text unit test should validate using installed mo file

We use custom gettext classes (e.g. GettextFactory &
NGettextFactory). We should exercise those classes with an installed
binary mo file to demonstrate we are actually returning the expected
translated strings for all strings defined as being translatable.

The test logic in install/po/test_i18n.py was recently enhanced to
make this type of testing easier and more complete.
tests/test_ipalib/test_text.py should import the new i18n test support
and run it.

Previously tests/test_ipalib/test_text.py made a feeble but incomplete
attempt to do the above but even that was often not run because the
test would skip because the necessary test files were not available
unless they had been manually created in the install/po subdir. It is
now possible to correct those deficiencies in the test.

This patch does the following:

* Moves the location of i18n test code and adjust references to it.
  install/po/test_i18n.py was moved to tests/i18n.py. This permits
  tests/test_ipalib/test_text.py to import the i18n test utilities
  in a clean fashion. The Makefile in install/po now calls this
  same file.

* Modfies test function in test_i18n.py to accept function pointers
  for retreiving a translation.

* Imports test_i18n.py from the install/po directory in the tree

* Creates a tmp directory for the test localedir

* Parses the current ipa.pot file in install/po and generates
  a test po and mo file with special unicode markers. It installs
  the test mo file in the tmp localedir. This is accomplished by
  calling create_po() from the test_i18n.py file.

* If any of the above does not work it raises nose.SkipTest with
  the reason, and skips the test.

* It sets up functions to get a translation and a plural translation
  via our text.GettextFactory class and text.NGettextFactory class
  respectively. This are the functions we use intenally to get
  translations. It set the localdir and lang which are used by those
  classes to match our test configuration. It then runs a validation
  test on every translation and it's plural found in the test.po file
  by calling po_file_iterate and passed it the function pointers to
  our internal routines.

* At the conclusion of the test it cleans up after itself.

Note: extraneous files are not created in the tree, only a tmp
directory is utilized.

Validating msgid's in C code was insufficient.

* Make the discovery of format conversions much more robust by authoring
  a new function parse_printf_fmt() that is able to discover each
  format conversion in a string and break it into it's individual
  subparts. One of those subparts is the argument selector index. In c
  code we need to know if the argumenet selector index is present to
  know if translator can reorder the substitution strings.

  This replaces the simplistic python_anonymous_substitutions_regexp
  which was insufficient to deal with other programming languages
  (e.g. c).

* Add get_prog_langs() function to return the set of programming
  languages a msgid appears in. This is necessar because the msdid
  validation is programming language specific.

https://fedorahosted.org/freeipa/ticket/2582

3 files changed, 228 insertions, 147 deletions

diff --git a/install/po/Makefile.in b/install/po/Makefile.in
index 4bee861a8..d65ba0c70 100644
--- a/install/po/Makefile.in
+++ b/install/po/Makefile.in
@@ -15,6 +15,7 @@ MSGINIT = @MSGINIT@
 MSGMERGE = @MSGMERGE@
 MSGCMP = @MSGCMP@
 TX = @TX@
+IPA_TEST_I18N = ../../tests/i18n.py
 
 DOMAIN = @GETTEXT_DOMAIN@
 MSGMERGE_UPDATE = $(MSGMERGE) --update
@@ -129,7 +130,7 @@ update-pot:
 	    echo "$(DOMAIN).pot unmodified" ; \
 	fi || :
 	@rm -f $(DOMAIN).pot.update $(DOMAIN).pot.update.tmp $(DOMAIN).pot.tmp
-	./test_i18n.py --show-strings --validate-pot $(DOMAIN).pot
+	$(IPA_TEST_I18N) --show-strings --validate-pot $(DOMAIN).pot
 
 msg-stats:
 	@pot_count=`$(MSGFMT) --statistics $(DOMAIN).pot 2>&1 | \
@@ -169,38 +170,14 @@ distclean: clean
 
