emacs-diffs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Emacs-diffs] /srv/bzr/emacs/trunk r112107: * frame.h (struct frame)


From: Eli Zaretskii
Subject: Re: [Emacs-diffs] /srv/bzr/emacs/trunk r112107: * frame.h (struct frame): Put menu_bar_window under #ifdef
Date: Sat, 23 Mar 2013 08:27:45 +0200

> Date: Sat, 23 Mar 2013 06:11:41 +0400
> From: Dmitry Antipov <address@hidden>
> CC: address@hidden
> 
> > For that matter, why ifdef this away at all?  On a TTY,
> > menu_bar_window is not a window, so the call to clear_glyph_matrix is
> > never made anyway.
> 
> See above: the only configuration where menu_bar_window is really used
> is --with-x --with-x-toolkit=none, and there are no reasons to have
> "constant nil" object in struct frame for other configurations.

I find such "optimizations" detrimental to maintainability.  The gains
of removing a single 32-bit or 64-bit member of a large structure are
nil and void, while the burden of remembering that the member doesn't
exist in some builds is significant, and makes some jobs harder than
they were before.  (E.g., how do you display this window in a .gdbinit
command?)  It also makes reading the code harder.

I thought the tendency in Emacs development was to _reduce_ the amount
of #ifdef's, not to add new ones.  This change goes against that
tendency.

So my vote is firmly for reverting this change.  Stefan?



reply via email to

[Prev in Thread] Current Thread [Next in Thread]