Important If you are using Python from a python.org 64-bit-only Python installer for macOS 10.9 and later, no further action is needed to use IDLE or tkinter. A built-in version of Tcl/Tk 8.6 will be used. If you are using Python (prior to 3.7) from a python.org 64-bit/32-bit Python installer for macOS 10.6 and later, you should only use IDLE or tkinter with an updated third-party Tcl/Tk 8.5 (not 8.6), like installed. If you are using macOS 10.6, do not use IDLE or Tkinter from the Apple-supplied Python 2.6.1 in macOS 10.6. If possible, install and use a newer version of Python and of Tcl/Tk. If you are using macOS 10.7 or later, the Apple-supplied Tcl/Tk 8.5 still has serious bugs that can cause application crashes. If you wish to use IDLE or Tkinter, install and use a newer version of Python and of Tcl/Tk.
The Cg Toolkit is a legacy NVIDIA toolkit no longer under active. Graphics APIs, and most operating systems (Windows XP, Vista and Windows 7, Mac OS X for.
Python's integrated development environment, and the it uses, depend on the which is not part of Python itself. For best results, it is important that the proper release of Tcl/Tk is installed on your machine.
For recent Python installers for macOS downloadable from this website, here is a summary of current recommendations followed by more detailed information. Python Release Installer Variant macOS Release Recommended Tcl/Tk Alternate Tcl/Tk Not Recommended all 10.6 to 10.14, 64-bit-only 10.9 to 10.14, 64-/32-bit 10.7 to 10.14 10.6. Tk On macOS There are currently three major variants of Tk in common use on macOS: Aqua Cocoa Tk A newer native implementation available as a universal 64-bit and 32-bit binary. This variant is the standard native macOS variant in Tk 8.6 and as of Tk 8.5.13. Aqua Cocoa support was backported to Tk 8.5 (prior to 8.5.13) and released by Apple starting with macOS 10.6 and by ActiveState starting with their 8.5.9.1 release. Aqua Carbon Tk Because it is implemented with older macOS Carbon interfaces, it is only available as a 32-bit binary (usually for Intel and PowerPC processors).
Aqua Carbon Tk 8.4 is included with macOS releases 10.4 through 10.14 and is also available from ActiveState. Aqua Carbon variants of Tk 8.5 had been available as an ActiveState Community Download prior to ActiveTcl 8.5.9. As of 8.5.13, the Tk project no longer supports Carbon builds of Tk 8.5. 32-bit-only Python installers downloadable from this website for older Python releases were linked with Aqua Carbon Tk 8.4. X11 Tk The traditional platform-independent UNIX Tk implementation which requires an X11 server, such as the Apple X11.app available as an optional component in all recent macOS releases. 64-bit and 32-bit binaries can be built. While the Python installers downloadable from this website do not support X11 Tk, other distributors of Python for macOS may do so.
ActiveTcl 8.5.18.0 ActiveState provides binary distributions of Tcl/Tk which are upward compatible with and generally more up-to-date than those provided by Apple in macOS releases. This version of Tcl/Tk includes fixes for some critical problems that you may encounter using tkinter or IDLE (see below). You can download an installer for this release from.
Download Open Graphics Toolkit For Mac
Note that ActiveState Community Edition binaries are not open source and are covered by an ActiveState license. You should read the license before downloading to verify that your usage complies with its terms of use.
This is an Aqua Cocoa Tk. Apple 8.5.9 This release is included in macOS 10.7 through at least macOS 10.14. As of this writing, there are at least two known issues with Tk 8.5.9 that are present in Apple 8.5.9 Tk but fixed in more recent upstream 8.5 releases. The more serious problem is an immediate crash in Tk when entering a composition character, like Option-u on a US keyboard.
(This problem is documented as.) There is also the more general problem of input manager support for composite characters which has also been fixed in more recent Tcl/Tk 8.5 releases. You can avoid these problems by using a current python.org installer that links with or by using a third-party distribution of Python that does not use Apple 8.5.9 Tk. This is an Aqua Cocoa Tk.
Note While Tcl and Tk are separate frameworks and libraries, they are closely related and are normally installed or updated simultaneously. You should not attempt to mix-and-match Tcl and Tk versions. References to a specific version of Tk assume the corresponding version of Tcl is installed as well. The Python for macOS installers downloaded from this website dynamically link at runtime to Tcl/Tk macOS frameworks. The Tcl/Tk major version is determined when the installer is created and cannot be overridden. All python.org installers for Python 3.7.x and 64-bit-only macOS installers for Python 3.6.x and 2.7.x link to their own built-in Tcl/Tk 8.6 frameworks and do not use external Tcl/Tk frameworks.
The Python 64-bit/32-bit macOS installers for Python 3.6.x and and 2.7.x dynamically link to Tcl/Tk 8.5 frameworks. The dynamically linking occurs when tkinter (Python 3) or Tkinter (Python 2) is first imported (specifically, the internal tkinter C extension module). By default, the macOS dynamic linker looks first in /Library/Frameworks for Tcl and Tk frameworks with the proper major version. This is the standard location for third-party or built from source frameworks, including the ActiveTcl releases. If frameworks of the proper major version are not found there, the dynamic linker looks for the same version in /System/Library/Frameworks, the location for Apple-supplied frameworks shipped with macOS. (Note, you should normally not modify or delete files in /System/Library.) As is common on macOS, the installed Pythons and the Tcl and Tk frameworks are built to run on multiple CPU architectures ( universal binaries) and across multiple macOS levels ( minimum deployment target). For Python to be able to dynamically link with a particular Tcl and Tk version, the available architectures in the Tcl/Tk frameworks must include the architecture that Python is running in and their minimum deployment target should be no greater than that of Python.