unable to import gtk.Window -> gio: TypeError: type 'gio.MemoryOutputStream' is not dynamically allocated but its base type '__main__.GPollableOutputStream' is dynamically allocated

[James Page]

Hit this after the most recent updates in vivid:

** (process:21949): WARNING **: couldn't make the type `gio.MemoryInputStream' ready

** (process:21949): WARNING **: couldn't make the type `gio.MemoryOutputStream' ready
TypeError: type 'gio.MemoryOutputStream' is not dynamically allocated but its base type '__main__.GPollableOutputStream' is dynamically allocated

** (process:21949): WARNING **: couldn't make the type `gio.unix.InputStream' ready

** (process:21949): WARNING **: couldn't make the type `gio.unix.OutputStream' ready
/usr/lib/python2.7/dist-packages/gtk-2.0/gtk/__init__.py:40: RuntimeWarning: tp_compare didn't return -1 or -2 for exception
  from gtk import _gtk
ImportError: could not import gio
ImportError: could not import gio
ImportError: cannot import name Object from gtk
Traceback (most recent call last):
  File "/usr/bin/terminator", line 102, in <module>
    TERMINATOR.create_layout(OPTIONS.layout)
  File "/usr/share/terminator/terminatorlib/terminator.py", line 240, in create_layout
    window, terminal = self.new_window()
  File "/usr/share/terminator/terminatorlib/terminator.py", line 166, in new_window
    window = maker.make('Window')
  File "/usr/share/terminator/terminatorlib/factory.py", line 94, in make
    output = func(**kwargs)
  File "/usr/share/terminator/terminatorlib/factory.py", line 100, in make_window
    import window
  File "/usr/share/terminator/terminatorlib/window.py", line 28, in <module>
    class Window(Container, gtk.Window):
AttributeError: 'module' object has no attribute 'Window'

As neither terminator or pyobject-2 has changed, I would suspect the python update that landed on the 26th.

ProblemType: Bug
DistroRelease: Ubuntu 15.04
Package: terminator 0.97-4
ProcVersionSignature: Ubuntu 3.18.0-13.14-generic 3.18.5
Uname: Linux 3.18.0-13-generic x86_64
ApportVersion: 2.16.1-0ubuntu2
Architecture: amd64
CurrentDesktop: Unity
Date: Fri Feb 27 09:56:01 2015
EcryptfsInUse: Yes
InstallationDate: Installed on 2014-11-25 (93 days ago)
InstallationMedia: Ubuntu 15.04 "Vivid Vervet" - Alpha amd64 (20141124)
PackageArchitecture: all
SourcePackage: terminator
UpgradeStatus: No upgrade log present (probably fresh install)

[James Page (james-page)]

Probably this:

     - Issue #22079: PyType_Ready() now checks that statically allocated type has
      no dynamically allocated bases.

[Matthias Klose (doko)]

http://bugs.python.org/issue22079

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Iain Lane (laney)]

Reverting the commit in python2.7, but likely the real fix is in pygobject-2

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Matthias Klose (doko)]

we should revert that in python3.4 as well, if we cannot fix this in pygobject

[Launchpad Janitor (janitor)]

This bug was fixed in the package python2.7 - 2.7.9-1ubuntu2

---------------
python2.7 (2.7.9-1ubuntu2) vivid; urgency=medium

  * Revert change from previous upload
    - Issue #22079: PyType_Ready() now checks that statically allocated type has
      no dynamically allocated bases.
    As it causes 'gio' from pygobject-2 to fail to initialise, which breaks
    several consuming applications. (LP: #1426294)
 -- Iain Lane <email address hidden> Fri, 27 Feb 2015 17:53:39 +0000

[Iain Lane (laney)]

On Fri, Feb 27, 2015 at 07:17:22PM -0000, Matthias Klose wrote:
> we should revert that in python3.4 as well, if we cannot fix this in
> pygobject

AIUI this is only available for python 2

--
Iain Lane [ <email address hidden> ]
Debian Developer [ <email address hidden> ]
Ubuntu Developer [ <email address hidden> ]

[Matthias Klose (doko)]

On 02/27/2015 11:44 PM, Iain Lane wrote:
> On Fri, Feb 27, 2015 at 07:17:22PM -0000, Matthias Klose wrote:
>> we should revert that in python3.4 as well, if we cannot fix this in
>> pygobject
>
> AIUI this is only available for python 2

the commit itself was done on all python branches.

