Re: Properly testing an OTP application after initialization
Roger Lipscombe <[email protected]>
| Newsgroups | gmane.comp.lang.erlang.general |
|---|---|
| Message-ID | <CAJgnQd-7pm2jhv6qBaCyFQ1cGT6NgvzkRpM4_-9-xKnaeaj1eQ@mail.gmail.com> |
I wrote something about how we use Common Test to test (almost) our entire system: https://blog.differentpla.net/blog/2020/09/14/erlang-common-test/ It might be useful to you. On Sun, 13 Sep 2020 at 18:18, Lincoln Bryant <[email protected]> wrote: > > Hello list, > > I'm writing an OTP application that has many things I would like to > test _after_ my application has started. The way I'm currently testing > is to wrap eunit in a utility function that I invoke in the shell, > which then goes through and does all sorts of stateful things against > my gen_servers and reports back. This seems to work okay, but seems > incompatible with e.g. 'rebar3 eunit' and generally seems like it may > not be the intended way to use the tool. > > Now that I want to start wiring the testing framework up to my CI > tool, I am wondering if I should do something like start my > application, connect a shell to it, run my tests, and shut down? or > should I take a step back entirely and consider using something like > Common Test instead? > > Thanks much, > Lincoln