bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Stephen Berman]

1. emacs -Q
2. Carry out the steps in NEWS to return to the pre-24 selection behavior:
   Change `mouse-drag-copy-region' to t.
   Change `x-select-enable-primary' to t.
   Change `x-select-enable-clipboard' to nil.
3. M-x gnus, answer y to continue to the Group buffer despite no server
   being found, then type `B f' to browse a foreign server,
   e.g. news.gmane.org, select a group and open an article.
4. In the Article buffer select some text (it doesn't matter whether by
   double clicking with mouse-1 or by the keyboard, though if the latter
   then after selecting point should at the end of the selection to have
   the same situation as when selecting by the mouse).
5. Type `M-w'.
6. Switch to a writable buffer, and type `C-y'.

=> The yanked text is not the selection from step 4, but -- if the
selection was made from text following the article headers -- a region
of the Article buffer ending at the start of the selection and beginning
either just below the last header line, or possibly including one or
more header lines; if the selection is within the headers, the yanked
text is the region beginning at the start of the selection and ending
just below the headers.

If you type `M-y' after step 6, then the yanked text is replaced by the
text selected in step 4.  If step 5 is omitted, C-y in step 6 yanks the
selected text as usual.

I have only observed this selection behavior in Gnus Article buffers,
and there it is reliably reproducible.  (I have observed it for some
time, probably since the new selection changes were introduced, but
didn't try to find a reproducible recipe till now.)

In GNU Emacs 24.0.50.1 (i686-suse-linux-gnu, GTK+ Version 2.20.1)
 of 2011-03-30 on escher
Windowing system distributor `The X.Org Foundation', version 11.0.10800000
configured using `configure  '--without-toolkit-scroll-bars' 'CFLAGS=-g -O2 -fno-optimize-sibling-calls''

Important settings:
  value of $LC_ALL: nil
  value of $LC_COLLATE: nil
  value of $LC_CTYPE: nil
  value of $LC_MESSAGES: nil
  value of $LC_MONETARY: nil
  value of $LC_NUMERIC: nil
  value of $LC_TIME: nil
  value of $LANG: en_US.UTF-8
  value of $XMODIFIERS: @im=local
  locale-coding-system: utf-8-unix
  default enable-multibyte-characters: t

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[David De La Harpe Golden]

On 31/03/11 23:05, Stephen Berman wrote:
> 1. emacs -Q
> 2. Carry out the steps in NEWS to return to the pre-24 selection behavior:
>     Change `mouse-drag-copy-region' to t.
>     Change `x-select-enable-primary' to t.
>     Change `x-select-enable-clipboard' to nil.

Please try:

(setq select-active-regions nil
       mouse-drag-copy-region t
       x-select-enable-primary t
       x-select-enable-clipboard nil)
(global-set-key [mouse-2] 'mouse-yank-at-click)

I'm not saying there isn't a real issue, what you describe does sound a 
bit similar to a problem that occasionally occurred with the new 
settings (without any changes to them) a while back, but please try with 
the above settings, which are AFAIK still* (whatever the NEWS file may 
currently say) the actual current recipe to restore the old behaviour 
(except on windows) - if you only did precisely what you said in your 
2., then AFAIK you were running with a doom-laden mix of old and new 
settings.

* bearing in mind I'm only beginning a personal catchup on about 3½ 
months of emacs developments.

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Stephen Berman]

On Fri, 01 Apr 2011 01:44:13 +0100 David De La Harpe Golden <david@harpegolden.net> wrote:

> On 31/03/11 23:05, Stephen Berman wrote:
>> 1. emacs -Q
>> 2. Carry out the steps in NEWS to return to the pre-24 selection behavior:
>>     Change `mouse-drag-copy-region' to t.
>>     Change `x-select-enable-primary' to t.
>>     Change `x-select-enable-clipboard' to nil.
>
> Please try:
>
> (setq select-active-regions nil
>       mouse-drag-copy-region t
>       x-select-enable-primary t
>       x-select-enable-clipboard nil)
> (global-set-key [mouse-2] 'mouse-yank-at-click)
>
> I'm not saying there isn't a real issue, what you describe does sound a bit
> similar to a problem that occasionally occurred with the new settings (without
> any changes to them) a while back, but please try with the above settings,
> which are AFAIK still* (whatever the NEWS file may currently say) the actual
> current recipe to restore the old behaviour (except on windows) - if you only
> did precisely what you said in your 2., then AFAIK you were running with a
> doom-laden mix of old and new settings.

