function search-pages--filter__bad_names() { name=$(jq .results[$entry].name "$run_path/build/json/$file" | sed "s|en-US|enUS|;" | jq .enUS | sed "s|^\"||; s|\"$||;"); if [ "$name" == "null" ]; then name=$(jq .results[$entry].name "$run_path/build/json/$file" | sed "s|en-GB|enGB|;" | jq .enGB | sed "s|^\"||; s|\"$||;"); fi # The API `name` parameter makes bad MediaWiki URL's - https://github.com/mozilla/addons-server/issues/8790 # Clean titles. Example of bad titles: "Ghostery – Privacy Ad Blocker" # - # Hard to filter out "-" since many add-ons use it, see for example https://addons.mozilla.org/en-US/firefox/search/?platform=linux&q=1-click # Flash Video Downloader - YouTube HD Download [4K] # Find all names that don't contain Aa-Zz 0-9 and white spaces should be much more reliable than this: name=$(echo "$name" | sed ' s/ — /\n/g; s/ – /\n/g; s/ - /\n/g; # looks identical to the prevoius line but is unique s/: /\n/g; s/(/\n/g; s/,/\n/g; s/™/\n/g; ' | head -n 1 | sed 's/ $//'); }