Fork me on GitHub

com.walmartlabs/lacinia 0.33.0 and lacinia-pedestal 0.12.0 Lacinia is an open-source implementation of Facebook's GraphQL specification, in Clojure. GraphQL is an outstanding approach to getting diverse clients and servers exchanging data cleanly and efficiently. GitHub repo: Documentation: Featured changes in 0.33.0: - Schema Definition Language now supports empty types and extend type - Handling of nulls and omitted variables now conforms to the spec - Introspection can be disabled at schema compile time - scalars have been simplified, to conform to the spec lacinia-pedestal adds support for accessing GraphQL as an HTTP endpoint GitHub repo: Documentation: Features changes in 0.12.0: - All GraphiQL assets are served from application, not from a CDN

parrot 32
clj 4

@U04VDKC4G are there any known breaking changes for lacinia-pedistal 0.11.0 -> 0.12.0. Build using it fails since the upgrade. From the test I get it's party working. Hope to have some time soon to look into it more.


The changes ate trivial, surprised theres a problem


I'll try to look into it asap, which is probably in about 14 hours from now.


Seems like there is a type issue. When I use the endpoint not working directly in GraphiQL it does work.


It's fixed by explicitly casting the string to an int.


Check the release notes, there were breaking changes (we're still alpha) related to scalars; lacinia 0.33.0 is more strict, to be closer to the spec. On our own internal apps, I had to override the Int scalar to be more forgiving, for compatibility with existing clients.


At some point soon, we need to lock it down as a 1.0 release.


I tried to find them. Also had some ID fields that I turned into Int to make it work.


Didn't share it here before, and just updated the lacinia-pedestal version for master. contains several Clojure applications which can be end-to-end performance tested as a whole. The acquired data can be viewed in different graphs. One part is a GraphQL endpoint mainly using subscriptions, writing and reading to Kafka.