[Iain Lane (laney)]

On Sat, Feb 28, 2015 at 09:43:14AM -0000, Matthias Klose wrote:
> On 02/27/2015 11:44 PM, Iain Lane wrote:
> > On Fri, Feb 27, 2015 at 07:17:22PM -0000, Matthias Klose wrote:
> >> we should revert that in python3.4 as well, if we cannot fix this in
> >> pygobject
> >
> > AIUI this is only available for python 2
>
> the commit itself was done on all python branches.

I know. But if you think that fixing pygobject-2 is enough then it
follows that you don't need the revert in python3.4 too, because
pygobject-2 isn't available there.

Unless you're aware of other instances of the same kind of bug.

--
Iain Lane [ <email address hidden> ]
Debian Developer [ <email address hidden> ]
Ubuntu Developer [ <email address hidden> ]

[Launchpad Janitor (janitor)]

Status changed to 'Confirmed' because the bug affects multiple users.

[Launchpad Janitor (janitor)]

This bug was fixed in the package python2.7 - 2.7.9-2ubuntu3

---------------
python2.7 (2.7.9-2ubuntu3) vivid; urgency=medium

  * Update to 20150401, taken from the 2.7 release branch.
    - Issue #23629: Fix the default __sizeof__ implementation for
      variable-sized objects.
    - Issue #21526: Tkinter now supports new boolean type in Tcl 8.5.
    - Issue #23838: linecache now clears the cache and returns an empty
      result on MemoryError.
    - Issue #23742: ntpath.expandvars() no longer loses unbalanced single
      quotes.
    - Issue #21802: The reader in BufferedRWPair now is closed even when
      closing writer failed in BufferedRWPair.close().
    - Issue #23671: string.Template now allows to specify the "self" parameter
      as keyword argument. string.Formatter now allows to specify the "self"
      and the "format_string" parameters as keyword arguments.
    - Issue #21560: An attempt to write a data of wrong type no longer cause
      GzipFile corruption.
    - Issue #23647: Increase imaplib's MAXLINE to accommodate modern mailbox
      sizes.
    - Issue #23539: If body is None, http.client.HTTPConnection.request now
      sets Content-Length to 0 for PUT, POST, and PATCH headers to avoid
      411 errors from some web servers.
    - Issue #23136: _strptime now uniformly handles all days in week 0,
      including Dec 30 of previous year.
    - Issue #23138: Fixed parsing cookies with absent keys or values in
      cookiejar.
    - Issue #23051: multiprocessing.Pool methods imap() and imap_unordered()
      now handle exceptions raised by an iterator.
    - Issue #22928: Disabled HTTP header injections in httplib.
    - Issue #23615: Module tarfile now can be reloaded with imp.reload().
    - Issue #23799: Added test.test_support.start_threads() for running and
      cleaning up multiple threads.
    - Issue #22390: test.regrtest now emits a warning if temporary files or
      directories are left after running a test.
    - Issue #23583: Added tests for standard IO streams in IDLE.
    - Issue #23583: Fixed writing unicode to standard output stream in IDLE.
  * Re-apply the fix for issue #22079. This is now a warning instead of an
    error. LP: #1426294.
  * Fix issue #23842, SystemError in os.minor, os.major. LP: #1435242.
 -- Matthias Klose <email address hidden> Thu, 02 Apr 2015 16:13:26 +0200