KATT (Klarna API Testing Tool) Is an HTTP-Based API Testing Tool for Erlang | Hacker News Reader

KATT (Klarna API Testing Tool) Is an HTTP-Based API Testing Tool for Erlang

github.com

57 pointsthunderbong15 comments

I've recently used Hurl [1] in some projects. It seems to have a similar goal and I like i a lot though it misses "imports" which makes modeling of more complex workflows hard.

[1]: https://hurl.dev

How does the "imports" feature supposed to work and how does it help modelling more complex workflows?

Also, would you be able to take a look at Step CI and tell me what you think about it, especially in comparison to Hurl?

https://github.com/stepci/stepci

It could be that Step CI already has the feature you're looking for!

> How does the "imports" feature supposed to work and how does it help modelling more complex workflows?

If hurl had something like an import feature, one could re-use HTTP requests. For example, one could create a file that logs into the service and use it in all other tests. It doesn't support that, so the only option you have is to re-write the login in every test _or_ build a script that outputs that. There is an open issue for that: https://github.com/Orange-OpenSource/hurl/issues/115

Gotcha!

https://docs.stepci.com/guides/organising-tests.html

We will also support reusable sessions/credentials soon

Why would I use this instead of other alternatives?

also, the name is misleading, the project is not made by Klarna

Regarding the name, quoting from the README:

"Despite the "Klarna" mention, this repository is not affiliated with Klarna AB. KATT was indeed born at Klarna, and Klarna AB holds copyright for parts of the code, but it is now being maintained outside the company, by its original authors and new contributors."

> Klarna AB holds copyright for parts of the code

Do they have a permission to redistribute the code?

This all feels legally complicated

The code is released under the Apache-2.0 license. Doesn't seem complicated to me at all.

Assuming that license related files in the repo are accurate this doesn't look all that complicated.

The License and notice files appear to be sufficient to understand the situation and there's really not a lot going on. The licensing files indicate that Klarna AB holds copyrights in the code (with some others mentioned) and that it is all licensed under Apache 2.0.

Naturally, the license related files in the repo can be inaccurate or incomplete. But when dealing with open source software you have to pick a point where you're going to trust these kinds of representations from strangers.

Perhaps a subtweet to Klarna's layoffs earlier this year

I can't help but think that this is going to result in some confusion.

"Have you tested this API?"

"Yes, I used Katt"

"Cat? You mean netcat or curl"

Katt means cat in Swedish and is pronounced "cut".

Where in Sweden have you ever heard someone pronounce 'katt' anywhere close to 'cut'? Maybe you misheard someone say 'kut' (baby seal)?

Maybe not a perfect comparison but I think it sounds very similar. "kut" on the other hand sounds like neither so not sure why you think I mean that.

cut - IPA(key): /kʌt/

https://en.wiktionary.org/wiki/cut#Pronunciation

katt - IPA(key): /ˈkatː/

https://en.wiktionary.org/wiki/katt#Pronunciation_3

(there's audio on both links)

I think your reply, and the further replies to it only serve to demonstrate my concerns even more.

I've had this happen before, where some tool or technology sounds like another tool. Combine that with people with varying accents, and you can quickly devolve into frustrating conversations.

"Okay, so the Foo system is having problems. Mary said they tested it using some cat scripts. Can we get Mary to share those, as we're unable to reproduce it"

(some time later)

"Oh, so Mary was using Katt with a K and not cat with a C."

Reply on news.ycombinator.com