1
0
Fork 0
forked from github/pelican

Merge pull request #2196 from mosra/absolute-url-merging

Make URL part joining aware of absolute URLs
This commit is contained in:
Justin Mayer 2018-02-08 08:19:26 -08:00 committed by GitHub
commit e7ac0a9272
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 131 additions and 64 deletions

View file

@ -11,7 +11,7 @@ import sys
import pytz
import six
from six.moves.urllib.parse import urlparse, urlunparse
from six.moves.urllib.parse import urljoin, urlparse, urlunparse
from pelican import signals
from pelican.settings import DEFAULT_CONFIG
@ -228,6 +228,87 @@ class Content(object):
key = key if self.in_default_lang else 'lang_%s' % key
return self._expand_settings(key)
def _link_replacer(self, siteurl, m):
what = m.group('what')
value = urlparse(m.group('value'))
path = value.path
origin = m.group('path')
# urllib.parse.urljoin() produces `a.html` for urljoin("..", "a.html")
# so if RELATIVE_URLS are enabled, we fall back to os.path.join() to
# properly get `../a.html`. However, os.path.join() produces
# `baz/http://foo/bar.html` for join("baz", "http://foo/bar.html")
# instead of correct "http://foo/bar.html", so one has to pick a side
# as there is no silver bullet.
if self.settings['RELATIVE_URLS']:
joiner = os.path.join
else:
joiner = urljoin
# However, it's not *that* simple: urljoin("blog", "index.html")
# produces just `index.html` instead of `blog/index.html` (unlike
# os.path.join()), so in order to get a correct answer one needs to
# append a trailing slash to siteurl in that case. This also makes
# the new behavior fully compatible with Pelican 3.7.1.
if not siteurl.endswith('/'):
siteurl += '/'
# XXX Put this in a different location.
if what in {'filename', 'attach'}:
if path.startswith('/'):
path = path[1:]
else:
# relative to the source path of this content
path = self.get_relative_source_path(
os.path.join(self.relative_dir, path)
)
if path not in self._context['filenames']:
unquoted_path = path.replace('%20', ' ')
if unquoted_path in self._context['filenames']:
path = unquoted_path
linked_content = self._context['filenames'].get(path)
if linked_content:
if what == 'attach':
if isinstance(linked_content, Static):
linked_content.attach_to(self)
else:
logger.warning(
"%s used {attach} link syntax on a "
"non-static file. Use {filename} instead.",
self.get_relative_source_path())
origin = joiner(siteurl, linked_content.url)
origin = origin.replace('\\', '/') # for Windows paths.
else:
logger.warning(
"Unable to find '%s', skipping url replacement.",
value.geturl(), extra={
'limit_msg': ("Other resources were not found "
"and their urls not replaced")})
elif what == 'category':
origin = joiner(siteurl, Category(path, self.settings).url)
elif what == 'tag':
origin = joiner(siteurl, Tag(path, self.settings).url)
elif what == 'index':
origin = joiner(siteurl, self.settings['INDEX_SAVE_AS'])
elif what == 'author':
origin = joiner(siteurl, Author(path, self.settings).url)
else:
logger.warning(
"Replacement Indicator '%s' not recognized, "
"skipping replacement",
what)
# keep all other parts, such as query, fragment, etc.
parts = list(value)
parts[2] = origin
origin = urlunparse(parts)
return ''.join((m.group('markup'), m.group('quote'), origin,
m.group('quote')))
def _update_content(self, content, siteurl):
"""Update the content attribute.
@ -251,69 +332,7 @@ class Content(object):
\2""".format(instrasite_link_regex)
hrefs = re.compile(regex, re.X)
def replacer(m):
what = m.group('what')
value = urlparse(m.group('value'))
path = value.path
origin = m.group('path')
# XXX Put this in a different location.
if what in {'filename', 'attach'}:
if path.startswith('/'):
path = path[1:]
else:
# relative to the source path of this content
path = self.get_relative_source_path(
os.path.join(self.relative_dir, path)
)
if path not in self._context['filenames']:
unquoted_path = path.replace('%20', ' ')
if unquoted_path in self._context['filenames']:
path = unquoted_path
linked_content = self._context['filenames'].get(path)
if linked_content:
if what == 'attach':
if isinstance(linked_content, Static):
linked_content.attach_to(self)
else:
logger.warning(
"%s used {attach} link syntax on a "
"non-static file. Use {filename} instead.",
self.get_relative_source_path())
origin = '/'.join((siteurl, linked_content.url))
origin = origin.replace('\\', '/') # for Windows paths.
else:
logger.warning(
"Unable to find '%s', skipping url replacement.",
value.geturl(), extra={
'limit_msg': ("Other resources were not found "
"and their urls not replaced")})
elif what == 'category':
origin = '/'.join((siteurl, Category(path, self.settings).url))
elif what == 'tag':
origin = '/'.join((siteurl, Tag(path, self.settings).url))
elif what == 'index':
origin = '/'.join((siteurl, self.settings['INDEX_SAVE_AS']))
elif what == 'author':
origin = '/'.join((siteurl, Author(path, self.settings).url))
else:
logger.warning(
"Replacement Indicator '%s' not recognized, "
"skipping replacement",
what)
# keep all other parts, such as query, fragment, etc.
parts = list(value)
parts[2] = origin
origin = urlunparse(parts)
return ''.join((m.group('markup'), m.group('quote'), origin,
m.group('quote')))
return hrefs.sub(replacer, content)
return hrefs.sub(lambda m: self._link_replacer(siteurl, m), content)
def get_siteurl(self):
return self._context.get('localsiteurl', '')

View file

@ -397,6 +397,54 @@ class TestPage(LoggedTestCase):
'</blockquote>'
)
def test_intrasite_link_absolute(self):
"""Test that absolute URLs are merged properly."""
args = self.page_kwargs.copy()
args['settings'] = get_settings(
STATIC_URL='http://static.cool.site/{path}',
ARTICLE_URL='http://blog.cool.site/{slug}.html')
args['source_path'] = 'content'
args['context']['filenames'] = {
'images/poster.jpg': Static('',
settings=args['settings'],
source_path='images/poster.jpg'),
'article.rst': Article('',
settings=args['settings'],
metadata={'slug': 'article',
'title': 'Article'})
}
# Article link will go to blog
args['content'] = (
'<a href="{filename}article.rst">Article</a>'
)
content = Page(**args).get_content('http://cool.site')
self.assertEqual(
content,
'<a href="http://blog.cool.site/article.html">Article</a>'
)
# Page link will go to the main site
args['content'] = (
'<a href="{index}">Index</a>'
)
content = Page(**args).get_content('http://cool.site')
self.assertEqual(
content,
'<a href="http://cool.site/index.html">Index</a>'
)
# Image link will go to static
args['content'] = (
'<img src="{filename}/images/poster.jpg"/>'
)
content = Page(**args).get_content('http://cool.site')
self.assertEqual(
content,
'<img src="http://static.cool.site/images/poster.jpg"/>'
)
def test_intrasite_link_markdown_spaces(self):
# Markdown introduces %20 instead of spaces, this tests that
# we support markdown doing this.