Re: Incompatible change without warning

Re: Incompatible change without warning

[Andrew M. Scott]

FYI, 

Kyle Jones' VM 7.19 (and prior) mailer has an (unfocus-frame) in its
vm-window.el file, VM didn't like the 2005-05-01 CVS removal of
unfocus-frame.

My local workaround was to comment out the (unfocus-frame) in my copy
of vm-window.el; which may not be an option to the average VM user.

I just sent details to bug-vm@wonderworks.com, though there hasn't
been active VM development since Sept 2004.

Andy Scott

Re: Incompatible change without warning

[Nick Roberts]

 > Kyle Jones' VM 7.19 (and prior) mailer has an (unfocus-frame) in its
 > vm-window.el file, VM didn't like the 2005-05-01 CVS removal of
 > unfocus-frame.

Since unfocus-frame does nothing but return nil, I don't know why it was
needed in the first place

 > My local workaround was to comment out the (unfocus-frame) in my copy
 > of vm-window.el; which may not be an option to the average VM user.

It might be better to replace it with nil since it's at the end of a
function (vm-warp-mouse-to-frame-maybe).

 > I just sent details to bug-vm@wonderworks.com, though there hasn't
 > been active VM development since Sept 2004.

I happen to use VM 7.19 and although it does indeed have unfocus-frame I have
not encountered any problems. VM isn't part of Emacs and doesn't really
respect it. It uses global hooks, tramples over global bindings and generally
does its own thing. Clearly Richard Stallman will decide what to do in this
case but I see it as VM's problem, not Emacs', if you have filed a bug report.

Rmail and MH-E are mail readers which *are* part of Emacs (which I must admit
I haven't used). Clearly hese will be supported by future Emacs releases.

Nick

Re: Incompatible change without warning

[Glenn Morris]

A function which was in the previous release of Emacs is now gone,
without ever being marked as obsolete in any released version of
Emacs. To me, this seems like a Bad Thing. What the function did, or
how external packages chose to use it, seem irrelevant.

Re: Incompatible change without warning

[Richard Stallman]

I guess we should reinstall unfocus-frame just so VM keep sworking.
We should write a comment saying the reason.

Eli, could you do this if no one else does it first?

Re: Incompatible change without warning

[Nick Roberts]

 > A function which was in the previous release of Emacs is now gone,
 > without ever being marked as obsolete in any released version of
 > Emacs. To me, this seems like a Bad Thing. What the function did, or
 > how external packages chose to use it, seem irrelevant.

The entry in subr.el suggested it was obsoleted in Emacs 19.32:

> (make-obsolete 'unfocus-frame "it does nothing." "19.32")

However, looking through the ChangeLogs, it looks like you are right.

On a related note, I see Juanma has obsoleted functions in subr.el without
a timestamp:

(define-obsolete-function-alias 'window-dot 'window-point)
...

I presume that should be:

(define-obsolete-function-alias 'window-dot 'window-point "22.1")
...

to make it easier to know when obsoletion really took place.


Nick

Re: Incompatible change without warning

[Richard Stallman]

    A function which was in the previous release of Emacs is now gone,
    without ever being marked as obsolete in any released version of
    Emacs. To me, this seems like a Bad Thing.

I would tend to agree.  Which function is it?

Re: Incompatible change without warning

[Juanma Barranquero]

> I presume that should be:
> 
> (define-obsolete-function-alias 'window-dot 'window-point "22.1")
> ...
> 
> to make it easier to know when obsoletion really took place.

You're right. I'll fix it (I'm waiting to know whether Richard wants
send-region and send-string to be de-obsoleted).

-- 
                    /L/e/k/t/u

Re: Incompatible change without warning

[Juanma Barranquero]

> The entry in subr.el suggested it was obsoleted in Emacs 19.32:
> 
> > (make-obsolete 'unfocus-frame "it does nothing." "19.32")
> 
> However, looking through the ChangeLogs, it looks like you are right.

I think these old functions where obsoleted a long time ago; but the
byte-compiler warnings and describe-function help machinery to show
them as obsolete is much more recent. So yeah, probably there wasn't a
byte-compiler warning on 21.1 saying `unfocus-frame' was obsolete; but
the function *was* obsolete anyway and suitable replacements had been
introduced (and, I  presume, documented).

-- 
                    /L/e/k/t/u

Re: Incompatible change without warning

[Richard Stallman]

Please put the definition of unfocus-frame  back in.
It should not be removed in this release.

    So yeah, probably there wasn't a
    byte-compiler warning on 21.1 saying `unfocus-frame' was obsolete; but
    the function *was* obsolete anyway and suitable replacements had been
    introduced (and, I  presume, documented).

You mustn't presume such a thing.

The fact is, this function had become a no-op.  There was no
"replacement".  But with nothing to call programmers' attention to the
unnecessary calls, they probably didn't remove them.  I am sure there
are many programs that still call this function.

In this release, it should only get a warning.  We can remove it in
23.1.

Re: Incompatible change without warning

[Juanma Barranquero]

On 5/20/05, Richard Stallman <rms@gnu.org> wrote:

> Please put the definition of unfocus-frame  back in.
> It should not be removed in this release.

OK, no problem.

> You mustn't presume such a thing.

Eh, don't shoot the messenger. I *didn't* remove unfocus-frame, I was
just talking about what I suppose was the frame of mind of the person
who did it :)

-- 
                    /L/e/k/t/u

Re: Incompatible change without warning

[Juanma Barranquero]

> OK, no problem.

But then, what about dot, dot-marker, dot-min, dot-max,
buffer-flush-undo, compiled-function-p and focus-frame, which were
deleted on the same commit? Do I restore them too?

-- 
                    /L/e/k/t/u

Re: Incompatible change without warning

[Richard Stallman]

    > You mustn't presume such a thing.

    Eh, don't shoot the messenger. I *didn't* remove unfocus-frame, I was
    just talking about what I suppose was the frame of mind of the person
    who did it :)

Sorry, it sounded like you were saying why you removed it,
so I thought you had done it.  (My memory doesn't tend to retain
things like who really did remove it.  I am getting old.)

Re: Incompatible change without warning

[Richard Stallman]

    But then, what about dot, dot-marker, dot-min, dot-max,
    buffer-flush-undo, compiled-function-p and focus-frame, which were
    deleted on the same commit? Do I restore them too?

Please restore focus-frame.

The rest were already marked obsolete in Emacs 21, so they can
now be deleted.

Re: Incompatible change without warning

[Juanma Barranquero]

> Please restore focus-frame.

Done (both focus- and unfocus-frame).

-- 
                    /L/e/k/t/u