Canopy 2.1 Known Issues

Please read the list of known issues in the Canopy 2.1.3 release notes (June 16, 2017). More recent issues are listed here:

Clarification of Canopy 2.1's change in "default User environment": 

  1. In Canopy 1.x, there was only one Canopy User Python 2.7 environment. You could optionally tell Canopy to make this be the default Python on your system, and Canopy would add it to your PATH environment variable.
  2. Similarly, when Canopy 2 is first started, it creates one Canopy Python environment (either Python 2.7 or 3.5, depending on which installer you run) named "User". 
  3. But in addition, in Canopy 2.1, you can create multiple environments (Python 2.7, 3.5, or 3.6). At any given time, one of these environments is active within the Canopy application
  4. Canopy's active environment is also the default python in any Canopy Command Prompt / Terminal that you open from the Canopy Tools menu.
  5. By default, the active environment in Canopy is the "User" environment mentioned in (2). You can change which environment is active from the Canopy Preferences dialog. See "Switching Python Environments" in the Canopy 2.1 documentation.
  6. Canopy no longer offers an option, as in (1), to make a Canopy Python environment be the default Python on your system by adding it to your PATH environment (though you can still do this manually if you wish; for more information on adding Canopy Python to your path, please see Make Canopy 2 User Python be your default Python.)

Other known issues:

  • Tiny fonts and icons on high-resolution displays: Fixing this will require converting Canopy to use the newest version of the Qt GUI toolkit. This is high on the roadmap but does not yet have an ETA. Meanwhile, you can work around this by reducing your machine's display resolution. We understand that this is not a great solution, and apologize for the inconvenience.
  • "RuntimeError:EDM lock is not held by the current process": This unusual error may occur while using the Canopy Package Manager. If it does, please quit and restart the Package Manager and try your action again. If the error repeats, please please quit and restart Canopy and try your action again, and let us know by email to [email protected], that restarting Package Manager did not suffice.
  • Crash when opening Editor Window on macOS: Some users report that Canopy crashes when clicking the Editor button from the Welcome Screen or when opening a file that would normally open the Editor window. For more details about this issue as well as workarounds to resolve the problem, please see macOS Crash When Opening Editor or Importing gdbm.
  • Canopy's editor and debugger do not work consistently with Python source files stored on virtual file systems such as Google Drive or Dropbox.
  • Matplotlib plots not showing with Error: TypeError: 'figure' is an unknown keyword argument: This problem is encountered due to a bug in the interaction between Canopy and the current provided version of the IPython kernel. The quick fix is to ensure that the backend settings under the Python Pane of the Canopy Preferences have the following settings: Pylab mode On and Backend set to "auto-select". This error should only occur when the backend setting is manually set to Qt4-interactive, all other settings should function correctly.

Fixed issues (as of Canopy 2.1.9):

  • Canopy does not start on Ubuntu 17.10: as a workaround, you can install the exact same Python distribution for command-line-only use (i.e. without the Canopy GUI which is exhibiting this bug), by using the Enthought Deployment Manager (EDM). For more information about this problem and potential workarounds, please see Canopy Won't Start on Ubuntu 17 and Other Linux Variants. Fixed in Canopy 2.1.8, March 2018.
  • Blurry text and images on high-resolution displays on Mac. This was fixed in Canopy 2.1.7, January, 2018.
  • "AttributeError:'NoneType' object has no attribute cent" within a Managed Common Install (MCI): The Canopy error reporter will pop up with this message if a non-admin user attempts to access the Package Manager in a MCI (admin-controlled). In a MCI, non-admin users cannot use the Package Manager and cannot switch between Python environments. See "Canopy 2 Managed Common Install".

Please do not enter support requests in article comments

Please use article comments for suggestions to improve the article. See "How to submit Enthought Canopy/EPD support requests".

Have more questions? Submit a request

Comments

Powered by Zendesk