Remove support for adblock style lists to prevent false positives

Signed-off-by: Adam Warner <me@adamwarner.co.uk>
This commit is contained in:
Adam Warner 2019-09-03 23:43:11 +01:00
parent e7af42a9f8
commit ad41bcca5a
No known key found for this signature in database
GPG key ID: 872950F3ECF2B173

View file

@ -353,46 +353,7 @@ gravity_ParseFileIntoDomains() {
# Determine how to parse individual source file formats # Determine how to parse individual source file formats
if [[ "${firstLine,,}" =~ (adblock|ublock|^!) ]]; then if [[ "${firstLine,,}" =~ (adblock|ublock|^!) ]]; then
# Compare $firstLine against lower case words found in Adblock lists # Compare $firstLine against lower case words found in Adblock lists
echo -ne " ${INFO} Format: Adblock" echo -ne " ${CROSS} Format: Adblock (list type not supported)"
# Define symbols used as comments: [!
# "||.*^" includes the "Example 2" domains we can extract
# https://adblockplus.org/filter-cheatsheet
abpFilter="/^(\\[|!)|^(\\|\\|.*\\^)/"
# Parse Adblock lists by extracting "Example 2" domains
# Logic: Ignore lines which do not include comments or domain name anchor
awk ''"${abpFilter}"' {
# Remove valid adblock type options
gsub(/\$?~?(important|third-party|popup|subdocument|websocket),?/, "", $0)
# Remove starting domain name anchor "||" and ending seperator "^"
gsub(/^(\|\|)|(\^)/, "", $0)
# Remove invalid characters (*/,=$)
if($0 ~ /[*\/,=\$]/) { $0="" }
# Remove lines which are only IPv4 addresses
if($0 ~ /^[0-9]+\.[0-9]+\.[0-9]+\.[0-9]+$/) { $0="" }
if($0) { print $0 }
}' "${source}" > "${destination}"
# Determine if there are Adblock exception rules
# https://adblockplus.org/filters
if grep -q "^@@||" "${source}" &> /dev/null; then
# Parse Adblock lists by extracting exception rules
# Logic: Ignore lines which do not include exception format "@@||example.com^"
awk -F "[|^]" '/^@@\|\|.*\^/ {
# Remove valid adblock type options
gsub(/\$?~?(third-party)/, "", $0)
# Remove invalid characters (*/,=$)
if($0 ~ /[*\/,=\$]/) { $0="" }
if($3) { print $3 }
}' "${source}" > "${destination}.exceptionsFile.tmp"
# Remove exceptions
comm -23 "${destination}" <(sort "${destination}.exceptionsFile.tmp") > "${source}"
mv "${source}" "${destination}"
fi
echo -e "${OVER} ${TICK} Format: Adblock"
elif grep -q "^address=/" "${source}" &> /dev/null; then elif grep -q "^address=/" "${source}" &> /dev/null; then
# Parse Dnsmasq format lists # Parse Dnsmasq format lists
echo -e " ${CROSS} Format: Dnsmasq (list type not supported)" echo -e " ${CROSS} Format: Dnsmasq (list type not supported)"