diff options
Diffstat (limited to 'nikola/post.py')
| -rw-r--r-- | nikola/post.py | 73 |
1 files changed, 41 insertions, 32 deletions
diff --git a/nikola/post.py b/nikola/post.py index 3e3b608..350014a 100644 --- a/nikola/post.py +++ b/nikola/post.py @@ -26,7 +26,7 @@ from __future__ import unicode_literals, print_function, absolute_import -import codecs +import io from collections import defaultdict import datetime import os @@ -318,8 +318,12 @@ class Post(object): deps = [] if self.default_lang in self.translated_to: deps.append(self.base_path) + deps.append(self.source_path) if lang != self.default_lang: - deps += [get_translation_candidate(self.config, self.base_path, lang)] + cand_1 = get_translation_candidate(self.config, self.source_path, lang) + cand_2 = get_translation_candidate(self.config, self.base_path, lang) + if os.path.exists(cand_1): + deps.extend([cand_1, cand_2]) return deps def compile(self, lang): @@ -327,10 +331,10 @@ class Post(object): def wrap_encrypt(path, password): """Wrap a post with encryption.""" - with codecs.open(path, 'rb+', 'utf8') as inf: + with io.open(path, 'r+', encoding='utf8') as inf: data = inf.read() + "<!--tail-->" data = CRYPT.substitute(data=rc4(password, data)) - with codecs.open(path, 'wb+', 'utf8') as outf: + with io.open(path, 'w+', encoding='utf8') as outf: outf.write(data) dest = self.translated_base_path(lang) @@ -354,7 +358,7 @@ class Post(object): """ dep_path = self.base_path + '.dep' if os.path.isfile(dep_path): - with codecs.open(dep_path, 'rb+', 'utf8') as depf: + with io.open(dep_path, 'r+', encoding='utf8') as depf: return [l.strip() for l in depf.readlines()] return [] @@ -416,24 +420,14 @@ class Post(object): All links in the returned HTML will be relative. The HTML returned is a bare fragment, not a full document. """ - def strip_root_element(el): - ''' Strips root tag from an Element. - - Required because lxml has an tendency to add <div>, <body> - root tags to strings which are generated by using - lxml.html.tostring() - - :param Element el: the root element to strip - ''' - return (el.text or '') + ''.join( - [lxml.html.tostring(child, encoding='unicode') - for child in el.iterchildren()]) if lang is None: lang = nikola.utils.LocaleBorg().current_lang file_name = self._translated_file_path(lang) - with codecs.open(file_name, "r", "utf8") as post_file: + with io.open(file_name, "r", encoding="utf8") as post_file: data = post_file.read().strip() + if self.compiler.extension() == '.php': + return data try: document = lxml.html.fragment_fromstring(data, "body") except lxml.etree.ParserError as e: @@ -448,13 +442,10 @@ class Post(object): if self.hyphenate: hyphenate(document, lang) - data = lxml.html.tostring(document, encoding='unicode') - # data here is a full HTML doc, including HTML and BODY tags - # which is not ideal (Issue #464) try: - data = strip_root_element(document.body) - except IndexError: # No body there, it happens sometimes - pass + data = lxml.html.tostring(document.body, encoding='unicode') + except: + data = lxml.html.tostring(document, encoding='unicode') if teaser_only: teaser = TEASER_REGEXP.split(data)[0] @@ -477,7 +468,7 @@ class Post(object): # This closes all open tags and sanitizes the broken HTML document = lxml.html.fromstring(teaser) try: - data = strip_root_element(document) + data = lxml.html.tostring(document.body, encoding='unicode') except IndexError: data = lxml.html.tostring(document, encoding='unicode') @@ -494,7 +485,7 @@ class Post(object): try: document = lxml.html.fromstring(data) demote_headers(document, self.demote_headers) - data = strip_root_element(document) + data = lxml.html.tostring(document.body, encoding='unicode') except (lxml.etree.ParserError, IndexError): data = lxml.html.tostring(document, encoding='unicode') @@ -527,7 +518,7 @@ class Post(object): # duplicated with Post.text() lang = nikola.utils.LocaleBorg().current_lang file_name = self._translated_file_path(lang) - with codecs.open(file_name, "r", "utf8") as post_file: + with io.open(file_name, "r", encoding="utf8") as post_file: data = post_file.read().strip() try: document = lxml.html.fragment_fromstring(data, "body") @@ -561,9 +552,10 @@ class Post(object): def source_link(self, lang=None): """Return absolute link to the post's source.""" + ext = self.source_ext(True) return "/" + self.destination_path( lang=lang, - extension=self.source_ext(), + extension=ext, sep='/') def destination_path(self, lang=None, extension='.html', sep=os.sep): @@ -588,6 +580,10 @@ class Post(object): if lang is None: lang = nikola.utils.LocaleBorg().current_lang + # Let compilers override extension (e.g. the php compiler) + if self.compiler.extension() != '.html': + extension = self.compiler.extension() + pieces = self.translations[lang].split(os.sep) pieces += self.folder.split(os.sep) if self._has_pretty_url(lang): @@ -604,8 +600,21 @@ class Post(object): else: return link - def source_ext(self): - return os.path.splitext(self.source_path)[1] + def source_ext(self, prefix=False): + """ + Return the source file extension. + + If `prefix` is True, a `.src.` prefix will be added to the resulting extension + if it’s equal to the destination extension. + """ + + ext = os.path.splitext(self.source_path)[1] + # do not publish PHP sources + if prefix and ext == '.html': + # ext starts with a dot + return '.src' + ext + else: + return ext # Code that fetches metadata from different places @@ -655,7 +664,7 @@ def get_metadata_from_file(source_path, config=None, lang=None): source_path = get_translation_candidate(config, source_path, lang) elif lang: source_path += '.' + lang - with codecs.open(source_path, "r", "utf8") as meta_file: + with io.open(source_path, "r", encoding="utf8") as meta_file: meta_data = [x.strip() for x in meta_file.readlines()] return _get_metadata_from_file(meta_data) except (UnicodeDecodeError, UnicodeEncodeError): @@ -723,7 +732,7 @@ def get_metadata_from_meta_file(path, config=None, lang=None): elif lang: meta_path += '.' + lang if os.path.isfile(meta_path): - with codecs.open(meta_path, "r", "utf8") as meta_file: + with io.open(meta_path, "r", encoding="utf8") as meta_file: meta_data = meta_file.readlines() # Detect new-style metadata. |