summaryrefslogtreecommitdiff
path: root/tests/test_intl.py
blob: b5cd1c9663c99af811918a6ea6ba77179a0348ab (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
# -*- coding: utf-8 -*-
"""
    test_intl
    ~~~~~~~~~

    Test message patching for internationalization purposes.  Runs the text
    builder in the test root.

    :copyright: Copyright 2010 by the Sphinx team, see AUTHORS.
    :license: BSD, see LICENSE for details.
"""

from subprocess import Popen, PIPE
import re
import os
from StringIO import StringIO

from sphinx.util.pycompat import relpath

from util import *
from util import SkipTest


warnfile = StringIO()
root = test_roots / 'test-intl'
doctreedir = root / '_build' / 'doctree'


def with_intl_app(*args, **kw):
    default_kw = {
        'srcdir': root,
        'doctreedir': doctreedir,
        'confoverrides': {
            'language': 'xx', 'locale_dirs': ['.'],
            'gettext_compact': False,
        },
    }
    default_kw.update(kw)
    return with_app(*args, **default_kw)


def setup_module():
    # Delete remnants left over after failed build
    (root / 'xx').rmtree(True)
    (root / 'xx' / 'LC_MESSAGES').makedirs()
    # Compile all required catalogs into binary format (*.mo).
    for dirpath, dirs, files in os.walk(root):
        dirpath = path(dirpath)
        for f in [f for f in files if f.endswith('.po')]:
            po = dirpath / f
            mo = root / 'xx' / 'LC_MESSAGES' / (
                    relpath(po[:-3], root) + '.mo')
            if not mo.parent.exists():
                mo.parent.makedirs()
            try:
                p = Popen(['msgfmt', po, '-o', mo],
                    stdout=PIPE, stderr=PIPE)
            except OSError:
                raise SkipTest  # most likely msgfmt was not found
            else:
                stdout, stderr = p.communicate()
                if p.returncode != 0:
                    print stdout
                    print stderr
                    assert False, \
                        'msgfmt exited with return code %s' % p.returncode
                assert mo.isfile(), 'msgfmt failed'


def teardown_module():
    (root / '_build').rmtree(True)
    (root / 'xx').rmtree(True)


@with_intl_app(buildername='text')
def test_simple(app):
    app.builder.build(['bom'])
    result = (app.outdir / 'bom.txt').text(encoding='utf-8')
    expect = (u"\nDatei mit UTF-8"
              u"\n***************\n" # underline matches new translation
              u"\nThis file has umlauts: äöü.\n")
    assert result == expect


@with_intl_app(buildername='text')
def test_subdir(app):
    app.builder.build(['subdir/contents'])
    result = (app.outdir / 'subdir' / 'contents.txt').text(encoding='utf-8')
    assert result.startswith(u"\nsubdir contents\n***************\n")


@with_intl_app(buildername='html', cleanenv=True)
def test_i18n_footnote_break_refid(app):
    """test for #955 cant-build-html-with-footnotes-when-using"""
    app.builder.build(['footnote'])
    result = (app.outdir / 'footnote.html').text(encoding='utf-8')
    # expect no error by build


@with_intl_app(buildername='text', cleanenv=True)
def test_i18n_footnote_regression(app):
    """regression test for fix #955"""
    app.builder.build(['footnote'])
    result = (app.outdir / 'footnote.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH FOOTNOTE"
              u"\n******************\n"  # underline matches new translation
              u"\nI18N WITH FOOTNOTE INCLUDE THIS CONTENTS [ref] [1] [100]\n"
              u"\n[1] THIS IS A AUTO NUMBERED FOOTNOTE.\n"
              u"\n[ref] THIS IS A NAMED FOOTNOTE.\n"
              u"\n[100] THIS IS A NUMBERED FOOTNOTE.\n")
    assert result == expect


