Better diffs

[SimonBaeumer]

It would be nice to print better diffs between the expected and actual result

[SimonBaeumer]

@dylanhitt What do you think about printing the diff as a table?

For each test a table:

Expected	Operator	Result
1	Equal	3
True	Equal	False
Hello	contained in	Hey World

As a summary:

Or as complete summary like this (maybe a readable like a sentence with the Operation column):

Test	Expected	Operation	Result
it should be equal	1	does not equal	3
it is true	True	does not equal	False
it contains hello	Hello	is not contained in	Hey World

Some more ideas:

• If a text is too long it will be cropped, i.e. after 100 characters
• If a text is more than, i.e. 10 lines, it will be cropped
• The operator is filled defined in the corresponding Matcher

The table would have the benefit that it is always consistent. The disadvantage is that it is not as verbose and detailed as it is currently. Imho it could be one of many output formats for the result and will be added along the current one.

[dylanhitt]

Let's start with that. I was thinking about how to deal with large text files. A bit of a cop out but it would help, we could scan the string until we reach the first point where there is a difference and pinpoint that spot something like hello w^rld. Of course we could print something cleaner than this example. But i'd say switching to table would be best.

I like the second table a lot actually. I didn't quite pick up that could be read like a sentence when I first read it. I need to stop skimming text 🙃

[SimonBaeumer]

Yes, a good diff library would be very cool. But I think we can't achieve a good (human) readability and displaying the difference in the results. For example what do we do if the diff is too big?

[dylanhitt]

You don't necessarily have to print the whole string just the context and the area around it that is different, maybe include the line number where it is first different. I'm not really proposing we do this now, it was mainly just spitting out ideas.