23.0.50; Middle ``w´´ in of permissions in dired-mode is red and bold: dired-warning

23.0.50; Middle ``w´´ in of permissions in dired-mode is red and bold: dired-warning

[Peter Dyballa]

Hello!

Since an update from CVS this noon (UTC) dired shows in case of files  
or directories like these

   .../emacs:
   total used in directory 9248 available 22772200
   drwxrwxr-x    66 pete  admin    2244 13 Okt 19:25 .
   drwxrwxr-x    20 pete  admin     680 27 Sep 13:15 ..
   -rw-rw-r--     1 pete  admin     210  4 Mär  2005 .arch-inventory
   -rw-rw-r--     1 pete  admin     500 22 Mär  2007 .cvsignore
   -rw-rw-r--     1 pete  admin      34 14 Aug  2006 .gdb_history
   -rw-rw-r--     1 pete  admin    1044 28 Apr  2004 BUGS
        ^
        |

the  ``w´´  in the marked column in red and bold in the dired-warning  
face. Even when launched as 'emacs -Q.'

In GNU Emacs 23.0.50.2 (powerpc-apple-darwin8.10.0, X toolkit, Xaw3d  
scroll bars)
  of 2007-10-13 on Latsche.local
Windowing system distributor `The XFree86 Project, Inc', version  
11.0.40400000
configured using `configure  '--without-gtk' '--without-sound' '-- 
without-pop' '--with-x-toolkit=lucid' '--with-xpm' '--with-jpeg' '-- 
with-tiff' '--with-gif' '--with-png' '--enable-locallisppath=/Library/ 
Application Support/Emacs/calendar22:/Library/Application Support/ 
Emacs/caml:/Library/Application Support/Emacs:/sw/share/emacs21/site- 
lisp/elib' 'CPPFLAGS=-no-cpp-precomp -I/usr/include/openssl -I/sw/ 
include/pango-1.0 -I/sw/lib/freetype219/include -I/sw/lib/freetype219/ 
include/freetype2 -I/sw/lib/fontconfig2/include -I/sw/include/ 
libpng12 -I/usr/local/include -I/sw/include' 'CXXFLAGS=-no-cpp- 
precomp -I/usr/include/openssl -I/sw/include/pango-1.0 -I/sw/lib/ 
freetype219/include -I/sw/lib/freetype219/include/freetype2 -I/sw/lib/ 
fontconfig2/include -I/sw/include/libpng12 -I/usr/local/include -I/sw/ 
include' 'CFLAGS=-pipe -bind_at_load -fPIC -mcpu=7450 -mtune=7450 - 
fast -mpim-altivec -ftree-vectorize -foptimize-register-move - 
freorder-blocks -freorder-blocks-and-partition -fthread-jumps - 
fpeephole -fno-crossjumping' 'LDFLAGS=-dead_strip -multiply_defined  
suppress -bind_at_load -L/sw/lib/freetype219/lib -L/sw/lib/ 
fontconfig2/lib -L/sw/lib/ncurses -L/usr/local/lib -L/sw/lib''

Important settings:
   value of $LC_ALL: nil
   value of $LC_COLLATE: nil
   value of $LC_CTYPE: de_DE.UTF-8
   value of $LC_MESSAGES: nil
   value of $LC_MONETARY: nil
   value of $LC_NUMERIC: nil
   value of $LC_TIME: nil
   value of $LANG: de_DE.UTF-8
   locale-coding-system: utf-8
   default-enable-multibyte-characters: t

Major mode: Dired by name

Minor modes in effect:
   show-paren-mode: t
   display-time-mode: t
   tooltip-mode: t
   mouse-wheel-mode: t
   menu-bar-mode: t
   file-name-shadow-mode: t
   global-font-lock-mode: t
   font-lock-mode: t
   blink-cursor-mode: t
   unify-8859-on-encoding-mode: t
   utf-translate-cjk-mode: t
   auto-compression-mode: t
   line-number-mode: t
   transient-mark-mode: t



--
Greetings

   Pete

Well done is better than well said.
                            -- Benjamin Franklin

Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[martin rudalics]

