Skip to content

Testing UI #2858

@pared

Description

@pared

It seems like we are having more and more UI issues.
Just to give a few examples:

#2839, #2691, #2602. #2854

It seems to me that there are two main problems that are the reason for this:

  1. we introduced external_repo after we had pretty well-established exception handling inside "normal" Repo (for example using relpaths inside error messages), and now we are having quite a few issues that are basically "rephrase X error because it looks strange/bad/uninformative when raised in external repo context"
  2. we do not pay too much attention to what is displayed in the log, and we do not test it

How to deal with that:
As to 1. : Eager: go through our exceptions and consider how they will look like in terminal in the context of external_repo. Lazy: just wait for issues and make changes accordingly. In both cases, tests would be useful for the future.
As to 2.: I think that we should befriend caplog, pay attention to it and check what is written to it, at least at the most important parts of tests that we write.

What do you think? @iterative/engineering

Metadata

Metadata

Assignees

No one assigned

    Labels

    discussionrequires active participation to reach a conclusion

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions