We could shortcut the following lines if there is only one README file

Do you mean like this ?
if len(readmes) > 1:
for i in order:
for j in readmes:
if i == j.name:
return j
elif len(readmes) == 1:
return readmes[0]
return None

Something like this yes

I wonder if we couldn't do something like:

readme_names = [j.name for j in readmes]
return readmes[min([order.index(f) for f in readme_names])]

(untested)

Finally, it would be cool to add some tests for this.

Other than that, it looks great, thanks! :)

I'll test your version

This fails when there's a readme that's not in the order list.
order.index('README.notinlist')
ValueError: 'README.notinlist' is not in list

What is we use:

readmes[min([order.index(f) for f in readme_names if f in order])] 

?

Your version picks README.Propaganda for the https://src.fedoraproject.org/rpms/desktop-backgrounds use-case.

The lists readmes and order are not in sync, so I suspect you meant

readmes[min([readme_names.index(f) for f in readme_names if f in order])]

But that still fails when there's only a README available that is not in the order list:

ValueError: min() arg is an empty sequence

I'd like to keep it 'as is' for now as I've tested my version with all combinations I could think of and concentrate on adding tests so that any code changes can be easily verified

I'd like to keep it 'as is' for now as I've tested my version with all combinations I could think of and concentrate on adding tests so that any code changes can be easily verified

Alright, thanks for double-checking and +1 for the tests :)

Could you rebase rather than merging?

tests are all passing, thanks! :)