> Since an update from CVS this noon (UTC) dired shows in case of files  
> or directories like these
> 
>   .../emacs:
>   total used in directory 9248 available 22772200
>   drwxrwxr-x    66 pete  admin    2244 13 Okt 19:25 .
>   drwxrwxr-x    20 pete  admin     680 27 Sep 13:15 ..
>   -rw-rw-r--     1 pete  admin     210  4 Mär  2005 .arch-inventory
>   -rw-rw-r--     1 pete  admin     500 22 Mär  2007 .cvsignore
>   -rw-rw-r--     1 pete  admin      34 14 Aug  2006 .gdb_history
>   -rw-rw-r--     1 pete  admin    1044 28 Apr  2004 BUGS
>        ^
>        |
> 
> the  ``wŽŽ  in the marked column in red and bold in the dired-warning  
> face. Even when launched as 'emacs -Q.'
>

See http://lists.gnu.org/archive/html/emacs-devel/2007-10/msg00588.html

Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Peter Dyballa]

Am 13.10.2007 um 20:22 schrieb martin rudalics:

>> Since an update from CVS this noon (UTC) dired shows in case of  
>> files  or directories like these
>>   .../emacs:
>>   total used in directory 9248 available 22772200
>>   drwxrwxr-x    66 pete  admin    2244 13 Okt 19:25 .
>>   drwxrwxr-x    20 pete  admin     680 27 Sep 13:15 ..
>>   -rw-rw-r--     1 pete  admin     210  4 Mär  2005 .arch-inventory
>>   -rw-rw-r--     1 pete  admin     500 22 Mär  2007 .cvsignore
>>   -rw-rw-r--     1 pete  admin      34 14 Aug  2006 .gdb_history
>>   -rw-rw-r--     1 pete  admin    1044 28 Apr  2004 BUGS
>>        ^
>>        |
>> the  ``wŽŽ  in the marked column in red and bold in the dired- 
>> warning  face. Even when launched as 'emacs -Q.'
>>
>
> See http://lists.gnu.org/archive/html/emacs-devel/2007-10/ 
> msg00588.html
>

Why are security holes now distributed with GNU Emacs from CVS?

--
Greetings

   Pete

      _o    o         o   o
    _<<     \\_/\_,   \\_ \\_/\_,
   (*)/(*) (*)   (*) (*) `-    (*)

Re: 23.0.50; Middle ``w▌▌ in of permissions in dired-mode is red and bold: dired-warning

[Glenn Morris]

Peter Dyballa wrote:

> Why are security holes now distributed with GNU Emacs from CVS?

This question is a bizarre non-sequitur.

Re: Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Peter Dyballa <Peter_Dyballa@Freenet.DE>
> Date: Sat, 13 Oct 2007 20:41:24 +0200
> Cc: emacs-pretest-bug@gnu.org
> 
> >>   -rw-rw-r--     1 pete  admin    1044 28 Apr  2004 BUGS
> >>        ^
> >>        |
> >> the  ``w´´  in the marked column in red and bold in the dired- 
> >> warning  face. Even when launched as 'emacs -Q.'
> >>
> >
> > See http://lists.gnu.org/archive/html/emacs-devel/2007-10/msg00588.html
> >
> 
> Why are security holes now distributed with GNU Emacs from CVS?

There are no security holes in the Emacs CVS, as I'm sure you
understand.

IMO, this whole feature of highlighting attributes is an annoyance,
and the only reason I could live with it until now is because it
previously had a not-so-in-my-face color.  Now someone made it glaring
red, which is too much for me.  So I now turned this off by default on
Windows, which _defaults_ to group- and world-writable files, even on
latest versions with strict security features, such as full-fledged
ACLs.

Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Peter Dyballa]

Am 13.10.2007 um 22:37 schrieb Eli Zaretskii:

> IMO, this whole feature of highlighting attributes is an annoyance,

This is my opinion, too. And there are probably a few thousands  
others that make deliberately files and (some) directories group  
writable. Another configure option to switch that annoyance off would  
be fine ... or I'll have to find a way to set that face to "normal,"  
i.e. regular and in foreground colour, in the system init file.

--
Greetings

   Pete

Bake pizza not war!

Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Miles Bader]

Peter Dyballa <Peter_Dyballa@Freenet.DE> writes:
>> IMO, this whole feature of highlighting attributes is an annoyance,
>
> This is my opinion, too. And there are probably a few thousands  others
> that make deliberately files and (some) directories group
> writable. Another configure option to switch that annoyance off would
> be fine ... or I'll have to find a way to set that face to "normal,"
> i.e. regular and in foreground colour, in the system init file.

Seriously, it's (intentionally, and explicitly) the _default_ in many places.

This "feature" should be disabled by default.

-Miles

-- 
`To alcohol!  The cause of, and solution to,
 all of life's problems' --Homer J. Simpson

Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> Cc: Eli Zaretskii <eliz@gnu.org>,  rudalics@gmx.at,  emacs-pretest-bug@gnu.org
> From: Miles Bader <miles@gnu.org>
> Date: Sun, 14 Oct 2007 08:24:42 +0900
> 
> Seriously, it's (intentionally, and explicitly) the _default_ in many places.
> 
> This "feature" should be disabled by default.

Then perhaps the new face dired-warn-writable that I introduced
yesterday should inherit from `default' on _all_ platforms, not only
on DOS/Windows.

Re: Re: 23.0.50; Middle ``wŽŽ in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> Cc: rudalics@gmx.at,
>  emacs-pretest-bug@gnu.org
> From: Peter Dyballa <Peter_Dyballa@Freenet.DE>
> Date: Sat, 13 Oct 2007 23:30:35 +0200
> 
> Am 13.10.2007 um 22:37 schrieb Eli Zaretskii:
> 
> > IMO, this whole feature of highlighting attributes is an annoyance,
> 
> This is my opinion, too. And there are probably a few thousands  
> others that make deliberately files and (some) directories group  
> writable. Another configure option to switch that annoyance off would  
> be fine ... or I'll have to find a way to set that face to "normal,"  
> i.e. regular and in foreground colour, in the system init file.

With my latest change, you can do that by customizing the new face
dired-warn-writable.

RE: Re: 23.0.50; Middle ``wZZ in of permissions in dired-mode is red and bold: dired-warning

[Drew Adams]

> With my latest change, you can do that by customizing the new face
> dired-warn-writable.

Good. So we've come almost full circle. There was already a separate face
for this in the original (dired+.el), except that the face did not have a
name that included "warn" (what are you WARNING about, anyway?). In the
original, there was a separate face for each file privilege: read, write,
execute (and also dir, link, and rare (bcsmpS)). There was no association of
a warning with writable files. And there still is no need for any warning or
for any alarmingly bright face for a write permission.

The attempt to reuse `font-lock-warning-face' for the totally unrelated
purpose of a file permission was misguided. After people complained and
sensed that that was a mistake, someone changed it briefly from
`font-lock-warning-face' to `font-lock-comment-delimiter-face'. That was
just as misguided. The problem was not just that a bright red face was
inappropriate; the problem was that there was no dedicated face for this, so
users had to reuse some existing face. And now we still see "warn" in the
face name, presumably for purely legacy reasons. What's that about?

There is a lesson in this. The real mistake was trying to avoid defining
separate faces for the different font-lock components of the Dired display.
I don't know what motivates such behavior, but I see it repeated over and
over. Perhaps it's for reasons of presumed economy; if so, it is a false
economy.

I really don't understand this propensity to reuse faces in contexts that
have no semantic connection. What if a user has a completely different frame
background for programming modes and for Dired? Dired has no notion of
comments, so why coopt a comment face to use in Dired? And what does a
warning have to do with a write permission? These things are totally
unrelated, but they end up being related by force because some implementors
are too conservative in creating new faces.

This artificial reuse produces unnecessary coupling - making a user search
for a face that fits both a file permission and either warnings or code
commenting - good luck! There is no logical connection, and there is no
reason to impose a connection. Even now, after you have removed that
coupling (thank you), there remains a vestigial trace of it in the "warn"
name you still use - `dired-warn-writable'? Yuck!

It is of course OK to reuse an existing face as a default value for a new
face. But I don't understand why there is so much resistance to defining new
faces.

