From b88c81df8f4c5168af454eaa4f92afa9349e4e13 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ionic=C4=83=20Biz=C4=83u?= <bizauionica@gmail.com>
Date: Wed, 3 Aug 2022 11:22:27 +0300
Subject: [PATCH] Throw if url is invalid. Add a length limit.

---
 lib/index.js  | 20 +++++++++++++++++---
 test/index.js | 42 ++++++++++++++++++++++++++++++++----------
 2 files changed, 49 insertions(+), 13 deletions(-)

diff --git a/lib/index.js b/lib/index.js
index b806431..fe5a052 100644
--- a/lib/index.js
+++ b/lib/index.js
@@ -32,18 +32,27 @@ import normalizeUrl from "normalize-url";
  *    - `search` (String): The url querystring value.
  *    - `href` (String): The input url.
  *    - `query` (Object): The url querystring, parsed as object.
+ *    - `parse_failed` (Boolean): Whether the parsing failed or not.
  */
 const parseUrl = (url, normalize = false) => {
 
     // Constants
     const GIT_RE = /(^(git@|http(s)?:\/\/)([\w\.\-@]+)(\/|:))(([\~,\.\w,\-,\_,\/]+)(.git){0,1}((\/){0,1}))/
 
-    if (typeof url !== "string" || !url.trim()) {
-        const err = new Error("Invalid url.")
+    const throwErr = msg => {
+        const err = new Error(msg)
         err.subject_url = url
         throw err
     }
 
+    if (typeof url !== "string" || !url.trim()) {
+        throwErr("Invalid url.")
+    }
+
+    if (url.length > parseUrl.MAX_INPUT_LENGTH) {
+        throwErr("Input exceeds maximum length. If needed, change the value of parseUrl.MAX_INPUT_LENGTH.")
+    }
+
     if (normalize) {
         if (typeof normalize !== "object") {
             normalize = {
@@ -56,7 +65,7 @@ const parseUrl = (url, normalize = false) => {
     const parsed = parsePath(url)
 
     // Potential git-ssh urls
-    if (parsed.protocol === "file") {
+    if (parsed.parse_failed) {
         const matched  = parsed.href.match(GIT_RE)
         if (matched) {
             parsed.protocols = ["ssh"]
@@ -65,10 +74,15 @@ const parseUrl = (url, normalize = false) => {
             parsed.host = matched[4]
             parsed.user = "git"
             parsed.pathname = `/${matched[6]}`
+            parsed.parse_failed = false
+        } else {
+            throwErr("URL parsing failed.")
         }
     }
 
     return parsed;
 }
 
+parseUrl.MAX_INPUT_LENGTH = 2048
+
 export default parseUrl;
diff --git a/test/index.js b/test/index.js
index 1a50435..288c36c 100644
--- a/test/index.js
+++ b/test/index.js
@@ -17,6 +17,7 @@ const INPUTS = [
           , hash: ""
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -32,6 +33,7 @@ const INPUTS = [
           , hash: ""
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -47,6 +49,7 @@ const INPUTS = [
           , hash: "some-hash?foo=bar"
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -62,6 +65,7 @@ const INPUTS = [
           , hash: ""
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -77,6 +81,7 @@ const INPUTS = [
           , hash: ""
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -92,6 +97,7 @@ const INPUTS = [
           , hash: ""
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -107,22 +113,24 @@ const INPUTS = [
           , hash: "http://a:1:1"
           , search: ""
           , query: {}
+          , parse_failed: false
         }
     ]
   , [
         ["git@github.my-enterprise.com:my-org/my-repo.git", false],
         {
             protocols: [ 'ssh' ]
-            , protocol: 'ssh'
-            , port: ''
-            , resource: 'github.my-enterprise.com'
-            , host: 'github.my-enterprise.com'
-            , user: 'git'
-            , password: ''
-            , pathname: '/my-org/my-repo.git'
-            , hash: ''
-            , search: ''
-            , query: {}
+          , protocol: 'ssh'
+          , port: ''
+          , resource: 'github.my-enterprise.com'
+          , host: 'github.my-enterprise.com'
+          , user: 'git'
+          , password: ''
+          , pathname: '/my-org/my-repo.git'
+          , hash: ''
+          , search: ''
+          , query: {}
+          , parse_failed: false
         }
     ]
   , [
@@ -138,6 +146,7 @@ const INPUTS = [
         , hash: ""
         , search: ""
         , query: {}
+        , parse_failed: false
       }
   ]
 ];
@@ -165,4 +174,17 @@ tester.describe("check urls", test => {
             parseUrl("")
         }).toThrow(/invalid url/i)
     })
+
+    test.should("throw if url is too long", () => {
+        parseUrl.MAX_INPUT_LENGTH = 10
+        test.expect(() => {
+            parseUrl("https://domain.com/")
+        }).toThrow(/input exceeds maximum length/i)
+    })
+
+    test.should("throw if url is invalid", () => {
+        test.expect(() => {
+            parseUrl("foo")
+        }).toThrow(/url parsing failed/i)
+    })
 });