start test suite for a few "happy paths", integration with travis

[hpk42]

From a discussion with @r10s today: before going for g-play it'd be good to have automated tests which run with each commit/push to the delta-core codebase. @flub has done some work on creating python bindings for delta-core and i have experience in writing tests (have eg authored http://pytest.org). So this issue here is about writing a few basic tests (create chat, send a message, receive a message) and integrating it into travis such that the tests run on PRs and pushed commits.

[t-oster]

nice idea. Any FakeSMTP/IMAP libraries which can run during the test and assert incoming/outgoing messages?

[flub]

The python stdlib has a stmpd implementation which can probably be used. imap might be some more work.

[t-oster]

well there is http://www.icegreen.com/greenmail/ but this need java

[ralphtheninja]

I'm going to try and push this further. Enough discussing 😉

#250 is a start to get Travis up and running.

[ralphtheninja]

I've tested deltachat-node with Greenmail but I can't get it to work. A client can connect to IMAP and SMTP (and IMAPS and SMTPS works as well), but sending messages back and forth doesn't work.

Perhaps I should clarify a bit. It works to send messages, but writing integration tests with multiple users is difficult, i.e. send a message from client1 and check that client2 received a message etc.

[r10s]

cc @hpk42

[ralphtheninja]

I'll try to write some more unit tests in deltachat-node using greenmail as backend just to see how for it can take me. Maybe we could have a more extensive setup for integration tests, more as in a "real" IMAP/SMTP environment in the future.

[ralphtheninja]

Also see deltachat/deltachat-node#98

[flub]

Starting the greenmail docker container was what I was thinking of doing for the tests as well, so it makes sense to use this in the node bindings tests too.

[hpk42]

python bindings have a test suite https://github.com/deltachat/deltachat-core/tree/master/python/tests
and the tests run against travis. creating a follow up issues #339 and closing this one here.