People have used dired+.el for over a decade without reporting any problem
with the faces - they could do what they wanted with them. The only user
requests wrt dired+.el faces over the years were for more, not fewer, to
distinguish additional aspects of the display. The last such request I
received was just this past September - for his own use, a user added a new
font-lock pattern and a new face, to color the whole file name when the file
was of a certain type. I chose not to add that to dired+.el, but it shows
that users want more control over the appearance, not less.

[Wager: The day we add an easy, interactive, WYSIWYG way for users to define
some of their own font-locking (that would be a useful feature, IMO), they
will font-lock more, not less.]

Make it easy for people to customize the appearance as they like, by having
separate faces for logically separate components. A warning and a
file-permission indicator have nothing in common. If you find yourself
coupling such things, you are likely making a mistake.

Wrt those who are annoyed by "too much" such Dired highlighting: I haven't
checked how this was recently implemented in vanilla Emacs, but in dired+ it
is relegated to a second level of highlighting, so you see it only if you
choose `font-lock-maximum-decoration'. And of course all Dired highlighting
goes away if you turn off `font-lock-mode'. I'd hope that there is similar
user control in vanilla Emacs now: from customizing individual Dired faces,
to choosing a font-lock level, to turning font-lock off altogether.

FWIW, dired+ is here: http://www.emacswiki.org/cgi-bin/wiki/DiredPlus.

Re: RE: Re: 23.0.50; Middle ``wZZ in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Cc: <rudalics@gmx.at>, <emacs-pretest-bug@gnu.org>
> Date: Sat, 13 Oct 2007 22:40:21 -0700
> 
> And now we still see "warn" in the
> face name, presumably for purely legacy reasons. What's that about?

It warns about group-writable and world-writable files.  Presumably,
some people see a potential security hole in such files.  (I say such
people need an OS that's better protected than your garden-variety
GNU/Linux or Unix, like SELinux, but that's me.)

> Wrt those who are annoyed by "too much" such Dired highlighting: I haven't
> checked how this was recently implemented in vanilla Emacs, but in dired+ it
> is relegated to a second level of highlighting, so you see it only if you
> choose `font-lock-maximum-decoration'.

This will not help, as `font-lock-maximum-decoration' defaults to t
since Emacs 20.1.

> And of course all Dired highlighting goes away if you turn off
> `font-lock-mode'.

The same happens in stock Emacs 22.1.

RE: RE: Re: 23.0.50; Middle ``wZZ in of permissions in dired-mode is red and bold: dired-warning

[Drew Adams]

> > Wrt those who are annoyed by "too much" such Dired
> > highlighting: ... you see it only if you
> > choose `font-lock-maximum-decoration'.
>
> This will not help, as `font-lock-maximum-decoration' defaults to t
> since Emacs 20.1.

Of course it will help. "Those who are annoyed by 'too much' such Dired
highlighting" can choose nil instead of the default value of t.

Re: 23.0.50; Middle ``wZZ in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: "Drew Adams" <drew.adams@oracle.com>
> Cc: <emacs-devel@gnu.org>
> Date: Sun, 14 Oct 2007 13:54:22 -0700
> 
> > > Wrt those who are annoyed by "too much" such Dired
> > > highlighting: ... you see it only if you
> > > choose `font-lock-maximum-decoration'.
> >
> > This will not help, as `font-lock-maximum-decoration' defaults to t
> > since Emacs 20.1.
> 
> Of course it will help. "Those who are annoyed by 'too much' such Dired
> highlighting" can choose nil instead of the default value of t.

Yeah, and lose the otherwise useful fontifications elsewhere in Emacs?
no, thanks.

In any case, at least my gripes were about the _defaults_, and by
default font-lock-maximum-decoration is on.  I know how to customize
my faces for a long time, but IMO I shouldn't be requested to
customize a face because someone else's paranoia.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> And now we still see "warn" in the
>> face name, presumably for purely legacy reasons. What's that about?
>
> It warns about group-writable and world-writable files.  Presumably,
> some people see a potential security hole in such files.  (I say such
> people need an OS that's better protected than your garden-variety
> GNU/Linux or Unix, like SELinux, but that's me.)