 maintainer-clean: distclean
 
-# We test our translations by taking the original untranslated string
-# (e.g. msgid) and prepend a prefix character and then append a suffix
-# character. The test consists of asserting that the first character in the
-# translated string is the prefix, the last character in the translated string
-# is the suffix and the everything between the first and last character exactly
-# matches the original msgid.
-#
-# We use unicode characters not in the ascii character set for the prefix and
-# suffix to enhance the test. To make reading the translated string easier the
-# prefix is the unicode right pointing arrow and the suffix left pointing arrow,
-# thus the translated string looks like the original string enclosed in
-# arrows. In ASCII art the string "foo" would render as:
-# -->foo<--
-#
-# Unicode right pointing arrow: u'\u2192', utf-8 = '\xe2\x86\x92'
-# Unicode left pointing arrow:  u'\u2190', utf-8 = '\xe2\x86\x90'
-#
-# The sed command below performs the prefix and suffix substitution.
-#
-# When msginit is invoked with an English target locale it copies the msgid
-# into the msgstr. This is an undocumented feature of msginit. Otherwise the
-# msgstr will be set to the empty string (i.e. untranslated). We depend on
-# the msgid being copied to the msgstr.
-
 test:
-	./test_i18n.py --test-gettext
+	$(IPA_TEST_I18N) --test-gettext
 
 validate-pot:
-	./test_i18n.py --show-strings --validate-pot $(DOMAIN).pot
+	$(IPA_TEST_I18N) --show-strings --validate-pot $(DOMAIN).pot
 
 validate-po:
-	./test_i18n.py --show-strings --validate-po $(po_files)
+	$(IPA_TEST_I18N) --show-strings --validate-po $(po_files)
 
 debug:
 	@echo Python potfiles:
diff --git a/install/po/test_i18n.py b/tests/i18n.py
index beb43ccaa..067bc5e39 100755
--- a/install/po/test_i18n.py
+++ b/tests/i18n.py
@@ -19,6 +19,8 @@
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
 
+# WARNING: Do not import ipa modules, this is also used as a
+# stand-alone script (invoked from install/po Makefile).
 import optparse
 import sys
 import gettext
@@ -86,14 +88,108 @@ _shell_substitution_regexp = re.compile(r'\$(\s*)([({]?)(\s*)\w+(\s*)([)}]?)')
 # group 4: whitespace between variable and ending delimiter
 # group 5: ending delimiter
 
-# We do not permit anonymous substitutions in translation strings
-# (e.g. '%s occurred' % error) because they do not provide the
-# necessary context to translators, they would only see
-# '%s occurred'. Instead a keyword substitution should be used
-# (e.g. '%(error)s occurred' % {'error': error_message})
+printf_fmt_re = re.compile(
+    r"%"                                     # start
+     "(\d+\$)?"                              # fmt_arg    (group  1)
+     "(([#0 +'I]|-(?!\d))*)"                 # flags      (group  2)
+     "(([+-]?([1-9][0-9]*)?)|(\*|\*\d+\$))?" # width      (group  4)
+     "(\.((-?\d*)|(\*|)|(\*\d+\$)))?"        # precision  (group  8)
+     "(h|hh|l|ll|L|j|z|t)?"                  # length     (group 13)
+     "([diouxXeEfFgGaAcspnm%])")             # conversion (group 14)
 
-# Python anonymous format substitutions, e.g. %s, %d, %f, etc.
-python_anonymous_substitutions_regexp = re.compile(r'%[srduoxf]\b') # e.g. %s
+#-------------------------------------------------------------------------------
+
+def get_prog_langs(entry):
+    '''
+    Given an entry in a pot or po file return a set of the
+    programming languges it was found in. It needs to be a set
+    because the same msgid may appear in more than one file which may
+    be in different programming languages.
+
+    Note: One might think you could use the c-format etc. flags to
+    attached to entry to make this determination, but you can't. Those
+    flags refer to the style of the string not the programming
+    language it came from. Also the flags are often omitted and/or are
+    inaccurate.
+
+    For now we just look at the file extension. If we knew the path to
+    the file we could use other heuristics such as looking for the
+    shbang interpreter string.
+
+    The set of possible language types witch might be returned are:
+
+    * c
+    * python
+
+    '''
+    result = set()
+
+    for location in entry.occurrences:
+        filename = location[0]
+        ext = os.path.splitext(filename)[1]
+
+        if ext in ('.c', '.h', '.cxx', '.cpp', '.hxx'):
+            result.add('c')
+        elif ext in ('.py'):
+            result.add('python')
+
+    return result
+
+def parse_printf_fmt(s):
+    '''
+    Parse a printf style format string and return a list of format
+    conversions found in the string.
+
+    Each conversion specification is introduced by the character %, and
+    ends with a conversion specifier.  In between there may be (in this
+    order) zero or more flags, an optional minimum field width, an
+    optional precision and an optional length modifier. See "man 3
+    printf" for details.
+
+    Each item in the returned list is a dict whose keys are the
+    sub-parts of a conversion specification. The key and values are:
+
+    fmt
+        The entire format conversion specification
+    fmt_arg
+        The positional index of the matching argument in the argument
+        list, e.g. %1$ indicates the first argument in the argument
+        will be read for this conversion, excludes the leading % but
+        includes the trailing $, 1$ is the fmt_arg in %1$.
+    flags
+        The flag characaters, e.g. 0 is the flag in %08d
+    width
+        The width field, e.g. 20 is the width in %20s
+    precision
+        The precisioin field, e.g. .2 is the precision in %8.2f
+    length
+        The length modifier field, e.g. l is the length modifier in %ld
+    conversion
+        The conversion specifier character, e.g. d is the conversion
+        specification character in %ld
+
+    If the part is not found in the format it's value will be None.
+    '''
+
+    result = []
+
+    # get list of all matches, but skip escaped %
+    matches = [x for x in printf_fmt_re.finditer(s) if x.group(0) != "%%"]
+
+    # build dict of each sub-part of the format, append to result
+    for match in matches:
+        parts = {}
+        parts['fmt']        = match.group(0)
+        parts['fmt_arg']    = match.group(1)
+        parts['flags']      = match.group(2) or None
+        parts['width']      = match.group(4) or None
+        parts['precision']  = match.group(8)
+        parts['length']     = match.group(13)
+        parts['conversion'] = match.group(14)
+
+        result.append(parts)
+
+    return result
 
