Fix Python tokenizer to not generate empty Text tokens.

mwparserfromhell/parser/tokenizer.py

@@ -424,7 +424,8 @@ class Tokenizer(object):
                 if brackets:
                     self._emit_text(before)
                     self._emit(tokens.ExternalLinkSeparator())
-                    self._emit_text(after)
+                    if after:
+                        self._emit_text(after)
                     self._context ^= contexts.EXT_LINK_URI
                     self._context |= contexts.EXT_LINK_TITLE
                     self._head += 1

tests/tokenizer/external_links.mwtest

@@ -15,7 +15,7 @@ output: [ExternalLinkOpen(brackets=True), Text(text="http://example.com/"), Exte
 name:   brackets_space
 label:  basic external link in brackets, with a space after
 input:  "[http://example.com/ ]"
-output: [ExternalLinkOpen(brackets=True), Text(text="http://example.com/"), ExternalLinkSeparator(), Text(text=""), ExternalLinkClose()]
+output: [ExternalLinkOpen(brackets=True), Text(text="http://example.com/"), ExternalLinkSeparator(), ExternalLinkClose()]
 
 ---