fix: Fix error where a lot of domains were not detected

Fix separator logic in trie serializer Affected domains: - `niteroi.br / *.nom.br` - `police.uk / *.sch.uk` - `budejju.no / nes.buskerud.no` - `hønefoss.no / os.hordaland.no` - `molde.no / heroy.more-og-romsdal.no` - `nordkapp.no / bo.nordland.no` - `osterøy.no / valer.ostfold.no` - `tananger.no / bo.telemark.no` - `vestby.no / sande.vestfold.no` - `haugesund.no / os.hedmark.no` Fixes #64
peerigon · May 27, 2019 · fc80789 · fc80789
1 parent 1d1a24a
commit fc80789
Show file tree

Hide file tree

Showing 5 changed files with 13 additions and 3 deletions.
diff --git a/src/tries/serializeTrie.js b/src/tries/serializeTrie.js
@@ -42,14 +42,15 @@ function lineToString(line, i, arr) {
         const prevLine = arr[i - 1];
 
         indexOfDifference = findIndexOfDifference(line, prevLine);
+
         if (indexOfDifference === -1) {
             // Identical lines
             return "";
         }
         if (indexOfDifference === 0) {
             // line and prevLine are completely different
             separatorFromPrev = SEPARATORS.RESET;
-        } else if (prevLine.length === line.length && indexOfDifference === line.length - 1) {
+        } else if (prevLine.length <= line.length && indexOfDifference === prevLine.length - 1) {
             // only the last part of line and prevLine are different
             separatorFromPrev = SEPARATORS.SAME;
         } else if (indexOfDifference > prevLine.length - 1) {

diff --git a/test/parseDomain.test.js b/test/parseDomain.test.js
@@ -172,6 +172,14 @@ describe("parseDomain(url)", () => {
         });
     });
 
+    it("should parse police.uk as tld", () => {
+        expect(parseDomain("example.police.uk")).to.eql({
+            subdomain: "",
+            domain: "example",
+            tld: "police.uk",
+        });
+    });
+
     it("should also work with custom top-level domains passed as regexps", () => {
         const options = {customTlds: /(\.local|localhost)$/};
 

diff --git a/test/snapshots.test.js b/test/snapshots.test.js
@@ -26,7 +26,7 @@ describe("snapshots", () => {
             fs.writeFileSync(pathToParsePubSuffixListSnapshot, JSON.stringify(parsedList));
         });
     });
-    describe("serialieTrie()", () => {
+    describe("serializeTrie()", () => {
         it("matches the approved snapshot", () => {
             const parsedList = JSON.parse(fs.readFileSync(pathToParsePubSuffixListSnapshot, "utf8"));
             const snapshot = JSON.parse(fs.readFileSync(pathToSerializeTrieSnapshot, "utf8"));

diff --git a/test/snapshots/serializeTrie.json b/test/snapshots/serializeTrie.json
diff --git a/test/tries/serializeTrie.test.js b/test/tries/serializeTrie.test.js
@@ -78,6 +78,7 @@ describe("serializeTrie()", () => {
             [["pl", "gov.pl", "ap.gov.pl", "net.pl"], "pl>gov>ap<net"],
             [["pl", "gov.pl", "ap.gov.pl", "uk", "ac.uk", "co.uk"], "pl>gov>ap|uk>ac,co"],
             [["jp", "岐阜.jp", "静岡.jp", "موقع"], "jp>岐阜,静岡|موقع"],
+            [["uk", "ac.uk", "police.uk", "*.sch.uk"], "uk>ac,police,sch>*"],
         ].forEach(args => {
             const parsedList = args[0];
             const expectedString = args[1];