 def validate_substitutions_match(s1, s2, s1_name='string1', s2_name='string2'):
     '''
@@ -233,22 +329,35 @@ def validate_substitution_syntax(s, s_name='string'):
     return errors
 
 
-def validate_anonymous_substitutions(s, s_name='string'):
+def validate_positional_substitutions(s, prog_langs, s_name='string'):
     '''
-    We do not permit multiple anonymous substitutions in translation
+    We do not permit multiple positional substitutions in translation
     strings (e.g. '%s') because they do not allow translators to reorder the
     wording. Instead keyword substitutions should be used when there are
     more than one.
     '''
     errors = []
 
+    fmts = parse_printf_fmt(s)
+    n_fmts = len(fmts)
 
-    matches = list(python_anonymous_substitutions_regexp.finditer(s))
-
-    if len(matches) > 1:
-        for match in python_anonymous_substitutions_regexp.finditer(s):
-            errors.append("%s has anonymous substitution '%s', use keyword substitution instead" %
-                          (s_name, match.group(0)))
+    errors = []
+    if n_fmts > 1:
+        for i, fmt_parts in enumerate(fmts):
+            fmt        = fmt_parts['fmt']
+            fmt_arg    = fmt_parts['fmt_arg']
+            width      = fmt_parts['width']
+
+            if width == '*':
+                errors.append("Error: * width arg in format '%s should be indexed" % fmt)
+
+            if fmt_arg is None:
+                if 'c' in prog_langs:
+                    errors.append("%s format '%s' is positional, should use indexed argument" %
+                                  (s_name, fmt))
+                else:
+                    errors.append("%s format '%s' is positional, should use keyword substitution" %
+                                  (s_name, fmt))
 
     if errors:
         if show_strings:
@@ -265,7 +374,7 @@ def validate_file(file_path, validation_mode):
 
     * validate_substitutions_match()
     * validate_substitution_syntax()
-    * validate_anonymous_substitutions()
+    * validate_positional_substitutions()
 
     Returns the number of entries with errors.
     '''
@@ -290,7 +399,8 @@ def validate_file(file_path, validation_mode):
         have_msgstr = msgstr.strip() != ''
         if validation_mode == 'pot':
             if have_msgid:
-                errors = validate_anonymous_substitutions(msgid, 'msgid')
+                prog_langs = get_prog_langs(entry)
+                errors = validate_positional_substitutions(msgid, prog_langs, 'msgid')
                 entry_errors.extend(errors)
         if validation_mode == 'po':
             if have_msgid and have_msgstr:
@@ -387,23 +497,28 @@ def validate_unicode_edit(msgid, msgstr):
 
 
 def test_translations(po_file, lang, domain, locale_dir):
-    try:
+    # The test installs the test message catalog under the xh_ZA
+    # (e.g. Zambia Xhosa) language by default. It would be nice to
+    # use a dummy language not associated with any real language,
+    # but the setlocale function demands the locale be a valid
+    # known locale, Zambia Xhosa is a reasonable choice :)
 
-        # The test installs the test message catalog under the xh_ZA
-        # (e.g. Zambia Xhosa) language by default. It would be nice to
-        # use a dummy language not associated with any real language,
-        # but the setlocale function demands the locale be a valid
-        # known locale, Zambia Xhosa is a reasonable choice :)
+    os.environ['LANG'] = lang
 
-        os.environ['LANG'] = lang
+    # Create a gettext translation object specifying our domain as
+    # 'ipa' and the locale_dir as 'test_locale' (i.e. where to
+    # look for the message catalog). Then use that translation
+    # object to obtain the translation functions.
 
