https://pact.io logo
#protobufs
Title
# protobufs
m

Matt (pactflow.io / pact-js / pact-go)

05/24/2022, 10:18 PM
Right now the plan is to: 1. Finish gRPC / protobufs plugins (work in progress now) 2. Improve the documentation and visibility of the plugin ecosystem - make examples, tutorials, workshops etc. so that maintainers and contributors can use and create them 3. Get the rust core into all key languages (JVM, JS, Golang, .NET, Python) 4. Add plugin support to all key languages Ruby is not on that list, but after (2) it could be done with community support
s

Santiago Martí

05/24/2022, 10:25 PM
Sounds great, thanks Matt!
m

Matt (pactflow.io / pact-js / pact-go)

05/24/2022, 10:26 PM
If you’re interested in picking up (3) for Ruby, there’s nothing stopping you now. See #libpact_ffi-users if you’d like to get involved! (and also probably pop your hand up in #pact-ruby)
s

Santiago Martí

05/25/2022, 8:35 PM
Thanks Matt, will look into it
👍 1
c

Cody A. Ray

06/13/2022, 6:47 PM
You don’t happen to have a rough ETA on GRPC for pact-go, do you? 🙏
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:36 AM
Hey Cody! It’s in a branch, let me get it merged and you can have a play with the next beta release.
🙌 1
c

Cody A. Ray

06/15/2022, 3:36 AM
Momma always said “just ask nicely, you never know” 😂 Thanks!
🙌 1
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:37 AM
What do you plan on testing, btw? If you can give us some idea of your use case that would be great to know
also see the supported/unsupported features here: https://github.com/pactflow/pact-protobuf-plugin/#supported-features
When do you think you’d have the headspace/capacity to play? Right away?
c

Cody A. Ray

06/15/2022, 3:40 AM
Yeah, we have an intern (Runlin) digging into all of this right now. I think day 2 of his internship he had a pact broker running (locally) already and has been writing tests on the HTTP side of the house. Happy to setup a zoom at some point to chat through everything. It definitely feels like right now we’re seeing what all ways we can abuse your tools 😂 We have a bunch of weird requirements we’re trying to balance that leads use quite a few of your tools together. • Ex: bidirectional contract testing looks pretty exciting, and I assume it’ll work better than the swagger-mock-validator we use today — unless that’s what is used under the hood. But we’re using the OSS broker, so this isn’t supported (yet?). • Using the wiremock-pact-generator to try to reduce the work on consumer teams. Runlin already has a POC using this too. But then we expect we’ll want BDCT to validate these much quicker, since my understanding is that a bunch of duplicative tests like we’d generate would be too slow to validate normally. But trying to see if this theory is true • And now we get to add GRPC into the mix, because a lot of our API calls are over GRPC and “faking” our contracts using grpcgateway seems… fallible.
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:46 AM
Ex: bidirectional contract testing looks pretty exciting, and I assume it’ll work better than the swagger-mock-validator we use today — unless that’s what is used under the hood. But we’re using the OSS broker, so this isn’t supported (yet?).
It is, but we have customised it to address a few things.
c

Cody A. Ray

06/15/2022, 3:47 AM
Yeah, I heard from a couple of our teams that tested that it has… issues. 🙈
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:47 AM
Using the wiremock-pact-generator to try to reduce the work on consumer teams. Runlin already has a POC using this too. But then we expect we’ll want BDCT to validate these much quicker, since my understanding is that a bunch of duplicative tests like we’d generate would be too slow to validate normally. But trying to see if this theory is true
💯 do not use the standard pact verification here. It won’t just be duplicative, but likely very flakey and hard for the provider team to verify reliably.
And now we get to add GRPC into the mix, because a lot of our API calls are over GRPC and “faking” our contracts using grpcgateway seems… fallible.
got it. The plugin stuff is still very new and the authoring interface is still a level below where we’ll want to be (you’ll see what I mean), so I expect us to get through the beta phase and get to DX improvements on that. But conceptually we want to get that working first
👌 1
c

Cody A. Ray

06/15/2022, 3:48 AM
Interesting. What would make them flakey? I assumed the main challenge on the providers (besides speed) would be handling a million different provider states.
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:50 AM
Interesting. What would make them flakey? I assumed the main challenge on the providers (besides speed) would be handling a million different provider states.
yes, that’s one. Because the wiremock pact generator doesn’t care for states (unless I’m mistaken), the provider won’t know how to deal with all of the scenarios.
1
c

Cody A. Ray

06/15/2022, 3:50 AM
wiremock itself has pretty solid state/scenario support, but yeah no idea how that translates into the pact.json
m

Matt (pactflow.io / pact-js / pact-go)

06/15/2022, 3:51 AM
to my knowledge it doesn’t. But that would be an obvious addition to the tool
nod hmm yes 1
(sorry, had a phone call)
let me know how the research goes on the supported features. I’m keen to know if the protobufs/gRPC features we support initially will cover the majority of use cases
2 Views