agustingianni/dork.py

## dork.py
from git import *
import time
import sys
import re
import urllib2
from multiprocessing import Pool
import multiprocessing
import cgi

WEBKIT_REPO = "/home/agustin/workspace/WebKit"

HTML5_HEADER = """<!DOCTYPE html>
<html>
<head>
<style>
h3 {
    font-family: verdana, arial, helvetica, sans-serif;
    color: blue;
}

p {
    font-family: verdana, arial, helvetica, sans-serif;
}

.bugzilla {

}

.keywords {

}

.message {
    color: grey;
    font-family: monospace;
}


</style>
"""

HTML5_FOOTER = """</body>
</html>"""

repo = Repo(WEBKIT_REPO)

interesting_commiters = ["evans", "schuh", "inferno"]
interesting_commit_keywords = ["crash", "security", "use after free", "use-after-free", "uninitialized memory", "buffer overflow"]

def is_interesting_commiter(commiter):
    for person in interesting_commiters:
        if person in commiter:
            return True

    return False

def has_interesting_bug_commit(bug_url):
    return
    if "bugs.webkit.org" in bug_url:
        buffer = urllib2.urlopen(bug_url).read()
        if "You are not authorized to access bug" in buffer:
            return True

    return False

def has_interesting_keywords(commit_message):
    commit_message = commit_message.lower()

    matched_keywords = []

    for keyword in interesting_commit_keywords:
        if keyword in commit_message:
            matched_keywords.append(keyword)

    return matched_keywords

print HTML5_HEADER

for c in repo.iter_commits('master'):
    # We do not search for commits before 2010
    if time.gmtime(c.committed_date).tm_year < 2010:
        break

    is_interesting = False

    # Get all the urls on the commit message
    interesting_urls = filter(has_interesting_bug_commit, re.findall(r'https?://\S+', c.message))
    interesting_keywords = has_interesting_keywords(c.message)

    if is_interesting_commiter(str(c.committer)) or \
        len(interesting_keywords) or \
        len(interesting_urls) != 0:

        print "<h3>" + "Interesting commit by %s on %s" % (c.committer, time.asctime(time.gmtime(c.committed_date))) + "</h3>"

        if len(interesting_urls):
            print "<p class='bugzilla'>" + "Bugzilla blocked urls" + "</p>"
            for interesting_url in interesting_urls:
                print "<p>" + "Private bugzilla at %s" % (interesting_url) + "</p>"

        if len(interesting_keywords):
            print "<p class='keywords'>" + "Interesting keywords matched:" + "</p>"
            print "<p>" + "%s" % ",".join(interesting_keywords) + "</p>"

        print "<p>" + "Commit message:" + "</p>"

        try:
            print "<pre class='message'>"
            for line in cgi.escape(c.message).split("\n"):
                print line

            print "</pre>"
        except:
            print "<p>UNICODE Fail</p>"

print HTML5_FOOTER
	from git import *
	import time
	import sys
	import re
	import urllib2
	from multiprocessing import Pool
	import multiprocessing
	import cgi

	WEBKIT_REPO = "/home/agustin/workspace/WebKit"

	HTML5_HEADER = """<!DOCTYPE html>
	<html>
	<head>
	<style>
	h3 {
	font-family: verdana, arial, helvetica, sans-serif;
	color: blue;
	}

	p {
	font-family: verdana, arial, helvetica, sans-serif;
	}

	.bugzilla {

	}

	.keywords {

	}

	.message {
	color: grey;
	font-family: monospace;
	}


	</style>
	"""

	HTML5_FOOTER = """</body>
	</html>"""

	repo = Repo(WEBKIT_REPO)

	interesting_commiters = ["evans", "schuh", "inferno"]
	interesting_commit_keywords = ["crash", "security", "use after free", "use-after-free", "uninitialized memory", "buffer overflow"]

	def is_interesting_commiter(commiter):
	for person in interesting_commiters:
	if person in commiter:
	return True

	return False

	def has_interesting_bug_commit(bug_url):
	return
	if "bugs.webkit.org" in bug_url:
	buffer = urllib2.urlopen(bug_url).read()
	if "You are not authorized to access bug" in buffer:
	return True

	return False

	def has_interesting_keywords(commit_message):
	commit_message = commit_message.lower()

	matched_keywords = []

	for keyword in interesting_commit_keywords:
	if keyword in commit_message:
	matched_keywords.append(keyword)

	return matched_keywords

	print HTML5_HEADER

	for c in repo.iter_commits('master'):
	# We do not search for commits before 2010
	if time.gmtime(c.committed_date).tm_year < 2010:
	break

	is_interesting = False

	# Get all the urls on the commit message
	interesting_urls = filter(has_interesting_bug_commit, re.findall(r'https?://\S+', c.message))
	interesting_keywords = has_interesting_keywords(c.message)

	if is_interesting_commiter(str(c.committer)) or \
	len(interesting_keywords) or \
	len(interesting_urls) != 0:

	print "<h3>" + "Interesting commit by %s on %s" % (c.committer, time.asctime(time.gmtime(c.committed_date))) + "</h3>"

	if len(interesting_urls):
	print "<p class='bugzilla'>" + "Bugzilla blocked urls" + "</p>"
	for interesting_url in interesting_urls:
	print "<p>" + "Private bugzilla at %s" % (interesting_url) + "</p>"

	if len(interesting_keywords):
	print "<p class='keywords'>" + "Interesting keywords matched:" + "</p>"
	print "<p>" + "%s" % ",".join(interesting_keywords) + "</p>"

	print "<p>" + "Commit message:" + "</p>"

	try:
	print "<pre class='message'>"
	for line in cgi.escape(c.message).split("\n"):
	print line

	print "</pre>"
	except:
	print "<p>UNICODE Fail</p>"

	print HTML5_FOOTER