Re: define-key #2 ---> ineffective if key is bound...

[Miles Bader <miles@gnu.org>]

"D. Goel" <deego@glue.umd.edu> writes:
> <-- thus, if the user wants to shuffle his O with P, then
> key-translation-map (rather than keyboard-translate-table) is the way to
> go.  In the current version, it seems that there is no way for me to
> shuffle O with P without side-effects, either by using
> keyboard-translate-table or by using key-translation-map, because in
> both cases, the function-key-map's up-arrow conversion (and other
> arrows) gets broken.

I agree with you, mostly because my mental model of the relationship
between these various translation-maps (function-key-map ->
key-translation-map -> normal keymaps) is a pipeline, where later stages
cannot influence earlier ones.

However, in practice, this is not true, except for simple cases, because
function-key-map and key-translation-map work by modifying the tail of a
growing data structure, character by character, and pay attention to more
than one character at the end.

I suppose it could be worked around, by having two copies of the current
input sequence, one which holds the original input events, and which
function-key-map also modifies, which would be copied before
key-translation-map was applies (and presumably would have to be re-run
from either the beginning or some other synchronization point if
function-key-map actually shrunk the input).

But the key-lookup machinery seems very complicated and fragile to me, so
it may not be so easy to make such a change...

-Miles
-- 
`...the Soviet Union was sliding in to an economic collapse so comprehensive
 that in the end its factories produced not goods but bads: finished products
 less valuable than the raw materials they were made from.'  [The Economist]