I don't understand the need of adding a new face `dired-warn-writable'?
It adds another unnecessary face, and still doesn't fix a problem
of the annoying bold red indication of group and world writable permissions.

If no one finds this indication useful, we can remove it completely
(I must admit it never helped me to find wrong permissions).  Otherwise,
we could revert the `dired-warning' face to less annoying colors, and
maybe turn this face off by default on Windows (is this possible to do
with a face condition?)

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> Cc: emacs-devel@gnu.org
> From: Juri Linkov <juri@jurta.org>
> Date: Mon, 15 Oct 2007 00:21:01 +0300
> 
> I don't understand the need of adding a new face `dired-warn-writable'?

To get me off the hook with people who use dired-warning-face.

I don't mind removing dired-warning-face, but its doc string is quite
general:

  "Face used to highlight a part of a buffer that needs user attention."

which to me means it could be used for something other than mode
bits.  "Perhaps someone customized dired-font-lock-keywords," says I,
in which case leaving that face alone won't do any particular harm.

> It adds another unnecessary face, and still doesn't fix a problem
> of the annoying bold red indication of group and world writable permissions.

It does fix that on platforms where group- and world-writable files
are the default.  On other platforms, you can customize it, or we can
make it inherit from `default' on _all_ platforms.  I didn't do it
because I was unsure how many users of Unix and GNU systems actually
want to pay attention to such files.

> If no one finds this indication useful

Obviously, someone did, as this face is with us since 2004.  And
Richard's comment there indicates that someone tried to highlight the
entire file name around 1997.

> Otherwise, we could revert the `dired-warning' face to less annoying
> colors

As long as the face's name is *-warning, I think making it inherit
from font-lock-warning-face is TRT: if we want to warn about
something, let that something stand out like any other warning.

That is why I renamed the face to something less general.  And if the
"warn" part isn't consistent enough with what I say above, it (the
"warn" part of the face's name) can go away.

> and maybe turn this face off by default on Windows (is this
> possible to do with a face condition?)

AFAIK, not really; you can only condition that on the value of
window-system.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> Cc: emacs-devel@gnu.org
>> From: Juri Linkov <juri@jurta.org>
>> Date: Mon, 15 Oct 2007 00:21:01 +0300
>> 
>> I don't understand the need of adding a new face `dired-warn-writable'?
>
> To get me off the hook with people who use dired-warning-face.
>
> I don't mind removing dired-warning-face, but its doc string is quite
> general:
>
>   "Face used to highlight a part of a buffer that needs user attention."
>
> which to me means it could be used for something other than mode
> bits.  "Perhaps someone customized dired-font-lock-keywords," says I,
> in which case leaving that face alone won't do any particular harm.
>
>> It adds another unnecessary face, and still doesn't fix a problem
>> of the annoying bold red indication of group and world writable permissions.
>
> It does fix that on platforms where group- and world-writable files
> are the default.  On other platforms, you can customize it, or we can
> make it inherit from `default' on _all_ platforms.  I didn't do it
> because I was unsure how many users of Unix and GNU systems actually
> want to pay attention to such files.
>
>> Otherwise, we could revert the `dired-warning' face to less annoying
>> colors
>
> As long as the face's name is *-warning, I think making it inherit
> from font-lock-warning-face is TRT: if we want to warn about
> something, let that something stand out like any other warning.
>
> That is why I renamed the face to something less general.  And if the
> "warn" part isn't consistent enough with what I say above, it (the
> "warn" part of the face's name) can go away.

When testing Emacs 22 with `emacs -Q', I see that dired buffers are
almost unusable by default.  The red bold face for write permissions
is very distracting that prevents focusing on other parts of the dired
buffer.  I'm afraid Emacs users will hate us for this if we release
Emacs 22.2 with this annoyance.

I propose a patch below for the 22 branch that renames the face
`dired-warn-writable' to more appropriate name `dired-perm-write'
that doesn't associate it with a warning, and changes its parent face
from `font-lock-warning-face' to `font-lock-comment-delimiter-face'
on non-Windows platforms.

The face `dired-warning' that inherits from `font-lock-warning-face'
still remains in dired.el, but it is not used anywhere.

