google/go-cmp

String diff for UUIDs can use strings.Join syntax which makes it more confusing

prashantv opened this issue · 4 comments

prashantv commented

This is a similar topic as #274 and #195, but when using cmp to diff UUIDs.

When the UUIDs are very different, the output looks like:

  	ID: strings.Join({
- 		"e93d2d8b-7d2a-4c62-967a-b189347a4fa2",
+ 		"7e93c401-4e27-3b83-7edc-2dd0803c39b3",
  	}, ""),

The strings.Join adds unnecessary noise here, ideally it would be displayed as:

-  	ID: "e93d2d8b-7d2a-4c62-967a-b189347a4fa2",
+ 	ID: "7e93c401-4e27-3b83-7edc-2dd0803c39b3",

However, if the UUIDs have common characters, the output gets really verbose, and isn't particularly useful:

  	ID: strings.Join({
- 		"217d",
  		"6",
- 		"7ef-d686-4107-ab35-347ed6c91a22",
+ 		"68416ab-a20e-48ae-afaa-8df7b91a4af3",
  	}, ""),

Is there a way to avoid this specialized string handling and have it never use the strings.Join output but show a simpler remove/add for string fields?

dsnet commented

Thanks for filing. This is a bug. There are more differences than similarities, that it shouldn't have chosen this particular representation.

  • 👍1
dsnet commented

For my own benefit, here's a reproduction: https://go.dev/play/p/7u4lcejtNsf

prashantv commented

Thanks for the additional context. Even in the case that the UUIDs have significant overlap, the strings.Join format is noisy, and causes other issues for code that tries to detect and replace UUIDs.

Ideally we could turn off this feature completely, is there existing options that can be passed to control this behaviour?

  • 👍1
richunger commented

Looks like this was introduced in 0.5.6. Works correctly in 0.5.5.