[wwwsearch-commits] r36044 - in wwwsearch/mechanize/trunk: mechanize test

[jjlee at codespeak.net]

Author: jjlee
Date: Fri Dec 29 21:40:06 2006
New Revision: 36044

Modified:
   wwwsearch/mechanize/trunk/mechanize/_html.py
   wwwsearch/mechanize/trunk/test/test_forms.doctest
Log:
Fix non-idempotent behaviour of Factory.forms() / .links() by making Factory not assume restarting behaviour of e.g. FormsFactory.forms() / LinksFactory.links() iterator functions.  Previously, if for example you got a ParseError during execution of .forms(), you could call it again and have it not raise an exception, because it started out where it left off.

Modified: wwwsearch/mechanize/trunk/mechanize/_html.py
==============================================================================
--- wwwsearch/mechanize/trunk/mechanize/_html.py	(original)
+++ wwwsearch/mechanize/trunk/mechanize/_html.py	Fri Dec 29 21:40:06 2006
@@ -509,8 +509,12 @@
         # this implementation sets .global_form as a side-effect, for benefit
         # of __getattr__ impl
         if self._forms_genf is None:
-            self._forms_genf = CachingGeneratorFunction(
-                self._forms_factory.forms())
+            try:
+                self._forms_genf = CachingGeneratorFunction(
+                    self._forms_factory.forms())
+            except:
+                self.set_response(self._response)
+                raise
             self.global_form = getattr(
                 self._forms_factory, "global_form", None)
         return self._forms_genf()
@@ -518,8 +522,12 @@
     def links(self):
         """Return iterable over mechanize.Link-like objects."""
         if self._links_genf is None:
-            self._links_genf = CachingGeneratorFunction(
-                self._links_factory.links())
+            try:
+                self._links_genf = CachingGeneratorFunction(
+                    self._links_factory.links())
+            except:
+                self.set_response(self._response)
+                raise
         return self._links_genf()
 
 class DefaultFactory(Factory):

Modified: wwwsearch/mechanize/trunk/test/test_forms.doctest
==============================================================================
--- wwwsearch/mechanize/trunk/test/test_forms.doctest	(original)
+++ wwwsearch/mechanize/trunk/test/test_forms.doctest	Fri Dec 29 21:40:06 2006
@@ -33,3 +33,27 @@
 >>> br.find_control(type="isindex").value = "blah"
 >>> br.click(type="isindex").get_full_url()
 'http://example.com/?blah'
+
+
+If something (e.g. calling .forms() triggers parsing, and parsing
+fails, the next attempt should not succeed!  This used to happen
+because the response held by LinksFactory etc was stale, since it had
+already been .read().  Fixed by calling Factory.set_response() on
+error.
+
+>>> import mechanize, sgmllib
+>>> br = mechanize.Browser()
+>>> r = mechanize._response.test_html_response("""\
+... <form>
+... <input type="text" name="foo" value="a"></input><!!!>
+... <input type="text" name="bar" value="b"></input>
+... </form>
+... """)
+>>> br.set_response(r)
+>>> try:
+...     br.select_form(nr=0)
+... except sgmllib.SGMLParseError:
+...     pass
+>>> br.select_form(nr=0)  # doctest: +IGNORE_EXCEPTION_DETAIL
+Traceback (most recent call last):
+SGMLParseError: expected name token