You're right.  So this is a -- rather insidious -- NEWS bug; fix below.
Thanks for setting me straight.

Steve Berman

*** /data/steve/bzr/emacs/trunk/etc/NEWS	2011-03-30 10:33:51.000000000 +0200
--- /data/steve/bzr/emacs/quickfixes/etc/NEWS	2011-04-01 10:38:09.000000000 +0200
***************
*** 332,337 ****
--- 332,338 ----
  
  *** To return to the previous behavior, do the following:
  
+ **** Change `select-active-regions' to nil.
  **** Change `mouse-drag-copy-region' to t.
  **** Change `x-select-enable-primary' to t (on X only).
  **** Change `x-select-enable-clipboard' to nil.

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Chong Yidong]

Stephen Berman <stephen.berman@gmx.net> writes:

> You're right.  So this is a -- rather insidious -- NEWS bug; fix below.

Applied; thanks.

Re: bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Lars Magne Ingebrigtsen]

Stephen Berman <stephen.berman@gmx.net> writes:

> 4. In the Article buffer select some text (it doesn't matter whether by
>    double clicking with mouse-1 or by the keyboard, though if the latter
>    then after selecting point should at the end of the selection to have
>    the same situation as when selecting by the mouse).
> 5. Type `M-w'.
> 6. Switch to a writable buffer, and type `C-y'.

The problem is tickled by the continuously updated text in the buffer,
but I have no idea why it's happening.

If I have `transient-mark-mode' switched on, and you `M-w' a line from
the article buffer, I get the following:

(car kill-ring) =>
"diff --git a/lisp/ChangeLog b/lisp/ChangeLog
"

Which is correct.

(x-selection-value-internal 'CLIPBOARD) =>
"diff --git a/lisp/ChangeLog b/lisp/ChangeLog
"

Which is also correct.

(x-selection-value-internal 'PRIMARY) =>
"
Signed-off-by: Kan-Ru Chen <kanru@kanru.info>
---
 lisp/ChangeLog |    4 ++++
 lisp/nnir.el   |    1 +
 2 files changed, 5 insertions(+), 0 deletions(-)

"

Which is WTF?!

The latter wins, so when you yank the text, you get that instead.  It's
not random what you get, it but it's not what you killed.

Does anybody who knows more about the Emacs kill'n'yank internals know
what this could mean?

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

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Eli Zaretskii]

> From: Lars Magne Ingebrigtsen <lmi@gnus.org>
> Date: Sun, 03 Apr 2011 19:37:16 +0200
> 
> If I have `transient-mark-mode' switched on, and you `M-w' a line from
> the article buffer, I get the following:
> 
> (car kill-ring) =>
> "diff --git a/lisp/ChangeLog b/lisp/ChangeLog
> "
> 
> Which is correct.
> 
> (x-selection-value-internal 'CLIPBOARD) =>
> "diff --git a/lisp/ChangeLog b/lisp/ChangeLog
> "
> 
> Which is also correct.
> 
> (x-selection-value-internal 'PRIMARY) =>
> "
> Signed-off-by: Kan-Ru Chen <kanru@kanru.info>
> ---
>  lisp/ChangeLog |    4 ++++
>  lisp/nnir.el   |    1 +
>  2 files changed, 5 insertions(+), 0 deletions(-)
> 
> "
> 
> Which is WTF?!

M-w and other keyboard gestures don't set PRIMARY in Emacs 24.

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Lars Magne Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> M-w and other keyboard gestures don't set PRIMARY in Emacs 24.

But it's the contents from PRIMARY that ends up in the buffer when you
say `C-y', apparently.

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

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Lars Magne Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> M-w and other keyboard gestures don't set PRIMARY in Emacs 24.

But like David De La Harpe Golden said, this may just be because of a
bad combination of the

(setq select-active-regions nil
      mouse-drag-copy-region t
      x-select-enable-primary t
      x-select-enable-clipboard nil)

options.  It does seem to affect quite a few people, though, and only in
the article buffer (with the automatically updating Date header).

