Release v6.7.3 (2023-11-14)

Published 2023-11-14 15:08:08Z

Release v6.7.3

This release is a maintenance release which improves the reliability of fylr and fixes some important bugs. It also adds more features to /inspect/files. Also in this release is the support of the file extension JFIF.


  • Support JFIF file format. This format is essentially a JPEG with more fixed configuration. This change adds recognition for .jfif extensions.

  • Base config: The login info text now allows Markdown format. With this, the information for the user on the login page can be formatted with basic attributes like bold and italics.

  • /inspect/migration: Add a new feature to the restore to link versions from the source and do not produce them in the restored instance. Also, add --upload-ignore-files-with-errors to make the restore more robust.


  • /inspect/files got some new buttons and a clean-up. Also, duplicate entries when searching for parents and children are now removed. A new feature allows to resync all files found in a search. Because of that the pre-defined searches for "Original in status done" and "Files in error state" have been removed. For status and version pulldowns we added the current database counts. The version pulldown allows to filter for original now.

  • Faster file workers: reduce clipping path metadata check. fylr metadata is not compiling has_clipping_path any more by default. The new parameter --clipping-path will tell it to run magick identify in order to determine whether a TIFF has a clipping path or not. The recipes for image conversion have been adapted to only ask "fylr convert" to --clip if the conversion is TIFF -> PNG. Before this was also done for other TIFF -> * conversions, but since we are cropping background it really only makes sense for PNG target.

  • Show a better error message if fylr doesn't start due to a misconfigured license file.

  • /inspect/migration: Add a checkbox to use --link-versions for fylr restore. With this, versions can be linked from the source and do not need to be produced by the restored instance, this is a helpful feature for test migrations.


  • Login of guests has been fixed.

  • Distorted colors of downloaded files in RGB format with applied color profile have been fixed.

  • PDF file worker now uses the configured temporary working directory in the exec server and no longer the OS temp dir.

  • fylr restore --continue now remembers ids of restored metadata mappings. This avoids an error which occurred while restoring collections with mappings (with the use of --continue).

Web Frontend

  • A new search field has been added in the expert search to search for text in the field metadata_fulltext of assets, allowing searches within the text of documents if it has been indexed.

  • Number inputs with default values in the base config are now displayed correctly, using placeholders.

  • A new option has been added when printing PDFs to include the change history of objects at the end of the PDF.

  • Corrections have been made to the filter panel in the main search.

  • Corrections have been made to the "individual pools" option in the collections configuration.

  • The representation of wildcard searches in the query badges of the main search has been fixed.

  • Sharing searches via URL has been corrected, now all search parameters are transferred correctly.

  • Error handling has been improved in the datamodel section when using objectstore.

  • A check has been added when using the CSV importer to import hierarchical objects; if a parent is specified and does not exist, the user will be notified.

  • An error has been fixed when using lookup:id in the JSON importer in objects that contain assets configured by URLs.

  • A bug where multimedia playback was not stopped correctly in fullscreen mode when changing objects has been fixed.

  • Now, when changing objects in fullscreen mode, the selection in the main search will change according to the object being viewed and the search view will be scrolled to the selected object.

  • Fixed a bug where files with filetypes not recognized by fylr could not be downloaded. Now, these files can be downloaded and exported in their original version.

Last updated