fails on this line:

./localeTools.py:324:17: W503 line break before binary operator
./localeTools.py:324:17: E129 visually indented line with same indent as next
logical line

[erick, 2017/10/29 03:45:55]

Acknowledged.

[erick, 2017/10/31 04:30:00]

Done.

[erick, 2017/10/31 04:30:00]

Done.

+                print >>sys.stderr, '%s %s ->' % (locale, key), \

[tlucas, 2017/10/25 15:09:23]

https://adblockplus.org/en/coding-style (section "Python"):

"Use the + operator when concatenating exactly two strings, use the format()
method for more complex string formatting, use the join() method when
concatenating pre-existing sequences."

-> don't use "foo%s" % ('bar')

Additionally, don't use "\" to enable line continuation. Make use of
parentheses, that way python will simply concatenate multi-line strings itself:

                print >>sys.stderr, (
                    '{} {} -> translation might be too long. More than 160 '
                    'chars or maxLength exceeded'
                ).format(locale, key)


for both: same below

[erick, 2017/10/29 03:45:55]

Acknowledged.

[erick, 2017/10/31 04:30:00]

Done.

[erick, 2017/10/31 04:30:01]

Done.

+                    'translation might be too long. More then', \
+                    '160 chars or maxLength exceeded'
+
+            for match in re.finditer(r'\$(\S+?)\$', value['message']):
+                variables.add(match.group(1))
+            expected = set(value.get('placeholders', {}).iterkeys())

[tlucas, 2017/10/25 15:09:23]

The above 3 lines seem rather complicated. Maybe something like this would be
more readable:

if not all([placeholder in value['messages'] for placeholder in expected]):
    ...

Addtionally, you are relying on the key/value pairs "placeholder" to be present
in every messages.json - are we sure that this is given? If not, i suggest
reading the placeholder from the default-language messages.json.

[erick, 2017/10/29 03:45:55]

Is a possible way to check for an empty set() then go through the
default-language json?

[tlucas, 2017/10/29 23:10:03]

Why would you want to potentially double cpu load?
Cycling through the placeholders expected in the default-language is perfectly
sufficient.

Technically: yes, but i strongly discourage your approach (if i got it right)

+            if variables != expected:
+                print >>sys.stderr, '%s %s: Variables' % (locale, key), \
+                    'used are %s, expected.' % variables, \
+                    'Variables: %s' % expected