bug#3907: bad result for insert-kbd-macro

[Stefan Monnier]

> I created a macro and called insert-kbd-macro on it.  This is the
> result:

> (fset 'allegati
>    [?e ?\C-[ ?< ?\C-s ?t ?e ?x ?t ?/ ?h ?t ?m ?l ?\C-m ?\C-[ ?{ ?\C-[ ?{ 
> ?\C-[ ?} ?\C-x ?q ?\C-[ ?x ?k ?e ?e ?p ?\C-m ?f ?i ?l ?e ?n ?a ?m ?e ?= ?\C-m 
> ?\C-m ?\C-m ?\C-[ ?% ?. ?* ?f ?i ?l ?e ?n ?a ?m ?e ?= ?\" ?\\ ?( ?[ ?^ ?\" ?] 
> ?+ ?\\ ?) ?\" ?\C-m ?  ?[ ?  ?\\ ?1 ?  ?] ?\C-m ?! ?\C-m ?\C-m ?\C-m ?\C-x ?q 
> ?\C-c ?\C-c ?o ?\C-[ ?\C-? ?i ?s ?i ?\C-m ?\C-[ ?p ?  ?d ?\C-[ ?> up ?\C-c 
> ?\C-c ?\C-m])

> 1) first, if I go to the end of the macro and hit C-x C-e, it appears
>    that the sexp is not correctly evaluated, as the parser sees
>    unbalanced parentheses

This is a bug in the emacs-lisp-mode, mostly.  But indeed, the printer
should also be more careful to escape the chars that cause problem for
the emacs-lisp-mode (i.e. print ?\[ rather than ?[ in your above
example).
This said, IIUC the printer used in Emacs-23 does pay attention this
problem and insert-kbd-macro gave me (for example):

   (setq last-kbd-macro
      [?a left ?\[])

so the problem seems to be fixed in Emacs-23.  Can you confirm?

> 2) second, insert-kbd-macro used to write the strings between double
>    quotes, which is generally more readable and allows for easy editing;
>    I do not know wht it used the vector style here

There's an `up' symbol in there, which couldn't fit in a string.
In general an event is not a char any more (it used to be, back in the
tty days when bytes and chars were the same thing), so a sequence of
events should preferably always be represented as an array of events,
rather than as a string (although, insert-kbd-macro still appears to use
a string when possible).


        Stefan