Open steelman opened 10 months ago
Hey @steelman , that is a good question, before checking I'd said that it was intentional by design.
If you look at the code here
class ListMailsFilter(HeaderMatchingFilter):
message = 'Tagging mailing list posts'
query = 'NOT tag:lists'
pattern = r"<(?P<list_id>[a-z0-9!#$%&'*+/=?^_`{|}~-]+)\."
header = 'List-Id'
tags = ['+lists', '+lists/{list_id}']
You can see it adds 2 tags (lists, lists/mylist).
I am asking because I just turned it on and found out that I've got message form several lists named dev tagged together as lists/dev, which doesn't make sense.
I have some like that too, in those case if I want to tag them out nicely, I'll use the filter module like so (documentation: configuration and filters) (I think the filter documentation need a good upgrade)
[Filter.37]
message = MyList/dev
query = from:'my-list-dev@list.mylist.net>'
tags = +myilst;+important
According to RFC2919 the part (up to the first
.
) that is used by the filter to create a tag is named list-label. Is there any technical reason (e.g. notmuch limitation) for not using the full list-id i.e. the text between<
and>
?I am asking because I just turned it on and found out that I've got message form several lists named
dev
tagged together aslists/dev
, which doesn't make sense.