RegEx matches domain names and IP addresses

[xCh12i5]

pi-hole/gravity.sh

Line 288 in ddbdb51

sed -nr -e 's/\.{2,}/./g' -e '/\./p' > ${destination}

In the gravity_ParseFileIntoDomains() function the regular expression sed -nr -e 's/\.{2,}/./g' -e '/\./p' matches domain names as well as IP addresses. In case it is intended to match exclusively domain names the regular expression should look like sed -r '/([^\.]+\.)+[^\.]{2,}/!d'.

Proof of concept with the original RegEx:
echo "google.com" | awk -F '#' '{print $1}' | awk -F '/' '{print $1}' | awk '($1 !~ /^#/) { if (NF>1) {print $2} else {print $1}}' | sed -nr -e 's/\.{2,}/./g' -e '/\./p'
google.com
echo "8.8.8.8" | awk -F '#' '{print $1}' | awk -F '/' '{print $1}' | awk '($1 !~ /^#/) { if (NF>1) {print $2} else {print $1}}' | sed -nr -e 's/\.{2,}/./g' -e '/\./p'
8.8.8.8

Proof of concept with the proposed RegEx:
echo "8.8.8.8" | tr -d '\r' | tr '[:upper:]' '[:lower:]' | sed -r '/(\/|#).*$/d' | sed -r 's/^.*\s+//g' | sed -r '/([^\.]+\.)+[^\.]{2,}/!d'
echo "google.com" | tr -d '\r' | tr '[:upper:]' '[:lower:]' | sed -r '/(\/|#).*$/d' | sed -r 's/^.*\s+//g' | sed -r '/([^\.]+\.)+[^\.]{2,}/!d'
google.com

[dschaper]

Does this manifest as a bug or an issue that needs to be resolved? What are the benefits for implementing the proposed changes?

[dschaper]

No response. Will implement. This is a better approach to select the domain name from the line in the hosts file instead of splitting and hoping the file is in IP FQDN format. Should be able to strip comments and then parse out the domain name only and remove a lot of the extra steps.

[AzureMarker]

See my comment here about reusing the domain and hostname detection implemented in the API: #2495 (comment)

[dschaper]

Thanks, I think this sed approach will work better for shell script. Once the API is done and released we can take a look at what shell can be replaced or eliminated.

[dschaper]

Thanks for the PR xCh12i5!