Closed dhpiggott closed 9 years ago
I don't have a preference on these things. Would love to have others weigh in.
Since no one else weighed it, let's change it and see how it goes. Could you write that PR?
(I think Spamassassin is probably smart enough to un-wrap it's own attachments, btw. So I don't think that's a problem.)
Great, I'll do that. I already have the commit in my own repo and have been running with it since I created this issue, so I'll cherry-pick it onto a branch of master shortly and open a PR (the commit is currently based on other changes you won't want hence I'll need to cherry-pick).
I recognise this is to some extent more a matter of opinion than a question of right/wrong but I wanted to just query these:
My concern is that these will make false positives painful to detect and correct. If spam were not marked as seen, detecting it would be as simple as noting a non-zero unread count on the Spam folder, and correcting it would be as simple as moving the false positive(s) out of the Spam folder. With Spam encapsulated as an attachment, to properly correct false positives will require first extracting the encapsulated original and then getting the MUA to add that to an IMAP folder (which no doubt breaks the process of unlearning it as spam too). And I doubt any webmail client supports that - I don't even know how I would do that with Evolution (it's admittedly not an issue I've run into yet).
Changing both of these would be trivial - I could easily make the change and open a PR but as I recognise it's to some extent a matter of opinion I thought it better to just open this for discussion first.