What is the minimum scope for v1?
Opened this issue · 3 comments
Guys, I was wondering what would be the initial scope of the project (or, if you prefer, when we will consider the first version as 'Done'):
1 - Image comparison && flagging is obviously the main feature and should be on v1.
2 - Should we provide a default binding with Selenium to take screenshots?
3 - Should we provide a CI-friendly interface?
4 - Should we build a web interface for Severino?
Please, submit your thoughts on these :)
- Are you talking about manual flagging or just identifying that two screenshots are different?
- I think so. This is pretty straightforward and will be the basis for all our comparisons.
- Generating a tap stating whether two pages with the same name are different should not be a problem. If we go for xUnit xml files we can even add some extra data adding a reference to the actual files.
- A web interface would be nice for training Severino and for seeing past runs, but I don't think this is mandatory for the first release.
What do you think?
1 - I was thinking about this... maybe we should "manual flag" the revision as a whole. Because if one screen is messed up, the whole set "fails" the acceptance critieria. I think we should flag revisions as "good" or "bad" and then change the status on a per-revision basis.
2 - ok
3 - agree creating a tap should be easy. let's do it IF it takes us a very little time, as expected.
4 - Ok, I'm going for a command line interface then :)
1 - Good idea. But maybe we shouldn't be strict about revisions. If you wanna run the tool against a local version, you wouldn't have a revision yet, right? Unless the revision is a timestamp :)
4 - I think the initial commands should be something like:
- capture
- compare (latest two folders by default, optionally <dir_a> <dir_b>)
- clean (all folders or oldest N folders)