Katrina Owen
Katrina Owen
Oooh, I like that!
@Goltergaul I didn't realize there were differences between the two rubies. I can see a few things we could do (there are likely other possible approaches as well). - don't...
> leaving out the pretty generate will make it too hard to spot differences in the approvals. You're absolutely right. > Maybe we could use a pure ruby json serializer...
> Not sure if the non-pretty one could also have differences though. Yeah, that's true. It could be worth a 30 minute experiment to find out.
A quick question to clarify. You're suggesting that we add more descriptive specs? https://github.com/approvals/approval_tests/blob/master/spec/approval_tests/approvals_spec.rb
It's important to NOT change the output. 1. A lot of the testing is for legacy things that do not have valid structure, and also sometimes for fragments that get...
(Sorry it took so long to respond. I'm at a conference, and I find it very hard to keep track of issues/pull requests during conferences).
> Spaces can be very significant in html, and changing it can break your app. Ah, see this is where I fall down because I don't know front-end stuff! OK,...
Could we make a rough guess using mean values from other tracks? Or is Prolog different enough that it would be a bad guess?