X-Git-Url: https://git.llucax.com/software/blitiri.git/blobdiff_plain/464df08f44fddca976f584e5daaf538adc4f443d..74a71431203c30c3b47ca985ec3d08e1ac38b080:/blitiri.cgi diff --git a/blitiri.cgi b/blitiri.cgi index e69789f..6ada91b 100755 --- a/blitiri.cgi +++ b/blitiri.cgi @@ -24,6 +24,12 @@ comments_path = "/tmp/blog/comments" # default templates. If they're not found, the built-in ones will be used. templates_path = "/tmp/blog/templates" +# Path where the cache is stored (must be writeable by the web server); +# set to None to disable. When enabled, you must take care of cleaning it up +# every once in a while. +#cache_path = "/tmp/blog/cache" +cache_path = None + # URL to the blog, including the name. Can be a full URL or just the path. blog_url = "/blog/blitiri.cgi" @@ -40,32 +46,10 @@ author = "Hartmut Kegan" # Article encoding encoding = "utf8" -# Captcha class -class Captcha (object): - def __init__(self, article): - self.article = article - words = article.title.split() - self.nword = hash(article.title) % len(words) % 5 - self.answer = words[self.nword] - self.help = 'gotcha, damn spam bot!' - - def get_puzzle(self): - nword = self.nword + 1 - if nword == 1: - n = '1st' - elif nword == 2: - n = '2nd' - elif nword == 3: - n = '3rd' - else: - n = str(nword) + 'th' - return "enter the %s word of the article's title" % n - puzzle = property(fget = get_puzzle) - - def validate(self, form_data): - if form_data.captcha.lower() == self.answer.lower(): - return True - return False +# Captcha method to use. At the moment only "title" is supported, but if you +# are keen with Python you can provide your own captcha implementation, see +# below for details. +captcha_method = "title" # @@ -101,6 +85,84 @@ except: data_path = os.path.realpath(data_path) templates_path = os.path.realpath(templates_path) + +# +# Captcha classes +# +# They must follow the interface described below. +# +# Constructor: +# Captcha(article) -> constructor, takes an article[1] as argument +# Attributes: +# puzzle -> a string with the puzzle the user must solve to prove he is +# not a bot (can be raw HTML) +# help -> a string with extra instructions, shown only when the user +# failed to solve the puzzle +# Methods: +# validate(form_data) -> based on the form data[2], returns True if +# the user has solved the puzzle uccessfully +# (False otherwise). +# +# Note you must ensure that the puzzle attribute and validate() method can +# "communicate" because they are executed in different requests. You can pass a +# cookie or just calculate the answer based on the article's data, for example. +# +# [1] article is an object with all the article's information: +# path -> string +# created -> datetime +# updated -> datetime +# uuid -> string (unique ID) +# title -> string +# author -> string +# tags -> list of strings +# raw_contents -> string in rst format +# comments -> list of Comment objects (not too relevant here) +# [2] form_data is an object with the form fields (all strings): +# author, author_error +# link, link_error +# catpcha, captcha_error +# body, body_error +# action, method + +class TitleCaptcha (object): + "Captcha that uses the article's title for the puzzle" + def __init__(self, article): + self.article = article + words = article.title.split() + self.nword = hash(article.title) % len(words) % 5 + self.answer = words[self.nword] + self.help = 'gotcha, damn spam bot!' + + @property + def puzzle(self): + nword = self.nword + 1 + if nword == 1: + n = '1st' + elif nword == 2: + n = '2nd' + elif nword == 3: + n = '3rd' + else: + n = str(nword) + 'th' + return "enter the %s word of the article's title" % n + + def validate(self, form_data): + if form_data.captcha.lower() == self.answer.lower(): + return True + return False + +known_captcha_methods = { + 'title': TitleCaptcha, +} + +# If the configured captcha method was a known string, replace it by the +# matching class; otherwise assume it's already a class and leave it +# alone. This way the user can either use one of our methods, or provide