@with_intl_app(buildername='html', cleanenv=True)
def test_i18n_footnote_backlink(app):
    """i18n test for #1058"""
    app.builder.build(['footnote'])
    result = (app.outdir / 'footnote.html').text(encoding='utf-8')
    expects = [
        '<a class="footnote-reference" href="#id5" id="id1">[100]</a>',
        '<a class="footnote-reference" href="#id4" id="id2">[1]</a>',
        '<a class="reference internal" href="#ref" id="id3">[ref]</a>',
        '<a class="fn-backref" href="#id2">[1]</a>',
        '<a class="fn-backref" href="#id3">[ref]</a>',
        '<a class="fn-backref" href="#id1">[100]</a>',
        ]
    for expect in expects:
        matches = re.findall(re.escape(expect), result)
        assert len(matches) == 1


@with_intl_app(buildername='text', warning=warnfile, cleanenv=True)
def test_i18n_warn_for_number_of_references_inconsistency(app):
    app.builddir.rmtree(True)
    app.builder.build(['refs_inconsistency'])
    result = (app.outdir / 'refs_inconsistency.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH REFS INCONSISTENCY"
              u"\n****************************\n"
              u"\n* FOR FOOTNOTE [ref2].\n"
              u"\n* reference FOR reference.\n"
              u"\n* ORPHAN REFERENCE: I18N WITH REFS INCONSISTENCY.\n"
              u"\n[1] THIS IS A AUTO NUMBERED FOOTNOTE.\n"
              u"\n[ref2] THIS IS A NAMED FOOTNOTE.\n"
              u"\n[100] THIS IS A NUMBERED FOOTNOTE.\n")
    assert result == expect

    warnings = warnfile.getvalue().replace(os.sep, '/')
    warning_fmt = u'.*/refs_inconsistency.txt:\\d+: ' \
          u'WARNING: inconsistent %s in translated message\n'
    expected_warning_expr = (
        warning_fmt % 'footnote references' +
        warning_fmt % 'references' +
        warning_fmt % 'references')
    assert re.search(expected_warning_expr, warnings)


@with_intl_app(buildername='html', cleanenv=True)
def test_i18n_link_to_undefined_reference(app):
    app.builder.build(['refs_inconsistency'])
    result = (app.outdir / 'refs_inconsistency.html').text(encoding='utf-8')

    expected_expr = ('<a class="reference external" '
                     'href="http://www.example.com">reference</a>')
    assert len(re.findall(expected_expr, result)) == 2

    expected_expr = ('<a class="reference internal" '
                     'href="#reference">reference</a>')
    assert len(re.findall(expected_expr, result)) == 0

    expected_expr = ('<a class="reference internal" '
                     'href="#i18n-with-refs-inconsistency">I18N WITH '
                     'REFS INCONSISTENCY</a>')
    assert len(re.findall(expected_expr, result)) == 1


@with_intl_app(buildername='html', cleanenv=True)
def test_i18n_keep_external_links(app):
    """regression test for #1044"""
    app.builder.build(['external_links'])
    result = (app.outdir / 'external_links.html').text(encoding='utf-8')

    # external link check
    expect_line = (u'<li>EXTERNAL LINK TO <a class="reference external" '
                   u'href="http://python.org">Python</a>.</li>')
    matched = re.search('^<li>EXTERNAL LINK TO .*$', result, re.M)
    matched_line = ''
    if matched:
        matched_line = matched.group()
    assert expect_line == matched_line

    # internal link check
    expect_line = (u'<li><a class="reference internal" '
                   u'href="#i18n-with-external-links">EXTERNAL '
                   u'LINKS</a> IS INTERNAL LINK.</li>')
    matched = re.search('^<li><a .* IS INTERNAL LINK.</li>$', result, re.M)
    matched_line = ''
    if matched:
        matched_line = matched.group()
    assert expect_line == matched_line

    # inline link check
    expect_line = (u'<li>INLINE LINK BY <a class="reference external" '
                   u'href="http://sphinx-doc.org">SPHINX</a>.</li>')
    matched = re.search('^<li>INLINE LINK BY .*$', result, re.M)
    matched_line = ''
    if matched:
        matched_line = matched.group()
    assert expect_line == matched_line

