Doesn't work under Java 11

[Engelberg]

Culprit is the version of fipp you rely on is an older fork that hasn't been updated to rely on the new rrb-vector. In Java 11, a toArray method was added so it is now overloaded in a way that requires a type hint to be resolved. This affected rrb-vector, which has now been fixed, but libs depending directly or indirectly on the old version don't work.

https://www.deps.co/blog/how-to-upgrade-clojure-projects-to-use-java-11/

[dgrnbrg]

I’d be happy to merge a or with the upgrade! Thanks for the report.

…

On Tue, Dec 18, 2018 at 4:49 AM Mark Engelberg ***@***.***> wrote: Culprit is the version of fipp you rely on is an older fork that hasn't been updated to rely on the new rrb-vector. In Java 11, a toArray method was added so it is now overloaded in a way that requires a type hint to be resolved. This affected rrb-vector, which has now been fixed, but libs depending directly or indirectly on the old version don't work. https://www.deps.co/blog/how-to-upgrade-clojure-projects-to-use-java-11/ — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#32>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAlwC8z7XxqD4kIv6wSyMYyot3IKjGZNks5u6LongaJpZM4ZX8Sp> .

[Engelberg]

Do you remember why you chose bigml/fipp over brandonbloom/fipp? Easiest fix would be to switch to brandonbloom/fipp, but I want to make sure you don't lose functionality you were expecting from the fork.

[dgrnbrg]

I think I originally used Brandon’s lib, but someone sent a PR to fix a different bug :)

…

On Tue, Dec 18, 2018 at 3:26 PM Mark Engelberg ***@***.***> wrote: Do you remember why you chose bigml/fipp over brandonbloom/fipp? Easiest fix would be to switch to brandonbloom/fipp, but I want to make sure you don't lose functionality you were expecting from the fork. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#32 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAlwCyp0Br4O0gAdE4EGTKYLRDqBPIOoks5u6U9ZgaJpZM4ZX8Sp> .

[Engelberg]

It all comes down to rrb-vector, which is a dependency of fipp and therefore puget-printer. The new version is compatible with Java 11, but also exhibits a clojurescript compilation problem under some circumstances (like when using cljsbuild as the build tool).

Interestingly, it looks like you aren't even using puget-printer when in clojurescript, so I'm not really sure why the clojurescript build even needs to compile puget-printer and therefore fipp and rrb-vector. It looks like in core.cljc you use conditional macro to require puget-printer only under clj, so I wouldn't expect puget-printer to even be compiled in clojurescript.

It seems to me that if it's possible to figure out why puget-printer is getting pulled into clojurescript, and just eliminate it from the cljs build, it would be safe to bump the dependency up to the latest puget-printer and have everything working on jdk 11, while compiling correctly on cljs.

[giodamelio]

In the mean time I was able to work around this issue (when using Leiningen) by forcing a newer version of rr-vector (the latest is 0.1.1 at the time of this comment). Just add the snippet below to you project.clj.

;; Force Spyscope to use a newer version of rrb-vector that is compatible with Java 11
:managed-dependencies [[org.clojure/core.rrb-vector "0.1.1"]]

[maris-orbidans-otm]

Doesn't work under java 15

openjdk version "15.0.2" 2021-01-19