From f56f4724c36a5261d53c8c78f30d62c12f85fe0e Mon Sep 17 00:00:00 2001 From: Alexandre Aubin Date: Sun, 22 Mar 2020 01:23:55 +0100 Subject: [PATCH] Attempt to anonymize data pasted to paste.yunohost.org (in particular domain names) --- src/yunohost/utils/yunopaste.py | 45 +++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) diff --git a/src/yunohost/utils/yunopaste.py b/src/yunohost/utils/yunopaste.py index 89c62d761..530295735 100644 --- a/src/yunohost/utils/yunopaste.py +++ b/src/yunohost/utils/yunopaste.py @@ -2,14 +2,23 @@ import requests import json +import logging +from yunohost.domain import _get_maindomain, domain_list +from yunohost.utils.network import get_public_ip from yunohost.utils.error import YunohostError +logger = logging.getLogger('yunohost.utils.yunopaste') def yunopaste(data): paste_server = "https://paste.yunohost.org" + try: + data = anonymize(data) + except Exception as e: + logger.warning("For some reason, YunoHost was not able to anonymize the pasted data. Sorry about that. Be careful about sharing the link, as it may contain somewhat private infos like domain names or IP addresses. Error: %s" % e) + try: r = requests.post("%s/documents" % paste_server, data=data, timeout=30) except Exception as e: @@ -24,3 +33,39 @@ def yunopaste(data): raise YunohostError("Uhoh, couldn't parse the answer from paste.yunohost.org : %s" % r.text, raw_msg=True) return "%s/raw/%s" % (paste_server, url) + + +def anonymize(data): + + # First, let's replace every occurence of the main domain by "domain.tld" + # This should cover a good fraction of the info leaked + main_domain = _get_maindomain() + data = data.replace(main_domain, "maindomain.tld") + + # Next, let's replace other domains. We do this in increasing lengths, + # because e.g. knowing that the domain is a sub-domain of another domain may + # still be informative. + # So e.g. if there's jitsi.foobar.com as a subdomain of foobar.com, it may + # be interesting to know that the log is about a supposedly dedicated domain + # for jisti (hopefully this explanation make sense). + domains = domain_list()["domains"] + domains = sorted(domains, key=lambda d: len(d)) + + count = 2 + for domain in domains: + if domain not in data: + continue + data = data.replace(domain, "domain%s.tld" % count) + count += 1 + + # We also want to anonymize the ips + ipv4 = get_public_ip() + ipv6 = get_public_ip(6) + + if ipv4: + data = data.replace(str(ipv4), "xx.xx.xx.xx") + + if ipv6: + data = data.replace(str(ipv6), "xx:xx:xx:xx:xx:xx") + + return data