I'm really replying to the ideas in multiple messages here, but I'm too
lazy to do proper quoting. I'm not picking on Brock here.
>>>>> "BR" == Brock Rozen <email@example.com> writes:
BR> In addition (this is strange, replying to my own message) -- if it
BR> could make these two things be considered the "same" subject: (that is,
BR> this is how it should be put on the TODO list -- whenever it's
BR> implemented is a different story)
As usual, what started as a request that I could solve in three lines of
code has ballooned into a huge thing with threading and prefix removal and
all sorts of stuff. I see some folks who _really_ need to learn how to
program; the hooks are all there for you to implement this kind of thing
Sorting by subject is easy, but not taking the oldest messages first can
lead to starvation (although the digest generator will try to push out
digests if the maximum spooled article age that you set is
exceeded). Stripping _just_ /re:\s*/i when sorting is relatively easy, but
it could easily get out of hand handling crap like AW: and SV: perpetrated
by folks who think Re: means something in English and should be localized.
All the information necessary to do proper threading (message ID's from
References: and In-Reply-To:, specifically) is available to the
digestifier. Note that these are unreliable; in mail, References: is only
occasionally used and In-Reply-To: doesn't do what you think. But anyway,
if someone wants to dump me even the algorithm for building a references
tree, I'll be happy to look at it. Yes, it should be easy, but then you
get into worrying about where to cut the tree when it exceeds the digest
size. I really don't want to think too much about this, especially when
the user agent will just properly thread the whole thing when the digest is
burst. At least my software does; I can't imagine people wanting to use
something that doesn't.
If I can find some time, I'll slot in sorting with Re: removal and see how
it goes. Threading is a fun problem but I just don't have the time.