I just heard, for the first time, that workflows a...
# suitescript
k
I just heard, for the first time, that workflows are actually converted to scripts before running them and then they are converted to JVM byte code. How accurate it that? If it is, does that mean suitescripts go through only 1 conversion while workflow goes through 2?
s
Yes, if true that would mean two 'conversions' . However a couple thoughts on that - 1. it would be prudent for NS to do the first conversion when you save the workflow. 2. SuiteScript (with NFT) is usually shorter and easier than doing a workflow. Hence I'd generally only use a workflow for things only workflows can do.
e
And it's only transpiled into Java on the server-side
and that's true for both WF and SS
a
This is a subject I would love to hear more detail about.
👍🏼 1
@Brendan Boyd
b
Oh, that's interesting. I hadn't heard that one before.
s
I haven't looked at Graal in detail, but being a modern javascript engine I would expect it to be efficient - nearly as much as pure Java
(assuming SS2.1)
b
Oh, I see, if everything's transpiling/compiling to GraalVM that kinda makes sense. Not sure about Netsuite, but I've seen really good performance metrics with Java and the Quarkus project on Kubernetes environments. I heard Netsuite is migrating more of it's workloads to Kubernetes, so all makes more sense now...
But yes, whether js or java, my understanding is it should all be running as bytecode very fast on Graal, no?
Can I take this to mean that SS 2.1 should run faster than 2.0 and 1.0? https://www.netsuite.com/blog/graal-runtime-technology-improves-netsuite-platform-developer-productivity
m
It's definitely true that client side workflows are converted to SuiteScript 1.0. I would assume the same with server side
message has been deleted
message has been deleted
r
TIL, you can use 1.0 syntax in Formula condition
m
Only for client side triggers. For server side it needs to be sql
s
@Brendan Boyd I have heard from NS once that SS2.1 is indeed faster than 2.0 (due to Graal) but I don't recall any hard numbers.
👍 1
e
Unlikely to receive any hard numbers either; publishing performance metrics breaks the TOS =/