one +# of his/her own. +if captcha_method in known_captcha_methods: + captcha_method = known_captcha_methods[captcha_method] + + # Default template default_main_header = """\ @@ -404,26 +466,53 @@ div.section h1 { """ + +# Cache decorator +# It only works if the function is pure (that is, its return value depends +# only on its arguments), and if all the arguments are hash()eable. +def cached(f): + # do not decorate if the cache is disabled + if cache_path is None: + return f + + def decorate(*args, **kwargs): + hashes = '-'.join( str(hash(x)) for x in args + + tuple(kwargs.items()) ) + fname = 'blitiri.%s.%s.cache' % (f.__name__, hashes) + cache_file = os.path.join(cache_path, fname) + try: + s = open(cache_file).read() + except: + s = f(*args, **kwargs) + open(cache_file, 'w').write(s) + return s + + return decorate + + # helper functions -def rst_to_html(rst): +@cached +def rst_to_html(rst, secure = True): settings = { 'input_encoding': encoding, 'output_encoding': 'utf8', 'halt_level': 1, 'traceback': 1, + 'file_insertion_enabled': secure, + 'raw_enabled': secure, } parts = publish_parts(rst, settings_overrides = settings, writer_name = "html") return parts['body'].encode('utf8') -def validate_rst(rst): +def validate_rst(rst, secure = True): try: - rst_to_html(rst) + rst_to_html(rst, secure) return None except SystemMessage, e: desc = e.args[0].encode('utf-8') # the error string desc = desc[9:] # remove ":" - line = int(desc[:desc.find(':')]) # get just the line number + line = int(desc[:desc.find(':')] or 0) # get the line number desc = desc[desc.find(')')+2:-1] # remove (LEVEL/N) try: desc, context = desc.split('\n', 1) @@ -435,20 +524,28 @@ def validate_rst(rst): def valid_link(link): import re - mail_re = r"^[^ \t\n\r@<>()]+@[a-z0-9][a-z0-9\.\-_]*\.[a-z]+$" scheme_re = r'^[a-zA-Z]+:' + mail_re = r"^[^ \t\n\r@<>()]+@[a-z0-9][a-z0-9\.\-_]*\.[a-z]+$" url_re = r'^(?:[a-z0-9\-]+|[a-z0-9][a-z0-9\-\.\_]*\.[a-z]+)' \ r'(?::[0-9]+)?(?:/.*)?$' - scheme = '' - rest = link + if re.match(scheme_re, link, re.I): scheme, rest = link.split(':', 1) - if (not scheme or scheme == 'mailto') and re.match(mail_re, rest, re.I): + # if we have an scheme and a rest, assume the link is valid + # and return it as-is; otherwise (having just the scheme) is + # invalid + if rest: + return link + return None + + # at this point, we don't have a scheme; we will try to recognize some + # common addresses (mail and http at the moment) and complete them to + # form a valid link, if we fail we will just claim it's invalid + if re.match(mail_re, link, re.I): return 'mailto:' + link - if not scheme and re.match(url_re, rest, re.I): - return 'http://' + rest - if scheme: - return link + elif re.match(url_re, link, re.I): + return 'http://' + link + return None def sanitize(obj): @@ -594,25 +691,23 @@ class Comment (object): self._link = '' self._raw_content = 'Removed comment' - - def get_author(self): + @property + def author(self): if not self.loaded: self.load() return self._author - author = property(fget = get_author) - def get_link(self): + @property + def link(self): if not self.loaded: self.load() return self._link - link = property(fget = get_link) - def get_raw_content(self): + @property + def raw_content(self): if not self.loaded: self.load() return self._raw_content - raw_content = property(fget = get_raw_content) - def set(self, author, raw_content, link = '', created = None): self.loaded = True @@ -680,6 +775,11 @@ class Comment (object): class CommentDB (object): def __init__(self, article): self.path = os.path.join(comments_path, article.uuid) + # if comments were enabled after the article was added, we + # will need to create the directory + if not os.path.exists(self.path): + os.mkdir(self.path, 0755) + self.comments = [] self.load(article) @@ -736,37 +836,35 @@ class Article (object): self._raw_content = '' self._comments = [] - - def get_title(self): + @property + def title(self): if not self.loaded: self.load() return self._title - title = property(fget = get_title) - def