    # unnamed link check
    expect_line = (u'<li>UNNAMED <a class="reference external" '
                   u'href="http://google.com">LINK</a>.</li>')
    matched = re.search('^<li>UNNAMED .*$', result, re.M)
    matched_line = ''
    if matched:
        matched_line = matched.group()
    assert expect_line == matched_line


@with_intl_app(buildername='text', warning=warnfile, cleanenv=True)
def test_i18n_literalblock_warning(app):
    app.builddir.rmtree(True)  #for warnings acceleration
    app.builder.build(['literalblock'])
    result = (app.outdir / 'literalblock.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH LITERAL BLOCK"
              u"\n***********************\n"
              u"\nCORRECT LITERAL BLOCK:\n"
              u"\n   this is"
              u"\n   literal block\n"
              u"\nMISSING LITERAL BLOCK:\n"
              u"\n<SYSTEM MESSAGE:")
    assert result.startswith(expect)

    warnings = warnfile.getvalue().replace(os.sep, '/')
    expected_warning_expr = u'.*/literalblock.txt:\\d+: ' \
            u'WARNING: Literal block expected; none found.'
    assert re.search(expected_warning_expr, warnings)


@with_intl_app(buildername='text')
def test_i18n_definition_terms(app):
    # regression test for #975
    app.builder.build(['definition_terms'])
    result = (app.outdir / 'definition_terms.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH DEFINITION TERMS"
              u"\n**************************\n"
              u"\nSOME TERM"
              u"\n   THE CORRESPONDING DEFINITION\n"
              u"\nSOME OTHER TERM"
              u"\n   THE CORRESPONDING DEFINITION #2\n")

    assert result == expect


@with_intl_app(buildername='text')
def test_seealso(app):
    app.builder.build(['seealso'])
    result = (app.outdir / 'seealso.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH SEEALSO"
              u"\n*****************\n"
              u"\nSee also: SHORT TEXT 1\n"
              u"\nSee also: LONG TEXT 1\n"
              u"\nSee also: SHORT TEXT 2\n"
              u"\n  LONG TEXT 2\n")
    assert result == expect


@with_intl_app(buildername='text')
def test_i18n_figure_caption(app):
    # regression test for #940
    app.builder.build(['figure_caption'])
    result = (app.outdir / 'figure_caption.txt').text(encoding='utf-8')
    expect = (u"\nI18N WITH FIGURE CAPTION"
              u"\n************************\n"
              u"\n   [image]MY CAPTION OF THE FIGURE\n"
              u"\n   MY DESCRIPTION PARAGRAPH1 OF THE FIGURE.\n"
              u"\n   MY DESCRIPTION PARAGRAPH2 OF THE FIGURE.\n")

    assert result == expect


@with_intl_app(buildername='html')
def test_i18n_index_entries(app):
    # regression test for #976
    app.builder.build(['index_entries'])
    result = (app.outdir / 'genindex.html').text(encoding='utf-8')

    def wrap(tag, keyword):
        start_tag = "<%s[^>]*>" % tag
        end_tag = "</%s>" % tag
        return r"%s\s*%s\s*%s" % (start_tag, keyword, end_tag)

    expected_exprs = [
        wrap('a', 'NEWSLETTER'),
        wrap('a', 'MAILING LIST'),
        wrap('a', 'RECIPIENTS LIST'),
        wrap('a', 'FIRST SECOND'),
        wrap('a', 'SECOND THIRD'),
        wrap('a', 'THIRD, FIRST'),
        wrap('dt', 'ENTRY'),
        wrap('dt', 'SEE'),
        wrap('a', 'MODULE'),
        wrap('a', 'KEYWORD'),
        wrap('a', 'OPERATOR'),
        wrap('a', 'OBJECT'),
        wrap('a', 'EXCEPTION'),
        wrap('a', 'STATEMENT'),
        wrap('a', 'BUILTIN'),
    ]
    for expr in expected_exprs:
        assert re.search(expr, result, re.M)