Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Test failures with Python 3.3 #1008

Closed
shimizukawa opened this issue Jan 3, 2015 · 11 comments
Closed

Test failures with Python 3.3 #1008

shimizukawa opened this issue Jan 3, 2015 · 11 comments
Labels

Comments

@shimizukawa
Copy link
Member

5 tests fail with Python 3.3. I use Docutils trunk and Sphinx trunk.

Output of tests in attachment.


@shimizukawa
Copy link
Member Author

From Dmitry Shachnev on 2012-10-25 15:10:05+00:00

This bug is also a big problem for Debian/Ubuntu as we are switching to Python 3.3 by default now.

@shimizukawa
Copy link
Member Author

From Dmitry Shachnev on 2012-10-26 13:22:41+00:00

I've been trying to debug this, the only issue I can find the reason to fail is: [this check](https://github.com/sphinx-doc/testing/blob/5b9f431fc92fc76c01fcb3b86c02c37298f1c2fc/sphinx/util/__init__.py#L200 now) always returns True in Python 3.3 (because every module now has a __loader__), so the first line of regexp in test_build_html.py needs to be adjusted to allow <string> as a module name.

Other issues are (I've no clue what may cause them):

  • autodoc attempts to import sphinx.application.TemplateBridge module;

  • strange

    >>  assert '1 added, 3 changed, 1 removed' in '1 added, 4 changed, 1 removed'
    

    in test_env.py.

@shimizukawa
Copy link
Member Author

From Takayuki Shimizukawa on 2012-10-30 01:11:07+00:00

I confirmed your reported test error with:

  • Python 3.3 (32bit) on Windows7 x64
  • Python 3.3 (32bit) on Ubuntu 12.04 x64
  • docutils trunk (0.10)
  • tox 1.4.2
  • virtualenv 1.8.2
  • nose 1.2.1
  • Sphinx rev eb3967d1e43b (default)

tested by:

$ tox -e py33

sometime caused test error but sometime passed. I'm confusing..

BTW, I fixed a few errors.

I'll continue to fix.

@shimizukawa
Copy link
Member Author

From Dmitry Shachnev on 2012-10-30 09:33:59+00:00

Thanks!

Note that some issues can be fixed by patch from issue #998.

@shimizukawa
Copy link
Member Author

From Takayuki Shimizukawa on 2012-10-30 13:26:11+00:00

I found error reason, but I don't know why working before Python-3.3...

Dmitry, please change your code and test again.

diff -r b1c22e958d7b sphinx/ext/autodoc.py
--- a/sphinx/ext/autodoc.py     Tue Oct 30 20:41:05 2012 +0900
+++ b/sphinx/ext/autodoc.py     Tue Oct 30 22:23:31 2012 +0900
@@ -1102,13 +1102,13 @@
 class MethodDocumenter(DocstringSignatureMixin, ClassLevelDocumenter):
     """
     Specialized Documenter subclass for methods (normal, static and class).
     """
     objtype = 'method'
     member_order = 50
-    priority = 0
+    priority = 1

@shimizukawa
Copy link
Member Author

From Dmitry Shachnev on 2012-10-30 16:48:24+00:00

Takayuki Shimizukawa: After that change, everything works! Thanks very much!

Can you please commit it and propose a pull request?

@shimizukawa
Copy link
Member Author

From Takayuki Shimizukawa on 2012-10-31 04:31:24+00:00

Dmitry Shachnev: Thanks for your confirmation!

I sent PR83.

@shimizukawa
Copy link
Member Author

From Georg Brandl on 2012-11-01 16:54:57+00:00

fix #1008 #1029 : intersphinx_mapping values are not stable if mapping have plural key/value set

→ <>

@shimizukawa
Copy link
Member Author

From Georg Brandl on 2012-11-01 16:54:57+00:00

fix #1008: can't get correct source code filename because all modules have 'loader' attribute at Python-3.3.

→ <>

@shimizukawa
Copy link
Member Author

From Georg Brandl on 2012-11-01 16:54:57+00:00

fix #440 #1008 : coarse timestamp resolution in some filesystem generate wrong outdated file-list

→ <<cset 718993a>>

@shimizukawa
Copy link
Member Author

From Georg Brandl on 2012-11-01 16:54:58+00:00

fix #1008: fix autodoc documenter resolve ordering.

note: MethodDocumenter and FunctionDocumenter is same priority then
Documenter.format_signature() choice documenter is not stable
(depend on memory condition?).

→ <<cset 88e1bbe>>

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
Projects
None yet
Development

No branches or pull requests

1 participant