mirror of
https://gitlab.archlinux.org/archlinux/aurweb.git
synced 2025-02-03 10:43:03 +01:00
199f34e42e
Fixes a few edge cases: - URLs within code blocks used to get redundant <> added, breaking bash code snippets like `curl https://...` into `curl <https://...>`. - Links written with markdown's <https://...> syntax also used to get an extra pair of brackets. Signed-off-by: Lukas Fleischer <lfleischer@archlinux.org>
139 lines
4.4 KiB
Python
Executable file
139 lines
4.4 KiB
Python
Executable file
#!/usr/bin/env python3
|
|
|
|
import re
|
|
import pygit2
|
|
import sys
|
|
import bleach
|
|
import markdown
|
|
|
|
import aurweb.config
|
|
import aurweb.db
|
|
|
|
repo_path = aurweb.config.get('serve', 'repo-path')
|
|
commit_uri = aurweb.config.get('options', 'commit_uri')
|
|
|
|
|
|
class LinkifyExtension(markdown.extensions.Extension):
|
|
"""
|
|
Turn URLs into links, even without explicit markdown.
|
|
Do not linkify URLs in code blocks.
|
|
"""
|
|
|
|
# Captures http(s) and ftp URLs until the first non URL-ish character.
|
|
# Excludes trailing punctuation.
|
|
_urlre = (r'(\b(?:https?|ftp):\/\/[\w\/\#~:.?+=&%@!\-;,]+?'
|
|
r'(?=[.:?\-;,]*(?:[^\w\/\#~:.?+=&%@!\-;,]|$)))')
|
|
|
|
def extendMarkdown(self, md, md_globals):
|
|
processor = markdown.inlinepatterns.AutolinkInlineProcessor(self._urlre, md)
|
|
md.inlinePatterns.add('linkify', processor, '_end')
|
|
|
|
|
|
class FlysprayLinksPreprocessor(markdown.preprocessors.Preprocessor):
|
|
_fsre = re.compile(r'\b(FS#(\d+))\b')
|
|
_sub = r'[\1](https://bugs.archlinux.org/task/\2)'
|
|
|
|
def run(self, lines):
|
|
return [self._fsre.sub(self._sub, line) for line in lines]
|
|
|
|
|
|
class FlysprayLinksExtension(markdown.extensions.Extension):
|
|
def extendMarkdown(self, md, md_globals):
|
|
preprocessor = FlysprayLinksPreprocessor(md)
|
|
md.preprocessors.add('flyspray-links', preprocessor, '_end')
|
|
|
|
|
|
class GitCommitsInlineProcessor(markdown.inlinepatterns.InlineProcessor):
|
|
"""
|
|
Turn Git hashes like f7f5152be5ab into links to AUR's cgit.
|
|
|
|
Only commit references that do exist are linkified. Hashes are shortened to
|
|
shorter non-ambiguous prefixes. Only hashes with at least 7 digits are
|
|
considered.
|
|
"""
|
|
|
|
_repo = pygit2.Repository(repo_path)
|
|
|
|
def __init__(self, md, head):
|
|
self._head = head
|
|
super().__init__(r'\b([0-9a-f]{7,40})\b', md)
|
|
|
|
def handleMatch(self, m, data):
|
|
oid = m.group(1)
|
|
if oid not in self._repo:
|
|
# Unkwown OID; preserve the orginal text.
|
|
return None, None, None
|
|
|
|
prefixlen = 12
|
|
while prefixlen < 40:
|
|
if oid[:prefixlen] in self._repo:
|
|
break
|
|
prefixlen += 1
|
|
|
|
el = markdown.util.etree.Element('a')
|
|
el.set('href', commit_uri % (self._head, oid[:prefixlen]))
|
|
el.text = markdown.util.AtomicString(oid[:prefixlen])
|
|
return el, m.start(0), m.end(0)
|
|
|
|
|
|
class GitCommitsExtension(markdown.extensions.Extension):
|
|
_head = None
|
|
|
|
def __init__(self, head):
|
|
self._head = head
|
|
super(markdown.extensions.Extension, self).__init__()
|
|
|
|
def extendMarkdown(self, md, md_globals):
|
|
processor = GitCommitsInlineProcessor(md, self._head)
|
|
md.inlinePatterns.add('git-commits', processor, '_end')
|
|
|
|
|
|
class HeadingTreeprocessor(markdown.treeprocessors.Treeprocessor):
|
|
def run(self, doc):
|
|
for elem in doc:
|
|
if elem.tag == 'h1':
|
|
elem.tag = 'h5'
|
|
elif elem.tag in ['h2', 'h3', 'h4', 'h5']:
|
|
elem.tag = 'h6'
|
|
|
|
|
|
class HeadingExtension(markdown.extensions.Extension):
|
|
def extendMarkdown(self, md, md_globals):
|
|
md.treeprocessors.add('heading', HeadingTreeprocessor(md), '_end')
|
|
|
|
|
|
def get_comment(conn, commentid):
|
|
cur = conn.execute('SELECT PackageComments.Comments, PackageBases.Name '
|
|
'FROM PackageComments INNER JOIN PackageBases '
|
|
'ON PackageBases.ID = PackageComments.PackageBaseID '
|
|
'WHERE PackageComments.ID = ?', [commentid])
|
|
return cur.fetchone()
|
|
|
|
|
|
def save_rendered_comment(conn, commentid, html):
|
|
conn.execute('UPDATE PackageComments SET RenderedComment = ? WHERE ID = ?',
|
|
[html, commentid])
|
|
|
|
|
|
def main():
|
|
commentid = int(sys.argv[1])
|
|
|
|
conn = aurweb.db.Connection()
|
|
|
|
text, pkgbase = get_comment(conn, commentid)
|
|
html = markdown.markdown(text, extensions=['fenced_code',
|
|
LinkifyExtension(),
|
|
FlysprayLinksExtension(),
|
|
GitCommitsExtension(pkgbase),
|
|
HeadingExtension()])
|
|
allowed_tags = (bleach.sanitizer.ALLOWED_TAGS +
|
|
['p', 'pre', 'h4', 'h5', 'h6', 'br', 'hr'])
|
|
html = bleach.clean(html, tags=allowed_tags)
|
|
save_rendered_comment(conn, commentid, html)
|
|
|
|
conn.commit()
|
|
conn.close()
|
|
|
|
|
|
if __name__ == '__main__':
|
|
main()
|