Version 1.0 of the Anaconda Launcher features Python 3 support, the ability to easily manage environments and channels, and a sandboxed environment for installing new packages.
Python 3 Support
The new version of Launcher drops its dependency on Qt and Python and now uses node-webkit to be entirely self-contained. This means you can install the Launcher with either Python 2.7 or Python 3.4 as the default Python version. Additionally, a number of bugs have been fixed:
- Links on the right hand side now work in Windows
- Icons display in Mac OSX 10.7
Manage Environments and Channels
No matter if you install the Python 2.7 or Python 3.4 version of Launcher, you can always create environments using conda that run other versions of Python (2.6, 2.7, 3.3, 3.4) and libraries, including NumPy, SciPy, and pandas.
The new version of Launcher makes managing these environments even easier. Create, clone, or remove environments from within the Launcher, allowing you to have separate installations of applications with different versions of Python (or any library) in each. Additionally, you can add and remove Binstar channels using the Launcher, making it easy to find more apps to install.
Install Sandboxed Apps
Managing dependencies can be difficult due to conflicting requirements, especially for packages with many dependencies. Launcher will now install certain applications into their own sandboxed environments, ensuring they have the right dependencies without conflicting with other packages you may have installed. Try it out with some of the new packages in the Anaconda repository, like Veusz.
If you want to package such applications yourself, grab the latest version of conda-build and add
app: own_environment: true
Get It Now
conda update launcher
or, for Anaconda users with Python 3,
conda install launcher