Released Version 1.14.0 and 1.13.2

[hylkevds]

Hi All,

Recently we've released version 1.14.0 and 1.13.2.

While 1.13.2 is simply a maintenance release with many updated dependencies, a few problems were also fixed:

• Fixed the search_path if Liquibase changes it. When applying database updates, Liquibase changes the search_path of the database connection it uses. If the DB admin set a manual search_path (possibly because PostGIS functions are in a different schema) this breaks the database. More information in this Liquibase Issue. FROST-Server now detects it if Liquibase changed the search_path, and changes it back.
• Fixed double URLEncode in $select of nextLinks. This is not noticeable for standards-conform nextLinks, since they do not contain characters that need to be URL-encoded, but with deep select and distinct select, the nextLink can contain characters that need to be URL-encoded.
• Fixed error returned when creating a Sensor with NULL metadata. Sensor.metadata must be set, but when setting it to an explicit JSON NULL, FROST though it was set, even though NULL is not a valid value. FROST now correctly returns an "Incomplete Entity" error in this case.
• Fixed $count when combined with $select=distinct:. FROST would return the non-distinct count, now it returns the correct count.

Of course these issues are also fixed in 1.14.0

Version 1.14.0 has two main features:

• Added support for Batch Requests in JSON format instead of MultiPart format. This makes Batch Requests much easier to create. (Thanks @pbaumard). The documentation for this feature can be found on the documentation site
• Increased the performance of the Datastream-update triggers that run on Observation insert/update/delete by up to 20%. To find out if the ObservedArea of a Datastream changed by the addition of a new Observation, a quite expensive query was needed. The Datastream now keeps track of which FeatureOfInterest was last used, and if a new Observation uses the same FeatureOfInterest, there is no need to update the ObservedArea.

We always appreciate feedback and comments, so let us know what you think!