[IMP] core: modernize tgz translations export

* don't use FS to do export
* update translations export dialog to fix outdated links (launchpad
  & 6.1 translations doc)

Note: can't use ``TarFile.gettarinfo`` to create the TarInfo object
because it wants an actual file object with a ``name`` attribute and
possibly even an fd.

closes #22798

odoo/addons/base/wizard/base_export_language_views.xml

@@ -24,13 +24,11 @@
                                 the rightmost column (value) contains the translations</li>
                             <li>PO(T) format: you should edit it with a PO editor such as
                                 <a href="http://www.poedit.net/" target="_blank">POEdit</a>, or your preferred text editor</li>
-                            <li>TGZ format: this is a compressed archive containing a PO file, directly suitable
-                                for uploading to Odoo's translation platform,
-                                <a href="https://translations.launchpad.net/openobject-addons" target="_blank">Launchpad</a></li>
+                            <li>TGZ format: bundles multiple PO(T) files as a single archive</li>
                             </ul>
                         </p>
                         <p>For more details about translating Odoo in your language, please refer to the
-                           <a href="http://doc.openerp.com/v6.1/contribute/07_improving_translations.html" target="_blank">documentation</a>.</p>
+                           <a href="https://github.com/odoo/odoo/wiki/Translations" target="_blank">documentation</a>.</p>
                     </div>
                     <footer states="choose">
                         <button name="act_getfile" string="Export" type="object" class="btn-primary"/>

odoo/tools/translate.py

@@ -687,23 +687,27 @@ def trans_export(lang, modules, buffer, format, cr):
                 writer.write(row['modules'], row['tnrs'], src, row['translation'], row['comments'])
 
         elif format == 'tgz':
-            rows_by_module = {}
+            rows_by_module = defaultdict(list)
             for row in rows:
                 module = row[0]
-                rows_by_module.setdefault(module, []).append(row)
-            tmpdir = tempfile.mkdtemp()
-            for mod, modrows in rows_by_module.items():
-                tmpmoddir = join(tmpdir, mod, 'i18n')
-                os.makedirs(tmpmoddir)
-                pofilename = (lang if lang else mod) + ".po" + ('t' if not lang else '')
-                buf = open(join(tmpmoddir, pofilename), 'wb')
-                _process('po', [mod], modrows, buf, lang)
-                buf.close()
+                rows_by_module[module].append(row)
 
-            tar = tarfile.open(fileobj=buffer, mode='w|gz')
-            tar.add(tmpdir, '')
-            tar.close()
+            with tarfile.open(fileobj=buffer, mode='w|gz') as tar:
+                for mod, modrows in rows_by_module.items():
+                    with io.BytesIO() as buf:
+                        _process('po', [mod], modrows, buf, lang)
+                        buf.seek(0)
 
+                        info = tarfile.TarInfo(
+                            join(mod, 'i18n', '{basename}.{ext}'.format(
+                                basename=lang or mod,
+                                ext='po' if lang else 'pot',
+                            )))
+                        # addfile will read <size> bytes from the buffer so
+                        # size *must* be set first
+                        info.size = len(buf.getvalue())
+
+                        tar.addfile(info, fileobj=buf)
         else:
             raise Exception(_('Unrecognized extension: must be one of '
                 '.csv, .po, or .tgz (received .%s).') % format)