fix: improved url pattern string (#30)

* fix: improved url pattern string

replaced 0-9 with %d
added %% which is used for matching safe string urls

* test: added test for encoded urls

Author: Justinas Grigas

lua/gx/handlers/url.lua

@@ -8,12 +8,12 @@ local M = {
 
 -- get url from line (with http/s)
 function M.handle(mode, line, _)
-  local pattern = "(https?://[a-zA-Z0-9_/%-%.~@\\+#=?&:]+)"
+  local pattern = "(https?://[a-zA-Z%d_/%%%-%.~@\\+#=?&:]+)"
   local url = helper.find(line, mode, pattern)
 
   -- match url without http(s)
   if not url then
-    pattern = "([a-zA-Z0-9_/%-%.~@\\+#]+%.[a-zA-Z0-9_/%-%.~@\\+#%=?&:]+)"
+    pattern = "([a-zA-Z%d_/%-%.~@\\+#]+%.[a-zA-Z%d_/%%%-%.~@\\+#=?&:]+)"
     url = helper.find(line, mode, pattern)
     if url then
       return "https://" .. url

test/spec/gx/handlers/url_handler_spec.lua

@@ -51,4 +51,21 @@ describe("url_parser_does_work", function()
     assert.equals("https://github.com", handler.handle("v", "// github.com"))
     assert.equals("https://github.com", handler.handle("v", 'print("github.com")'))
   end)
+
+  it("encoded urls", function()
+    assert.equals(
+      "https://company.example.com/team-name/repo-name/-/merge_requests/new?merge_request%5Bsource_branch%5D=branch-name",
+      handler.handle(
+        "v",
+        "https://company.example.com/team-name/repo-name/-/merge_requests/new?merge_request%5Bsource_branch%5D=branch-name"
+      )
+    )
+    assert.equals(
+      "http://example.com/?encoded=%E3%81%93%E3%82%93%E3%81%AB%E3%81%A1%E3%81%AF%21%22%C2%A3%24%25%5E%26%2A%28%29-_%3D%2B%7B%5B%7D%5D%23~%27",
+      handler.handle(
+        "v",
+        "http://example.com/?encoded=%E3%81%93%E3%82%93%E3%81%AB%E3%81%A1%E3%81%AF%21%22%C2%A3%24%25%5E%26%2A%28%29-_%3D%2B%7B%5B%7D%5D%23~%27"
+      )
+    )
+  end)
 end)