Lars Ingebrigtsen <larsi@gnus.org> writes:

> Felix Dietrich <felix.dietrich@sperrhaken.name> writes:
>
>> The first, relating to readability: I found it quit difficult to follow
>> the flow of the functionʼs code due to its nesting and idiosyncratic
>> order of condition checking.  I want to, therefore, propose a
>> restructured version, which I believe to be functionally equivalent.
>
> Yes, the code is pretty confusing...  so reading the patches, I'm not
> quite sure whether the new one is equivalent or not.  So I think we'd
> want to have a number of tests in mailcap-tests.el to test that the
> results really are equivalent before and after the change, too.

I agree that adding some automated tests is a good idea, but, since I am
not familiar with writing these in Emacs Lisp, I will first have to
spent a bit of time with the ERT manual and the existing tests, before I
cans start writing the tests.  Meanwhile, to shorten the wait ;), here
is a series of patches that go through the transformation steps of
‘mailcap-add-mailcap-entry’ and demonstrate that it is still the same
function (unless I misstepped).  (I donʼt know if this will end up all
that readable and useful; it seemed like a good idea when I started and
might end up as time better spent learning ERT.)