`B t' needs to show the rule that matched

`B t' needs to show the rule that matched

[Karl Kleinpaste]

While I appreciate having `B t' (gnus-summary-respool-trace) for
testing changes to nnmail-split-methods, what I really need is for
that command to tell me which rule matched to produce the intended
destination group, much as `V t' (gnus-score-find-trace) tells me what
rule in which scorefile induced what scoring modification.  I'd like
to see this when I find a message that was split erroneously, so I can
modify the split methods to get it right.

How hard would it be to convince `B t' to announce the split rule?

Re: `B t' needs to show the rule that matched

[Lars Magne Ingebrigtsen]

Karl Kleinpaste <karl@charcoal.com> writes:

> While I appreciate having `B t' (gnus-summary-respool-trace) for
> testing changes to nnmail-split-methods, what I really need is for
> that command to tell me which rule matched to produce the intended
> destination group, much as `V t' (gnus-score-find-trace) tells me what
> rule in which scorefile induced what scoring modification.  I'd like
> to see this when I find a message that was split erroneously, so I can
> modify the split methods to get it right.
>
> How hard would it be to convince `B t' to announce the split rule?

I've had a look over the code (look for `nnmail-split-trace' in
nnmail.el), and it doesn't look trivial.  For advanced splits, for
instance, basically everything matches.  It's all one big rule.

(setq
 nnmail-split-fancy
 '(| ("sender" "4ad recording artists list" "4ad")
      ...))

Of course, one could be creative and say that the top-level `|'
"doesn't count", but there might not be a `|' at the top.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen

Re: `B t' needs to show the rule that matched

[Kai Großjohann]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> Of course, one could be creative and say that the top-level `|'
> "doesn't count", but there might not be a `|' at the top.

How about always printing the bottom-most split rule?  Well, err.
The leaves are just group names, these are probably not so
interesting.  So you'd print the bottom-most non-leaf.

kai
-- 
Simplification good!  Oversimplification bad!  (Larry Wall)

Re: `B t' needs to show the rule that matched

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE (Kai Großjohann) writes:

> How about always printing the bottom-most split rule?  Well, err.
> The leaves are just group names, these are probably not so
> interesting.  So you'd print the bottom-most non-leaf.

I've now done something similar to that, because I didn't want to
maintain a stack.  :-)  Seems to give better output than before.

-- 
(domestic pets only, the antidote for overdose, milk.)
   larsi@gnus.org * Lars Magne Ingebrigtsen