bug#10308: 24.0.92; regression: *Help* ruined for toggling read-only

bug#10308: 24.0.92; regression: *Help* ruined for toggling read-only

[Drew Adams]

In all Emacs releases prior to Emacs 24 (and even in 24 up until the
middle of 2011), you could toggle buffer `*Help*' and edit text there
relatively normally.  Only a very few of the ordinary keyboard keys,
such as `RET', had any special meaning.  In particular, typically
self-inserting keys inserted, and deleting keys deleted.
 
Turning off read-only in *Help* has long been useful for composing
and testing *Help* displays (e.g. doc strings).  I, and I know
others as well, have used this feature for decades.  This
regression removes it.

Now, the typically self-inserting keys do nothing - no message, nada.
`C-h k' for a typical key such as `f' says it is undefined.  Why?
What good is that?  Isn't it enough that *Help* is by default
read-only?  Why fiddle with the key bindings like that?

And `DEL' (Backspace) gives the ridiculous error message `Beginning
of buffer' (huh?), no matter where point is when you hit it.  Yet
`C-d', `M-d', etc. still delete text normally.  And `M-q' still
works (thank goodness).

And `q' quits, and `z' kills the buffer , even when the buffer has
been toggled to writable!  This is crazy.  The help-mode keys
should have no such effects when the buffer is made writable.  Try
Emacs 23, to see the difference - when read-only, `q' quits; when
writable, it self-inserts - useful.

What is the point in a user making the buffer writable now?  What
can you do with it writable that you could not do with it read-only?
In effect, we've lost a useful writable help mode, since it is now
nearly the same as read-only *Help* (only a few keys such as `C-d'
are exceptionally still usable to edit).

Now, to be able to edit *Help* you need to change its mode, instead
of just hitting `C-x C-q'.  And if you are iteratively fine-tuning
doc for *Help*, then each time you refresh it from developing code,
you need to change the mode again before you can edit it.

Please revert this regression.  It is inconsistent (e.g. `DEL' vs
`C-d'), makes Emacs less useful, and serves no constructive purpose
(AFAICT - what can you do with a writable *Help* now?).  It is an
incompatible change affecting users, yet it is not even mentioned
in the NEWS.  AFAIK, it was not proposed for discussion on
emacs-devel either.

*Help* is not just a passive output message - it is a useful Emacs 
buffer.  And part of that use is editing its content.  This is
Emacs.  Let's not put obstacles in the way of users trying to
interact with the most flexible, self-documenting, customizable
editor.  If intentional, this change is misguided.
 

In GNU Emacs 24.0.92.1 (i386-mingw-nt5.1.2600) of 2011-12-06 on MARVIN
 Windowing system distributor `Microsoft Corp.', version 5.1.2600
 configured using `configure --with-gcc (4.6) --no-opt --cflags
 -ID:/devel/emacs/libs/libXpm-3.5.8/include
 -ID:/devel/emacs/libs/libXpm-3.5.8/src
 -ID:/devel/emacs/libs/libpng-dev_1.4.3-1/include
 -ID:/devel/emacs/libs/zlib-dev_1.2.5-2/include
 -ID:/devel/emacs/libs/giflib-4.1.4-1/include
 -ID:/devel/emacs/libs/jpeg-6b-4/include
 -ID:/devel/emacs/libs/tiff-3.8.2-1/include
 -ID:/devel/emacs/libs/gnutls-2.10.1/include --ldflags
 -LD:/devel/emacs/libs/gnutls-2.10.1/lib'
 

bug#10308: 24.0.92; regression: *Help* ruined for toggling read-only

[Chong Yidong]

"Drew Adams" <drew.adams@oracle.com> writes:

> In all Emacs releases prior to Emacs 24 (and even in 24 up until the
> middle of 2011), you could toggle buffer `*Help*' and edit text there
> relatively normally.  Only a very few of the ordinary keyboard keys,
> such as `RET', had any special meaning.  In particular, typically
> self-inserting keys inserted, and deleting keys deleted.

If you want to edit the buffer, switch to a major mode suitable for
editing buffers.  Closing as wontfix.

bug#10308: 24.0.92; regression: *Help* ruined for toggling read-only

[Drew Adams]

> If you want to edit the buffer, switch to a major mode suitable for
> editing buffers.  Closing as wontfix.

There is no mode so suitable for editing *Help* as its own mode.

Changing the mode is not a reasonable workaround, as I reported:

>> Now, to be able to edit *Help* you need to change its mode,
>> instead of just hitting `C-x C-q'.  And if you are
>> iteratively fine-tuning doc for *Help*, then each time you
>> refresh it from developing code, you need to change the mode
>> again before you can edit it.

I get the impression that you have never used the capability of making *Help*
writable to develop a *Help* screen interactively & iteratively.  Otherwise,
this should be a no-brainer.

But even if you do not use this feature personally, surely you can understand
that others might.  And surely you can understand that what you have replaced it
with is useless to _everyone_.  Can you honestly make any argument in favor of
the new behavior when writable?  Is there any use case for it?  It is not even
consistent (e.g. DEL vs C-d).

It _appears_ that you did not read the bug report, or did not read it well.
Please (re-)read it.

When toggled to writable, you keep key bindings that make no sense in that
context.  You do not restore the normal keys, which do make sense.  There is now
no point in making the buffer writable.  You have taken away that functionality
for no reason - nothing useful or usable replaces it when the buffer is
writable.

And it is an incompatible change, but it is not even mentioned in NEWS.  How can
you in good conscience think there is no bug here?