-        # Create a gettext translation object specifying our domain as
-        # 'ipa' and the locale_dir as 'test_locale' (i.e. where to
-        # look for the message catalog). Then use that translation
-        # object to obtain the translation functions.
+    t = gettext.translation(domain, locale_dir)
 
-        t = gettext.translation(domain, locale_dir)
+    get_msgstr = t.ugettext
+    get_msgstr_plural = t.ungettext
 
+    return po_file_iterate(po_file, get_msgstr, get_msgstr_plural)
+
+def po_file_iterate(po_file, get_msgstr, get_msgstr_plural):
+    try:
         # Iterate over the msgid's
         if not os.path.isfile(po_file):
             print >>sys.stderr, 'file does not exist "%s"' % (po_file)
@@ -422,8 +537,8 @@ def test_translations(po_file, lang, domain, locale_dir):
             if entry.msgid_plural:
                 msgid = entry.msgid
                 msgid_plural = entry.msgid_plural
-                msgstr = t.ungettext(msgid, msgid_plural, 1)
-                msgstr_plural = t.ungettext(msgid, msgid_plural, 2)
+                msgstr = get_msgstr_plural(msgid, msgid_plural, 1)
+                msgstr_plural = get_msgstr_plural(msgid, msgid_plural, 2)
 
                 try:
                     n_translations += 1
@@ -448,7 +563,7 @@ def test_translations(po_file, lang, domain, locale_dir):
 
             else:
                 msgid = entry.msgid
-                msgstr = t.ugettext(msgid)
+                msgstr = get_msgstr(msgid)
 
                 try:
                     n_translations += 1
diff --git a/tests/test_ipalib/test_text.py b/tests/test_ipalib/test_text.py
index 1931ca4fe..9f60785ff 100644
--- a/tests/test_ipalib/test_text.py
+++ b/tests/test_ipalib/test_text.py
@@ -22,10 +22,13 @@ Test the `ipalib.text` module.
 """
 
 import os
+import shutil
+import tempfile
 import re
 import nose
 import locale
 from tests.util import raises, assert_equal
+from tests.i18n import create_po, po_file_iterate
 from ipalib.request import context
 from ipalib import request
 from ipalib import text
@@ -35,93 +38,6 @@ singular = '%(count)d goose makes a %(dish)s'
 plural = '%(count)d geese make a %(dish)s'
 
 
-# Unicode right pointing arrow
-prefix = u'\u2192'               # utf-8 == '\xe2\x86\x92'
-# Unicode left pointing arrow
-suffix = u'\u2190'               # utf-8 == '\xe2\x86\x90'
-
-def get_msgid(po_file):
-    'Get the first non-empty msgid from the po file'
-
-    msgid_re = re.compile(r'^\s*msgid\s+"(.+)"\s*$')
-    f = open(po_file)
-    for line in f.readlines():
-        match = msgid_re.search(line)
-        if match:
-            msgid = match.group(1)
-            f.close()
-            return msgid
-    f.close()
-    raise ValueError('No msgid found in %s' % po_file)
-
-def test_gettext():
-    '''
-    Test gettext translation
-
-    We test our translations by taking the original untranslated
-    string (e.g. msgid) and prepend a prefix character and then append
-    a suffix character. The test consists of asserting that the first
-    character in the translated string is the prefix, the last
-    character in the translated string is the suffix and the
-    everything between the first and last character exactly matches
-    the original msgid.
-
-    We use unicode characters not in the ascii character set for the
-    prefix and suffix to enhance the test. To make reading the
-    translated string easier the prefix is the unicode right pointing
-    arrow and the suffix left pointing arrow, thus the translated
-    string looks like the original string enclosed in arrows. In ASCII
-    art the string "foo" would render as: "-->foo<--"
-    '''
-
-    localedir='install/po/test_locale'
-    test_file='install/po/test.po'
-
-    lang = os.environ['LANG']
-    os.environ['LANG'] = 'xh_ZA'
-
-    # Tell gettext that our domain is 'ipa', that locale_dir is
-    # 'test_locale' (i.e. where to look for the message catalog)
-    _ = text.GettextFactory('ipa', localedir)
-
-    # We need a translatable string to test with, read one from the
-    # test po file
-    if not file_exists(test_file):
-        raise nose.SkipTest(
-           'Test language not available, run "make test_lang" in install/po'
-        )
-    msgid = get_msgid(test_file)
-
-    # Get the localized instance of the msgid, it should be a Gettext
-    # instance.
-    localized = _(msgid)
-    assert(isinstance(localized, text.Gettext))
-
-    # Get the translated string from the Gettext instance by invoking
-    # unicode on it.
-    translated = unicode(localized)
-
-    # Perform the verifications on the translated string.
-
-    # Verify the first character is the test prefix
-    assert(translated[0] == prefix)
-
-    # Verify the last character is the test suffix
-    assert(translated[-1] == suffix)
-
-    # Verify everything between the first and last character is the
-    # original untranslated string
-    assert(translated[1:-1] == msgid)
-    
-    # Reset the language and assure we don't get the test values
-    context.__dict__.clear()
-    os.environ['LANG'] = lang
-
-    translated = unicode(localized)
-
-    assert(translated[0] != prefix)
-    assert(translated[-1] != suffix)
-
 def test_create_translation():
     f = text.create_translation
     key = ('foo', None)
@@ -129,6 +45,79 @@ def test_create_translation():
     assert context.__dict__[key] is t
 