get_author(self): + @property + def author(self): if not self.loaded: self.load() return self._author - author = property(fget = get_author) - def get_tags(self): + @property + def tags(self): if not self.loaded: self.load() return self._tags - tags = property(fget = get_tags) - def get_raw_content(self): + @property + def raw_content(self): if not self.loaded: self.load() return self._raw_content - raw_content = property(fget = get_raw_content) - def get_comments(self): + @property + def comments(self): if not self.loaded: self.load() return self._comments - comments = property(fget = get_comments) - def __cmp__(self, other): if self.path == other.path: @@ -957,7 +1055,7 @@ def render_comments(article, template, form_data): if not form_data: form_data = CommentFormData() form_data.action = blog_url + '/comment/' + article.uuid + '#comment' - captcha = Captcha(article) + captcha = captcha_method(article) print template.get_comment_form(article, form_data, captcha.puzzle) def render_html(articles, db, actyear = None, show_comments = False, @@ -983,7 +1081,7 @@ def render_artlist(articles, db, actyear = None): print template.get_main_header() print '

Articles

' for a in articles: - print '
  • %(title)s
  • ' \ + print '
  • %(title)s
  • ' \ % { 'url': blog_url, 'uuid': a.uuid, 'title': a.title, @@ -1054,6 +1152,7 @@ def handle_cgi(): atom = False style = False post = False + post_preview = False artlist = False comment = False @@ -1063,10 +1162,11 @@ def handle_cgi(): atom = path_info == '/atom' tag = path_info.startswith('/tag/') post = path_info.startswith('/post/') + post_preview = path_info.startswith('/preview/post/') artlist = path_info.startswith('/list') comment = path_info.startswith('/comment/') and enable_comments - if not style and not atom and not post and not tag \ - and not comment and not artlist: + if not style and not atom and not post and not post_preview \ + and not tag and not comment and not artlist: date = path_info.split('/')[1:] try: if len(date) > 1 and date[0]: @@ -1080,6 +1180,15 @@ def handle_cgi(): elif post: uuid = path_info.replace('/post/', '') uuid = uuid.replace('/', '') + elif post_preview: + art_path = path_info.replace('/preview/post/', '') + art_path = urllib.unquote_plus(art_path) + art_path = os.path.join(data_path, art_path) + art_path = os.path.realpath(art_path) + common = os.path.commonprefix([data_path, art_path]) + if common != data_path: # something nasty happened + post_preview = False + art_path = art_path[len(data_path)+1:] elif tag: t = path_info.replace('/tag/', '') t = t.replace('/', '') @@ -1103,15 +1212,20 @@ def handle_cgi(): render_style() elif post: render_html( [db.get_article(uuid)], db, year, enable_comments ) + elif post_preview: + article = Article(art_path, datetime.datetime.now(), + datetime.datetime.now()) + render_html( [article], db, year, enable_comments ) elif artlist: articles = db.get_articles() articles.sort(cmp = Article.title_cmp) render_artlist(articles, db) elif comment: form_data = CommentFormData(author.strip().replace('\n', ' '), - link.strip().replace('\n', ' '), captcha, body) + link.strip().replace('\n', ' '), captcha, + body.replace('\r', '')) article = db.get_article(uuid) - captcha = Captcha(article) + captcha = captcha_method(article) redirect = False valid = True if not form_data.author: @@ -1132,11 +1246,14 @@ def handle_cgi(): form_data.body_error = 'please, write a comment' valid = False else: - error = validate_rst(form_data.body) + error = validate_rst(form_data.body, secure=False) if error is not None: (line, desc, ctx) = error - form_data.body_error = 'error at line %d: %s' \ - % (line, desc) + at = '' + if line: + at = ' at line %d' % line + form_data.body_error = 'error%s: %s' \ + % (at, desc) valid = False if valid: c = article.add_comment(form_data.author, @@ -1231,7 +1348,10 @@ def handle_cmd(): if os.environ.has_key('GATEWAY_INTERFACE'): + i = datetime.datetime.now() handle_cgi() + f = datetime.datetime.now() + print '' % (f-i) else: sys.exit(handle_cmd())