On Sun, Dec 09, 2001 at 06:25:03PM -0500, David T-G
<[EMAIL PROTECTED]> wrote:
> Daniel --
>
> ...and then Daniel Eisenbud said...
> %
> % On Sun, Dec 09, 2001 at 02:50:44PM -0500, David T-G
><[EMAIL PROTECTED]> wrote:
> % > % subject. The question mark denotes a missing reference. So if a
> % >
> % > This I also understand -- but it seems to be too over-the-top. I've been
> ...
> %
> % This is what $hide_missing is for. Set it. But for getting rid of long
>
> But it was -- in all of the displays I pasted in! For some threads in a
> large mailbox I could see a difference with $hide_missing (oh how I
> am starting to wish for vi-like shortcut settings :-) but for many
> I couldn't and only for Nicolas' message (msg 10) in this test case.
Right, but the specific scenario you described in that paragraph
(keeping just one message from late in a thread) would be taken care of
by $hide_missing.
> ...
> % > and wonder why the second thread isn't collapsed and if it has something
> % > to do with the missing reference at the top of the thread. If I open
> % > that thread in 1.3.23 I see
> %
> % I'll look into the thread-collapsing thing.
>
> Thanks. BTW, I can't even collapse it manually, and if I'm not on the
> first message (number 3) then my cursor jumps there when I try.
This is consistent with my hypothesis of what was wrong. The
message-collapsing code is utter crap, IMHO, and I didn't have the
energy to rip it all out and rewrite it from scratch before, so I tried
to just change what I needed to change, and I obviously missed this
detail. I bet it will be an easy fix (and I still do mean to rewrite
the whole thing sometime.)
> ...
> % > subject (though Prahlad's message, interestingly enough, did not appear
> % > to have any missing refs according to 1.3.24 above), the threading looks
> % > manageable and appears in the same order.
> %
> % Prahlad's message does have a missing parent in the new display: it's
> % the missing parent that's at the top of the thread.
>
> Ahhh... So it was attached by subject because its parent is so early
> that it's not anywhere else.
>
> No, wait a minute... It should have the same parent reference as the
> first message, right? In the 1.3.24 picture there's no question mark --
> even when I turn off $hide_missing.
The question mark at the top of the thread is the parent of his message
in 1.3.24. Really. Go look at the thread tree again -- his message is
a sibling of the top message in the thread, which also has that question
mark as its parent.
> % [...]
> % > I'd prefer to have $hide_missing hide *all* of the missing reference
> % > indicators and give me a display like in 1.3.23 -- all of the clips above
> % > are with $hide_missing set!
> %
> % Hiding all the missing messages will make the thread display less
> % comprehensible, since things that aren't siblings will be grouped
> % together in a way that looks like they're siblings, but they won't be
>
> Hmmm... I suppose I can get the idea of how this would cause problems,
> but I can't yet see an example. But, as I see below, perhaps I won't
> have to worry about that...
For instance,
1 foo
2 |-?->bar
3 | `->baz
4 |->grault
5 `->quux
if $sort_aux is set to date, it is possible that bar is before grault
and quux, but baz is after both of them. So with the question mark
removed, the thread tree would look like
1 foo
2 |->bar
3 |->baz
4 |->grault
5 `->quux
and baz would appear to be in the wrong place. Additionally,
next-subthread and previous-subthread won't appear to work quite right,
though some of that may be that they haven't been quite properly adapted
to the new threading code -- I'll think a bit more about what they
should do. Part of the problem with these commands is that it's not
entirely clear what the right thing for them to do is, even under
simpler circumstances.
-Daniel
--
Daniel E. Eisenbud
[EMAIL PROTECTED]
"We should go forth on the shortest walk perchance, in the spirit of
undying adventure, never to return,--prepared to send back our embalmed
hearts only as relics to our desolate kingdoms."
--Henry David Thoreau, "Walking"
