That looks decent. Could you add unit tests for this method? Even if they need to be wrapped in a feature detect for element.files, that would still be a big improvement over no coverage at all.

I tried to create test for it, but I'm not sure how to mock the file input functionality, since JS cannot update the value if the file input...

Do you have an idea how to test it?

You could start with an input of type file, select that from the DOM, then overwrite the files property. Not sure if its writable, but if it is, this might work (not tested):

<input type="file" id="accept-input">

var element = $( "#accept-input" )[0];
// Assign something here for each assertion, not sure what the right format is
element.files = [ "myfile.docx" ];
ok( $.validator.method.accept( null, element, "docx" ), "something something is valid" );

I have no idea what the right format for files and param are, but I hope this helps. Thanks for looking into it!

Hmmm, the input property files is readonly for security reasons.

Makes sense. Maybe we can mock the DOM element from scratch? Along with the files property it needs to be pass the this.optional() check and respond to $(element).attr("type") with "file".

I had the same idea.. I will try it :)

I've just created the tests, the last one is for invalid mime type. I'm not sure if the last one check for dependency-mismatch is ok...

Thanks for the update! This is promising, but needs a bit of refactoring and cleanup.

Yes, you are right, this is raw version, I will clean it up.

Never noticed the new commits until now - there are no notifications for pushes... reviewing again.

Still a few minor issues with the tests. Please let me know when you push new commits, so that I can review them immediately.

ping @jzaefferer

Hope thats more clear. I'll be away the next few days, will check again next week.

Sorry for the delay, I had urgent work to do.

ping @jzaefferer

@jzaefferer did you have a chance to see this?

Sorry for the delay. Could you rebase this against master? Thanks.

@mishal As @jzaefferer requested in #1512, I have squashed/rebased this PR into #1531. Authorship and the original diff should be maintained but now in 1 commit and on top of the latest master.

Thanks, closing, please merge #1531