do not catch UnicodeError, handle that intern

git-svn-id: https://linkchecker.svn.sourceforge.net/svnroot/linkchecker/trunk/linkchecker@3269 e7d03fd6-7b0d-0410-9947-9c21f3af8025

ChangeLog

@@ -65,6 +65,12 @@
     Removed: linkcheck/checker/{ignored,error}url.py
     Added: linkcheck/checker/unknownurl.py
 
+  * Convert the "label too long" domain name parse error into
+    a more friendly error message.
+    Type: bugfix
+    Changed: linkcheck/checker/{__init__,urlbase,httpurl,fileurl}.py,
+      linkchecker
+
 3.4 "The Chumscrubbers" (released 4.2.2006)
 
   * Ignore decoding errors when retrieving the robots.txt URL.

linkcheck/checker/__init__.py

@@ -34,8 +34,6 @@ import linkcheck.dns.exception
 # Catch these exception on syntax checks.
 ExcSyntaxList = [
     linkcheck.LinkCheckerError,
-    # .encode('idna') raises this
-    UnicodeError,
 ]
 
 # Catch these exceptions on content and connect checks. All other

linkcheck/checker/fileurl.py

@@ -107,7 +107,7 @@ class FileUrl (urlbase.UrlBase):
         base_url = re.sub("^file://(/?)([a-zA-Z]):", r"file:///\2|", base_url)
         # norm base url again after changing
         if self.base_url != base_url:
-            base_url, is_idn = linkcheck.url.url_norm(base_url)
+            base_url, is_idn = linkcheck.checker.urlbase.url_norm(base_url)
             self.base_url = unicode(base_url)
 
     def build_url (self):

linkcheck/checker/httpurl.py

@@ -310,7 +310,8 @@ class HttpUrl (internpaturl.InternPatternUrl, proxysupport.ProxySupport):
                 "Redirected to %r", newurl)
             self.add_info(_("Redirected to %(url)s.") % {'url': newurl},
                           tag="http-redirect")
-            redirected, is_idn = linkcheck.url.url_norm(newurl)
+            # norm base url - can raise UnicodeError from url.idna_encode()
+            redirected, is_idn = linkcheck.checker.urlbase.url_norm(newurl)
             assert None == linkcheck.log.debug(linkcheck.LOG_CHECK,
                 "Norm redirected to %r", redirected)
             urlparts = linkcheck.strformat.url_unicode_split(redirected)

linkcheck/checker/urlbase.py

@@ -53,6 +53,17 @@ def urljoin (parent, url, scheme):
     return urlparse.urljoin(parent, url)
 
 
+def url_norm (url):
+    """
+    Wrapper for url.url_norm() to convert UnicodeError in LinkCheckerError.
+    """
+    try:
+        return linkcheck.url.url_norm(url)
+    except UnicodeError:
+        msg = _("URL has unparsable domain name: %s") % sys.exc_info()[1]
+        raise linkcheck.LinkCheckerError(msg)
+
+
 class UrlBase (object):
     """
     An URL with additional information like validity etc.
@@ -294,8 +305,8 @@ class UrlBase (object):
         Construct self.url and self.urlparts out of the given base
         url information self.base_url, self.parent_url and self.base_ref.
         """
-        # norm base url
-        base_url, is_idn = linkcheck.url.url_norm(self.base_url)
+        # norm base url - can raise UnicodeError from url.idna_encode()
+        base_url, is_idn = url_norm(self.base_url)
         if is_idn:
             self.add_warning(_("""URL %r has a unicode domain name which
                           is not yet widely supported. You should use

linkchecker

@@ -715,7 +715,12 @@ for url in args:
         # syntactic sugar
         url = "ftp://%s" % url
     url_data = get_url_from(url, 0, aggregate, assume_local=True)
-    linkcheck.add_intern_pattern(url_data, config)
+    try:
+        linkcheck.add_intern_pattern(url_data, config)
+    except UnicodeError:
+        linkcheck.log.error(linkcheck.LOG_CMDLINE,
+            _("URL has unparsable domain name: %s"), sys.exc_info()[1])
+        sys.exit(1)
     aggregate.urlqueue.put(url_data)
 # set up profiling/psyco
 if do_profile and not has_profile: