Testing

[ghost]

Thoughts on adding some tests to help prevent things like #326 from happening again?

[SpaceK33z]

How would you test a websocket connection failing? Sure, it's possible to test small part of this repo, but I think it's very hard to add fully automatic tests.

I think it would be useful to document what you should (manually) test after changing something.

For instance, in example/ there are two html files index.html and inlined.html. You need to verify both of these still work (With PR #302 I forgot to test inlined.html, which caused #326).

[ghost]

Even a light process sounds good given there are thousands of people using
this software every day.
On Sat, Dec 5, 2015 at 8:27 AM Kees Kluskens [email protected]
wrote:

How would you test a websocket connection failing? Sure, it's possible to
test small part of this repo, but I think it's very hard to add fully
automatic tests.

I think it would be useful to document what you should (manually) test
after changing something.

For instance, in example/ there are two html files index.html and
inlined.html. You need to verify both of these still work (With PR #302
#302 I forgot to test
inlined.html, which caused #326
#326).

—
Reply to this email directly or view it on GitHub
#334 (comment)
.

[SpaceK33z]

True that. Do you have time to put together a PR? Maybe start with client/ first, I think that is the most easy and a fairly important part.

[ghost]

Given I lost a couple hours last week I'll consider it. I'd love to hear
more ideas on approach to find something effective and simple.
On Sat, Dec 5, 2015 at 10:50 AM Kees Kluskens [email protected]
wrote:

True that. Do you have time to put together a PR? Maybe start with client/
first, I think that is the most easy and a fairly important part.

—
Reply to this email directly or view it on GitHub
#334 (comment)
.

[ghost]

@SpaceK33z Do you have time to put together a roadmap, so we can add Unit Tests to it for one-dot-oh?

[SpaceK33z]

@JHabdas, 9 months later, I'm starting to work on this :). I've opened #581 for this to have a clean start.

Closing this in favor of #581.