Doing the settings above makes the problem disappear, but it's
unfortunate that other combination(s) of the options has such a
confusing outcome, I think.
      
-- 
(domestic pets only, the antidote for overdose, milk.)
  bloggy blog http://lars.ingebrigtsen.no/

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Eli Zaretskii]

> From: Lars Magne Ingebrigtsen <larsi@gnus.org>
> Cc: bug-gnu-emacs@gnu.org
> Date: Sun, 03 Apr 2011 20:44:49 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > M-w and other keyboard gestures don't set PRIMARY in Emacs 24.
> 
> But it's the contents from PRIMARY that ends up in the buffer when you
> say `C-y', apparently.

Not by default.  C-y should give you the clipboard or the kill-ring.
Isn't that what you reported?

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Lars Magne Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> Not by default.  C-y should give you the clipboard or the kill-ring.
> Isn't that what you reported?

No, `C-y' gave me PRIMARY, which (with my settings, and has been
reported by other users) was different in the article buffer from
`kill-ring' and CLIPBOARD (and only in the article buffer).

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

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Eli Zaretskii]

> From: Lars Magne Ingebrigtsen <larsi@gnus.org>
> Cc: bug-gnu-emacs@gnu.org
> Date: Sun, 03 Apr 2011 23:44:04 +0200
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > Not by default.  C-y should give you the clipboard or the kill-ring.
> > Isn't that what you reported?
> 
> No, `C-y' gave me PRIMARY, which (with my settings, and has been
> reported by other users) was different in the article buffer from
> `kill-ring' and CLIPBOARD (and only in the article buffer).

So what is the bug? that C-y gave you PRIMARY or that it was different
from the kill ring and the clipboard?

Re: bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Lars Magne Ingebrigtsen]

Eli Zaretskii <eliz@gnu.org> writes:

> So what is the bug? that C-y gave you PRIMARY or that it was different
> from the kill ring and the clipboard?

The bug is that `C-y' gives me a text I didn't `M-w'.  The kill
ring/primary/clipboard thing was just my attempt at debugging the
situation, in the hope that somebody would understand what the cause is.

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

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[Eli Zaretskii]

> From: Lars Magne Ingebrigtsen <lmi@gnus.org>
> Newsgroups: gnu.emacs.bug
> Date: Mon, 04 Apr 2011 11:24:55 +0200
> Mail-Copies-To: never
> 
> Eli Zaretskii <eliz@gnu.org> writes:
> 
> > So what is the bug? that C-y gave you PRIMARY or that it was different
> > from the kill ring and the clipboard?
> 
> The bug is that `C-y' gives me a text I didn't `M-w'.

That's the documented effect of setting x-select-enable-primary to a
non-nil value, AFAIK.  I think we can close this bug, then.

bug#8400: 24.0.50; Strange selection behavior in Gnus Article buffer

[David De La Harpe Golden]

On 04/04/11 12:51, Eli Zaretskii wrote:
>> From: Lars Magne Ingebrigtsen<lmi@gnus.org>
>> Newsgroups: gnu.emacs.bug
>> Date: Mon, 04 Apr 2011 11:24:55 +0200
>> Mail-Copies-To: never
>>
>> Eli Zaretskii<eliz@gnu.org>  writes:
>>
>>> So what is the bug? that C-y gave you PRIMARY or that it was different
>>> from the kill ring and the clipboard?
>>
>> The bug is that `C-y' gives me a text I didn't `M-w'.
>
> That's the documented effect of setting x-select-enable-primary to a
> non-nil value, AFAIK.  I think we can close this bug, then.
>

Yeah. If x-select-enable-primary is non-nil, then as well as C-y getting 
from primary, C-w/M-w also set primary. However, if you've also asked 
for primary to be set by some other means and used those other means, 
then C-y will sometimes quite correctly give you the text from primary 
that got there by means other than M-w.

I still haven't quite convinced myself that the text Stephen Berman was 
getting with his mix of settings was the expected one even given those 
settings. But on the whole, I think the main bug here was the docs bug.