Open louis-bompart opened 6 days ago
:x: Title should follow the conventional commit spec:
<type>(optional scope): <description>
Example: feat(headless): add result-list controller
File | Old (kb) | New (kb) | Change (%) |
---|---|---|---|
case-assist | 235.9 | 235.9 | 0 |
commerce | 339.1 | 339.1 | 0 |
search | 409.5 | 409.5 | 0 |
insight | 395.4 | 395.4 | 0 |
recommendation | 248.5 | 248.5 | 0 |
ssr | 403.5 | 403.5 | 0 |
ssr-commerce | 351.4 | 351.4 | 0 |
Use case | SSR (#) | CSR (#) | Progress (%) |
---|---|---|---|
search | 39 | 44 | 89 |
recommendation | 0 | 4 | 0 |
case-assist | 0 | 6 | 0 |
insight | 0 | 27 | 0 |
commerce | 0 | 15 | 0 |
tinker with one playwright test to see if we could cram Pact in there...
Mayyybe? Lot of effort to implement on our side, but if we tool ourselves up, and we document Pact ('cause sorry, the DX of their matches is pretty poor and doesn't do any effort to play nice with frontend developers' brains)
@y-lakhdar , just throwing that here as 'yep, that's possible, but', if that helps ya. @acote-coveo , poking you as the 'contract-guru' 🧙, I'm mostly interested to see if there's value for an API maintainer such as yourself in
pacts/Atomic-SearchUI.json
@lvu285 , if you discuss contract testing with other teams, we have a (very dirty) POC.
Why in Playwright? If we look at Headless's test landscape, we don't have much E2E. They are done at the level of Atomic (so Atomic+Headless+API, the whole whammy).
Decoupling Headless from Atomic would be another ordeal, so for the sake of the POC, I went with 'Headless&Atomic' and put the API aside.
Next steps:
Evaluate 'new test' cost & gain v.s. cache approach
To be investigated https://pactumjs.github.io/guides/contract-testing.html