On Sun, 2012-11-25 at 19:41:12 -0800, Russ Allbery wrote: > Henrique de Moraes Holschuh <h...@debian.org> writes: > > Well, the software to do it is around for more than 15 years. Google > > for "procmail duplicate suppression". > > This works exactly backwards of how useful duplicate suppression would > actually work.
> The personal copy is useless; the mailing list copy will get filed into > the proper folder and is the one that you want to keep. Well, obviously that depends on how one handles mail, but I can see how getting a copy when you don't read a specific mailing list very often, serves at least to draw attention that there's mail on the list, which can always be removed afterwards if it's unneeded. > The duplicate suppression that you want is to get rid of the personal copy > and keep the list copy, but that's more complex to do right, because you > have to essentially quarantine the personal copy while you wait for the > list copy that's supposed to replace it, and then deliver the personal > copy if the list copy never arrives. You certainly have to go to more > effort than just mainining a database of message IDs and throwing away the > message the second time you see it. That's one of the reasons I stopped removing/stashing dupes long time ago, the other being, the annoyance of having broken/partial threads on multiple mail boxes. I switched to just marking and showing them “distinctively”, so that I can know immediately if there's other instances around, and can choose where to read them from. As in: ,--- procmailrc --- # Check for duped mails :0 Whc : Admin/msgid.lock | formail -D 524288 Admin/msgid.cache # If it's a dupe, mark it :0 aBfh | formail -a "X-Duped: yes" `--- ,--- muttrc --- color index blue black "~h '^X-Duped: yes'" `--- In any case, I don't mind much adapting to either mailing list usage (w/ or w/o explicit CC), but I've increasingly been finding that the no-CC policy is not w/o fault, being inconsistent (because you don't know off-hand who's subscribed, so on initial mails you might need to CC people directly), prone to missing the recipient (because this is not a usual convention, and people not subscribed might mail a list and expect being CCed, but doing so might incur being chastized in the name of the CoC!), requires being more attentive in case you need to CC when people request it explicitly (and the danger of being chastized for not doing so), and slighty annoying when people complain due to being accidentally CCed or not CCed at all. Thanks, Guillem -- To UNSUBSCRIBE, email to debian-devel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/20121126090759.ga12...@gaara.hadrons.org
