gitweb: Ensure to strip any prefix before git clone URL

Some gitweb instances can have some string prefixes before the displayed
git clone URLs so ensure to strip them to properly extract URLs.

Related to swh/infra/sysadm-environment#5051.

swh/lister/gitweb/lister.py

@@ -130,6 +130,12 @@ class GitwebLister(StatelessLister[Repositories]):
         urls = []
         for row in bs.find_all("tr", {"class": "metadata_url"}):
             url = row.contents[-1].string.strip()
+            for scheme in ("http", "https", "git"):
+                # remove any string prefix before origin
+                pos = url.find(f"{scheme}://")
+                if pos != -1:
+                    url = url[pos:]
+                    break
 
             if "," in url:
                 urls_ = [s.strip() for s in url.split(",") if s]

swh/lister/gitweb/tests/data/https_git.distorted.org.uk/~mdw_firewall

@@ -43,8 +43,8 @@ summary | <a href="https://git.distorted.org.uk/~mdw/firewall/shortlog">shortlog
 <tr id="metadata_desc"><td>description</td><td>Firewall scripts for distorted.org.uk.</td></tr>
 <tr id="metadata_owner"><td>owner</td><td>Mark Wooding</td></tr>
 <tr id="metadata_lchange"><td>last change</td><td><span class="datetime">Thu, 16 Mar 2023 18:09:32 +0000</span> (18:09 +0000)</td></tr>
-<tr class="metadata_url"><td>URL</td><td>https://git.distorted.org.uk/~mdw/firewall</td></tr>
-<tr class="metadata_url"><td></td><td>git://git.distorted.org.uk/~mdw/firewall</td></tr>
+<tr class="metadata_url"><td>URL</td><td>fallback: https://git.distorted.org.uk/~mdw/firewall</td></tr>
+<tr class="metadata_url"><td></td><td>fast: git://git.distorted.org.uk/~mdw/firewall</td></tr>
 </table>
 <div class="header">
 <a class="title" href="https://git.distorted.org.uk/~mdw/firewall/shortlog">shortlog</a>
@@ -164,4 +164,4 @@ window.onload = function () {
 };
 </script>
 </body>
-</html>
\ No newline at end of file
+</html>