Darren Hoo writes:
> "Eric M. Ludlam" <eric@...> writes:
>> I agree. I'm not sure what the implications are, but it seems like the
>> right solution. Databases like the Emacs Lisp one should still work,
>> since it uses defvar mode local, and the per-buffer ones should work,
>> since they set stuff up per buffer.
>> I think we should do it anyway because the doc says to use
>> setq-mode-local to apply new values, so we should be ok.
>> I'm not sure what would happen if someone uses custom to customize the
>> value, but if someone uses (setq ...) in their .emacs, that setting
>> would stop working.
> user (setq ...) before loading CEDET will do, otherwise it will not
> work as user would expect. So make-buffer-local-variable of a customizable
> variable is a not good idea to me, since different setting of this variable
> may affect performance drastically.
I committed the change to use make-variable-buffer-local. I was
confident that I could use a :set function in combination with
custom-local-buffer to avoid problems with customize, but that was
premature. In fact, I'm not at all sure what custom-local-buffer really
However, custom uses 'set-default' as the default :set function, and if
the user also uses setq-mode-local, it should override the value from
customize. I think this is what the user would expect?
Anyway, revert at will. :-)