Undocumented breaking change in minibuffer behavior

Undocumented breaking change in minibuffer behavior

[Vitalie Spinu]

Hi everyone, 

I noticed that emacs 24 ignores the global (t) value of truncate-lines
(as a result of this fix
http://comments.gmane.org/gmane.emacs.diffs/111667 ). This change is not
documented in the news.

I've been truncating lines for years now, and the continuous flickering
in emacs 24 is really disturbing (especially on laptops with small
display). I am a heavy user of IDO which would always re-size the
minibuffer to two lines. The same problem with long eldoc lines.

I wonder if there should be an option like minibuffer-truncate-lines to
undo the  "fix"? I've never been bothered by the reported bug.

Or may be there should be a way to set the minibuffer height to two
lines? I see `max-mini-window-height', but not
`min-mini-window-height'. Would it be meaningful to implement such a
variable?

Thanks, 
Vitalie.

Re: Undocumented breaking change in minibuffer behavior

[Johan Bockgård]

Vitalie Spinu <spinuvit@gmail.com> writes:

> I noticed that emacs 24 ignores the global (t) value of truncate-lines
> (as a result of this fix
> http://comments.gmane.org/gmane.emacs.diffs/111667 ). This change is not
> documented in the news.
>
> I've been truncating lines for years now, and the continuous flickering
> in emacs 24 is really disturbing (especially on laptops with small
> display). I am a heavy user of IDO which would always re-size the
> minibuffer to two lines. The same problem with long eldoc lines.
>
> I wonder if there should be an option like minibuffer-truncate-lines to
> undo the  "fix"?

  ;; Emulate Emacs 23
  (add-hook 'minibuffer-setup-hook
            (lambda ()
              (setq truncate-lines (default-value 'truncate-lines))))

Re: Undocumented breaking change in minibuffer behavior

[Glenn Morris]

Bug fixes are usually not NEWSworthy, and it never occurred to me that
anyone would actually want the previous behaviour.

It would be easy to change it so that eg truncate-lines = 'nomini
means "truncate lines unless in a minibuffer".

Re: Undocumented breaking change in minibuffer behavior

[Vitalie Spinu]

>>>> Glenn Morris <rgm@gnu.org>
>>>> on Wed, 14 Mar 2012 23:55:30 -0400 wrote:

  > Bug fixes are usually not NEWSworthy, and it never occurred to me that
  > anyone would actually want the previous behaviour.

I expect there are plenty of people who are setting truncate-lines
globally to t, and who will spot the change. So a line in a news
probably won't harm.

  > It would be easy to change it so that eg truncate-lines = 'nomini
  > means "truncate lines unless in a minibuffer".

I realized that I would also prefer the new default with the exception
of IDO completion. So I modified Johan's advice to work for IDO only:

(add-hook 'ido-minibuffer-setup-hook
	  (lambda ()
	    (setq truncate-lines (default-value 'truncate-lines))))

Thanks, 
Vitalie.