 
+class test_TestLang(object):
+    def setUp(self):
+        self.tmp_dir = None
+        self.saved_lang  = None
+
+        self.lang = 'xh_ZA'
+        self.domain = 'ipa'
+
+        self.ipa_i18n_dir = os.path.join(os.path.dirname(__file__), '../../install/po')
+
+        self.pot_basename = '%s.pot' % self.domain
+        self.po_basename = '%s.po' % self.lang
+        self.mo_basename = '%s.mo' % self.domain
+
+        self.tmp_dir = tempfile.mkdtemp()
+        self.saved_lang  = os.environ['LANG']
+
+        self.locale_dir = os.path.join(self.tmp_dir, 'test_locale')
+        self.msg_dir = os.path.join(self.locale_dir, self.lang, 'LC_MESSAGES')
+
+        if not os.path.exists(self.msg_dir):
+            os.makedirs(self.msg_dir)
+
+        self.pot_file = os.path.join(self.ipa_i18n_dir, self.pot_basename)
+        self.mo_file = os.path.join(self.msg_dir, self.mo_basename)
+        self.po_file = os.path.join(self.tmp_dir, self.po_basename)
+
+        result = create_po(self.pot_file, self.po_file, self.mo_file)
+        if result:
+            raise nose.SkipTest('Unable to create po file "%s" & mo file "%s" from pot file "%s"' %
+                                (self.po_file, self.mo_file, self.pot_file))
+
+        if not file_exists(self.po_file):
+            raise nose.SkipTest('Test po file unavailable, run "make test" in install/po')
+
+        if not file_exists(self.mo_file):
+            raise nose.SkipTest('Test mo file unavailable, run "make test" in install/po')
+
+        self.po_file_iterate = po_file_iterate
+
+    def tearDown(self):
+        if self.saved_lang is not None:
+            os.environ['LANG'] = self.saved_lang
+
+        if self.tmp_dir is not None:
+            shutil.rmtree(self.tmp_dir)
+
+    def test_test_lang(self):
+        print "test_test_lang"
+        # The test installs the test message catalog under the xh_ZA
+        # (e.g. Zambia Xhosa) language by default. It would be nice to
+        # use a dummy language not associated with any real language,
+        # but the setlocale function demands the locale be a valid
+        # known locale, Zambia Xhosa is a reasonable choice :)
+
+        os.environ['LANG'] = self.lang
+
+        # Create a gettext translation object specifying our domain as
+        # 'ipa' and the locale_dir as 'test_locale' (i.e. where to
+        # look for the message catalog). Then use that translation
+        # object to obtain the translation functions.
+
+        def get_msgstr(msg):
+            gt = text.GettextFactory(localedir=self.locale_dir)(msg)
+            return unicode(gt)
+
+        def get_msgstr_plural(singular, plural, count):
+            ng = text.NGettextFactory(localedir=self.locale_dir)(singular, plural, count)
+            return ng(count)
+
+        result = self.po_file_iterate(self.po_file, get_msgstr, get_msgstr_plural)
+        assert result == 0
+
 class test_LazyText(object):
 
     klass = text.LazyText