1152843 - Newish messages, marked as read on another device, are not processed by the bayes junk filter

Status: NEW

(MailNews Core :: Filters, defect)

Description

[Kent James (:rkent)]

Since I frequently do a preliminary scan of messages on my phone, sometimes I'll read a spam message. After that, the junk processing does not attempt to classify the message. This is invisible to most users, but very clear if you use JunQuilla which shows the junk processing status of each message.

Not sure if processing read messages for junk should be the silent choice, or if we need an option for this.

Comment 1

[Wayne Mery (:wsmwk)]

I'm thinking maybe yes, with phones having become so prolific. However, how far back in time should it check "already read" messages - if that's even possible to control?

Comment 2

[Kent James (:rkent)]

I believe it is possible to distinguish new messages from existing messages, even if they are read, but that would be a concern. Normal filters may also have this issue.

Comment 3

[Wayne Mery (:wsmwk)]

I assume bug 308148 will also help this bug

Comment 4

[Wayne Mery (:wsmwk)]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #3)
> I assume bug 308148 will also help this bug

Are we good now that bug 308148 is fixed?

Comment 5

[Kent James (:rkent)]

(In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #4)
> (In reply to Wayne Mery (:wsmwk, use Needinfo for questions) from comment #3)
> > I assume bug 308148 will also help this bug
> 
> Are we good now that bug 308148 is fixed?

No, I don't think so. The decision to do Bayes filtering on a message is independent of the decision to do Imap filtering. I would have to investigate the issues.