bug#5774: 23.1.92; Emacs manual, node Persistent Marks

bug#5774: 23.1.92; Emacs manual, node Persistent Marks

[Drew Adams]

emacs -Q

In GNU Emacs 23.1.92.1 (i386-mingw-nt5.1.2600)
 of 2010-02-20 on LENNART-69DE564
Windowing system distributor `Microsoft Corp.', version 5.1.2600
configured using `configure --with-gcc (3.4) --no-opt --cflags -Ic:/g/include
-fno-crossjumping'

1. The node says:

  "When Transient Mark mode is off, the mark is _never_ deactivated..."

This is wrong. And it confuses users. Check `C-h v mark-active' with t-m-mode
off and see if what it says makes sense in combination with this Info doc (no).

The truth is that when t-m-mode is off there is no notion of an active mark.
That is what should be said (and it was said like that in the past). The notion
of an active mark and region is a t-m-mode notion.

2. Replace undo (`C-x u') in the example given by a more typical command wrt
t-m-mode (e.g. query-replace). Things are said in this paragraph that are not
general but are specific to `undo', and that's not the message we are trying to
get across here. (You typically cannot use `C-u' to make such commands operate
on the region when the region is inactive.)

3. "`C-u C-x C-x'  Activate the mark without changing it". Remove "without
changing it". That doesn't add anything except confusion. All the `C-u' does is
activate the mark, period. `C-u C-x C-x' still swaps point and mark ("changing"
the mark).

bug#5774: 23.1.92; Emacs manual, node Persistent Marks

[Chong Yidong]

> "When Transient Mark mode is off, the mark is _never_ deactivated..."
>
> This is wrong.  And it confuses users. Check `C-h v mark-active' with
> t-m-mode off and see if what it says makes sense in combination with
> this Info doc (no).

It is correct.  If tmm is off, the mark starts out inactive, and the
variable `mark-active' is nil.  Once you type C-SPC, `mark-active'
becomes t, and remains t regardless of (almost) anything you do.

> 2. Replace undo (`C-x u') in the example given by a more typical command
> wrt t-m-mode (e.g. query-replace).
>
> 3. "`C-u C-x C-x'  Activate the mark without changing it". Remove "without
> changing it". That doesn't add anything except confusion.

I have edited the text to remove the possible confusion.

bug#5774: 23.1.92; Emacs manual, node Persistent Marks

[Drew Adams]

> > "When Transient Mark mode is off, the mark is _never_ 
> > deactivated..."
> >
> > This is wrong.  And it confuses users. Check `C-h v 
> > mark-active' with t-m-mode off and see if what it says
> > makes sense in combination with this Info doc (no).
> 
> It is correct.  If tmm is off, the mark starts out inactive, and the
> variable `mark-active' is nil.  Once you type C-SPC, `mark-active'
> becomes t, and remains t regardless of (almost) anything you do.

`mark-active' is an internal variable that is irrelevant to this doc and a
conceptual understanding. What is relevant for the user's conceptual model is
observable behavior, not implementation.

It makes no difference what the value of `mark-active' is when t-m-mode is off.
Your code can set it to anything at all, including nil, and that will change
nothing in terms of observable behavior.

What is important to get across here is that activeness of the mark has meaning
_only_ for transient-mark-mode and similar, derived modes. (That includes
temporary t-m-mode.)

Inactive/active affects behavior only for t-m-mode and the like. When t-m-mode
is off there is _no recognition_ of the mark states active/inactive - no
observable difference or change in behavior. And that's as it should be.

> > 2. Replace undo (`C-x u') in the example given by a more 
> > typical command wrt t-m-mode (e.g. query-replace).
> >
> > 3. "`C-u C-x C-x'  Activate the mark without changing it". 
> > Remove "without changing it". That doesn't add anything
> > except confusion.

4. This node should be renamed. "Persistent Marks" implies persistence, which
typically means saving to disk and across sessions. Some 3rd-party Emacs
libraries do provide for persistent marks, but that's not what this is about.
Please rename the node - perhaps something like `Transient Mark Mode Off' would
be a much clearer indicator of the content.

bug#5774: 23.1.92; Emacs manual, node Persistent Marks

[Lars Ingebrigtsen]

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

> 1. The node says:
>
>   "When Transient Mark mode is off, the mark is _never_ deactivated..."
>
> This is wrong. And it confuses users. Check `C-h v mark-active' with t-m-mode
> off and see if what it says makes sense in combination with this Info doc (no).

[...]

> `mark-active' is an internal variable that is irrelevant to this doc and a
> conceptual understanding. What is relevant for the user's conceptual model is
> observable behavior, not implementation.
>
> It makes no difference what the value of `mark-active' is when t-m-mode is off.
> Your code can set it to anything at all, including nil, and that will change
> nothing in terms of observable behavior.

True, but I can't find anything in the manual that states this in this
way any more after grepping for a minute.  Closing, but please reopen if
this problem is still in the manual.

-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#5774: 23.1.92; Emacs manual, node Persistent Marks

[Drew Adams]

> True, but I can't find anything in the manual that states this in
> this way any more after grepping for a minute.  Closing, but please
> reopen if this problem is still in the manual.

Yes, thanks.  I believe this was fixed.  There is no longer
a `Persistent Marks' node.