yarray / avipost.web Goto Github PK

Selenium + Cucumber caused some trouble, mainly:

1. The error messages of selenium are vague
2. The control flow of cucumber are not clear
3. Confused API, easy to misuse.

Candidates:

For Selenium:

1. Use other browser automation tools like Casper.
2. Use headless fake browser like Zombie.
3. Do not test DOM. Make it as thin as possible, than only test the interface.
4. Do not test real "document", use jsdom to fake it.
5. Stick to selenium, find ways to solve problems 1-3.
6. Use other language bindings of Selenium.

For Cucumber-js

1. Use other libraries like yada
2. Use libraries in other languages
3. Use parse to roll our own

revise JsDoc. The goal is to provide clear type hints for tern (also means we are clear about the interfaces)

Changes include new logics for auth, and modified fixture creation strategy.

After finishing this issue, the tests should reflect the expected application logic, and only fail on assertion.

Facebook's Flow is impressing and is still not mature. Its benefits now are not worthy the visual noise now. If at least it supports const/let we can rethink about it, otherwise just leave it for the future.

After establishing e2e test in backend, it seems unintuitive of nighwatch's test. Is it possible to do cucumber like e2e test with selenium for front end?

https://www.npmjs.com/package/hardy is highly related but seems inactive and incomplete.

Best cucumber port for node.js seems to be https://github.com/cucumber/cucumber-js and quite active.

These as demo projects may be a good start:

https://github.com/pataniqa/web-testing-examples
https://github.com/keithamus/WebCumberNode
https://github.com/jbpros/cukecipes

Also we can use other languages, like python - even ruby since cucumber comes from it.

Several problems:

No oauth setup (manually?)
Flush will destroy oauth tables
No user creation

Since we use sass now, it's better to use mixin rather than letting helper class pollute our html

Think about the "ideal" UI engine for us, including:

What parts should a UI engine contain?
What are the best ways to do that?

The conclusion should contain a conceptual model. The solving of this question should learn from other UI frameworks and can explain them with the final model, to justify their suitability for us.

Other UI frameworks/libraries

• React
• Masonry
• Bootstrap
• Webix

If linter is reasonable, change code, otherwise configure linter

Add auth logic when requesting backend

Mainly integrate it into the build procedure

Since the ci framework for backend has been established, it's time to use real backend to test the frontend.