Cyberduck

We previously only had mirror repository for Cyberduck on GitHub and managed pull requests for changes on our own Git server and mirrored all changes to a SVN repository which was used to display a timeline of changes in Trac.

We are now using the repository on GitHub as the primary source root and accept pull requests at the same place. Previous milestone history has been preserved. This will make contributions more straightforward and simplify the development setup for many.

Issues

We now use Github as well to manage all issues containing bug reports or feature requests. We have migrated all previous 11919 tickets opened in Trac since 2005 to GitHub.

Documentation

Additionally we will also retire the current documentation in the Wiki and move it to docs.cyberduck.io. Contributions to the documentation written in Markdown are welcome can be made by opening a pull request.

As the latest version of macOS now numbered 12 will arrive later today both Cyberduck and Mountain Duck are fully compatible.

Cyberduck CLI is now available for the growing ARM ecosystem. Run the universal file transfer tool on AWS EC2 A1 instance types or Raspberry Pis for example. Both DEB and RPM package types are available in 32bit and 64bit for ARM. The regular installation instructions apply.

We have made some adjustments in version 7.7 of Cyberduck to support the new window appearance introduced in macOS 11. Switching between bookmarks and browser view is now always visible in the toolbar.

The bookmark proxy icon is now displayed inline with the other toolbar controls including the bookmark nickname replacing the previous window title.

Additionally the input field to search for bookmarks or files is now located in the toolbar and treated accordingly.

When horizontal space is limited, the toolbar can display the Search button in place of the search bar. When people click the Search button, the bar expands; when they click elsewhere in the window, the search bar collapses and the toolbar displays the button again.

We have made available a feature in Cyberduck 7.6 and later to restore files from Glacier storage class in S3. You can temporarily restore files from Glacier using File → Restore for files with the storage class set to Glacier. The file will be restored using standard retrieval and expire in 2 days after retrieval. Restoring takes some time and attempting to download an item not yet restored will lead to an error The operation is not valid for the object’s storage class. 

With Cyberduck 7 we introduce the possibility of a multi-segmented download. Multi-Segmented downloads offer the functionality to download one file over several connections. This has the advantage that some files can be downloaded faster and the download process is less error-prone.

In previous versions of Cyberduck, there was only the possibility to run downloads over a single serial connection, the multisegmented download feature multiplies the connections making a parallel download process possible. The following diagram visualizes the two types of download described.

You can expect at least a doubled download speed if you use the multisegment download functionality. Some additional time is required to concatenate the downloaded segments after the download itself is complete. We have collected the following data in download tests with downloads from Amazon S3 and a Minio server running locally.

	AWS EC21	Local System2
Regular Download	20 MByte/s	200 Mbyte/s
Segmented Download	40 Mbyte/s	1 GByte/s

¹ AWS EC2: T2.Medium against a regular S3 endpoint

² Local Minio S3 Server feeding data from local NVMe-PCIe-SSD

The described feature can be turned on or off in Preferences → Transfers → General → Downloads.

Refer to the full changelog for many more enhancements and bug fixes.