Release v6.10.2 (2024-05-15)

Published 2024-05-15 13:00:22Z

Server

Sorting improvements: If you are using fields of type string a re-index is recommended (see below).

/api/eas/rput?skip_head_request=1|0: This new parameter for the eas/rput allows to skip the HEAD request which fylr performs to check the Remote URL. When using skip_head_request the filename parameter is mandatory. This is useful for migrations from source systems which cannot handle the load fylr produces when requesting HEAD on many URLs. Skipping the HEAD is now hard wired default for fylr restore and available as a checkbox in our front end JSON Importer.
/api/search: Support for sort.numeric. With this option, columns of type string can be sorted naturally (1, 2, 10 instead of 1, 10, 2). This option was present in easydb 5 but missed implementation in fylr. For databases which use columns of type string a re-index is required in order to make use of this feature, without re-index the sorting by string columns with numeric: true will be ignored.

Standard rendering: Some cases of standard rendering were broken due to an over complicated algorithm on how to interpret format configurations for standard rendering. Now, we use a very straightforward interpretation with no exceptions. The format defined by the value will be used for output, no matter which format the following value used. In rare cases (e.g. nested standard with parentheses or brackets and comma) this might affect how the standard info is rendered. Most of these cases can be fixed by changing the comma format to space to restore the previous behaviour.
fylr restore: Improved handling of checks for --include-password. Also, allow to restore without purge (only data model and base config are uploaded). Fixed removing the .lck file if an error occurs.
/inspect/migration: Add a new option No Purge. Also, make All Versions the default.
/xmlmapping/tags: Improved mappable tags like Track1:ImageWidth, Track2:ImageWidth, Track3:ImageWidth. This improves mapping capabilities for videos.

fylr convert: Convert images with broken color profiles. If such an error is encountered we ignore the color profile and try again. This fixes issues with some JPEGs where preview production failed. Affected images need to be re-synced using /inspect/files.
fylr convert: Support black and white TIFF. For these TIFF a special case has been added where we don't use magick convert to read the TIFF but the Go TIFF library. Affected images need to be re-synced using /inspect/files.
Re-Index objects on file errors: If an error occurs on an original file, we now re-index the object. This fixes an everlasting pending state in the standard cache which lead to an hourglass icon showing in the standard search result for erroneous files.
Event Stream Listener: Catch panics on startup. In rare cases, fylr would panic at startup due to already closed listeners. This can happen if a 3rd party application tries to quickly re-connect websockets upon a starting fylr with low timeouts.
DELETE /api/user|group: When deleting users or groups, the deletion no longer fails if the record is references as who in a transition. Now, the relations to the transition are removed upon deleting the user or group. This may lead to empty transitions which do not execute on any user or group interactions. Manual deletion or amendment is needed for such transitions after deletion.

JSON Importer: A new option "skip_head_request" has been added when using the "Remote URL" upload type. This prevents the fylr server from making a head request when downloading the file. If this option is enabled, the user must specify the file name in the JSON payload.

PDF Creator: A design issue with nested fields has been fixed.
PDF Print: A bug has been fixed where fields to be printed were not correctly hidden when using the "Select fields" option in the print manager.

Standard View: The configuration of the standard view mode in the main search has been refactored. The new configuration panel allows users to show or hide specific elements in the main search display. This gives users much more control over the design of records displayed in the main search.
CSV Importer Linked Mappings: The CSV importer has been improved and now allows the use of the same linked object field for multiple mappings if it is going to be mapped to specific elements within a nested structure. This makes the CSV importer more compatible with CSVs exported from fylr.
Specific Nested Indexes on CSV Importer: Detection of user-specified indices in nested table name columns has been improved. This enhances the compatibility of CSVs exported from fylr with the CSV importer.
Tag Editor: The design of the tag editor has been improved.

Last updated 2 months ago