Index: lisp/dired.el
===================================================================
RCS file: /sources/emacs/emacs/lisp/dired.el,v
retrieving revision 1.359.2.10
diff -c -r1.359.2.10 dired.el
*** lisp/dired.el	22 Feb 2008 18:47:26 -0000	1.359.2.10
--- lisp/dired.el	13 Mar 2008 02:19:00 -0000
***************
*** 344,356 ****
  (defvar dired-warning-face 'dired-warning
    "Face name used for a part of a buffer that needs user attention.")
  
! (defface dired-warn-writable
    '((((type w32 pc)) :inherit default)  ;; These default to rw-rw-rw.
!     (t (:inherit font-lock-warning-face)))
    "Face used to highlight permissions of group- and world-writable files."
    :group 'dired-faces
    :version "22.2")
! (defvar dired-warn-writable-face 'dired-warn-writable
    "Face name used for permissions of group- and world-writable files.")
  
  (defface dired-directory
--- 344,356 ----
  (defvar dired-warning-face 'dired-warning
    "Face name used for a part of a buffer that needs user attention.")
  
! (defface dired-perm-write
    '((((type w32 pc)) :inherit default)  ;; These default to rw-rw-rw.
!     (t (:inherit font-lock-comment-delimiter-face)))
    "Face used to highlight permissions of group- and world-writable files."
    :group 'dired-faces
    :version "22.2")
! (defvar dired-perm-write-face 'dired-perm-write
    "Face name used for permissions of group- and world-writable files.")
  
  (defface dired-directory
***************
*** 414,423 ****
     ;; fields with keymaps to frob the permissions, somewhat a la XEmacs.
     (list (concat dired-re-maybe-mark dired-re-inode-size
  		 "[-d]....\\(w\\)....")	; group writable
! 	 '(1 dired-warn-writable-face))
     (list (concat dired-re-maybe-mark dired-re-inode-size
  		 "[-d].......\\(w\\).")	; world writable
! 	 '(1 dired-warn-writable-face))
     ;;
     ;; Subdirectories.
     (list dired-re-dir
--- 414,423 ----
     ;; fields with keymaps to frob the permissions, somewhat a la XEmacs.
     (list (concat dired-re-maybe-mark dired-re-inode-size
  		 "[-d]....\\(w\\)....")	; group writable
! 	 '(1 dired-perm-write-face))
     (list (concat dired-re-maybe-mark dired-re-inode-size
  		 "[-d].......\\(w\\).")	; world writable
! 	 '(1 dired-perm-write-face))
     ;;
     ;; Subdirectories.
     (list dired-re-dir

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Date: Thu, 13 Mar 2008 04:19:37 +0200
> Cc: emacs-devel@gnu.org
> 
> When testing Emacs 22 with `emacs -Q', I see that dired buffers are
> almost unusable by default.  The red bold face for write permissions
> is very distracting that prevents focusing on other parts of the dired
> buffer.  I'm afraid Emacs users will hate us for this if we release
> Emacs 22.2 with this annoyance.
> 
> I propose a patch below for the 22 branch that renames the face
> `dired-warn-writable' to more appropriate name `dired-perm-write'
> that doesn't associate it with a warning, and changes its parent face
> from `font-lock-warning-face' to `font-lock-comment-delimiter-face'
> on non-Windows platforms.

I object.  I think now is not the time for making behavioral changes
in the release branch.  We are too close to the release.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> When testing Emacs 22 with `emacs -Q', I see that dired buffers are
>> almost unusable by default.  The red bold face for write permissions
>> is very distracting that prevents focusing on other parts of the dired
>> buffer.  I'm afraid Emacs users will hate us for this if we release
>> Emacs 22.2 with this annoyance.
>>
>> I propose a patch below for the 22 branch that renames the face
>> `dired-warn-writable' to more appropriate name `dired-perm-write'
>> that doesn't associate it with a warning, and changes its parent face
>> from `font-lock-warning-face' to `font-lock-comment-delimiter-face'
>> on non-Windows platforms.
>
> I object.  I think now is not the time for making behavioral changes
> in the release branch.  We are too close to the release.

This is not a behavioral change.  This is a fix for the grave usability bug
you introduced to GNU/Linux users.  It seems you realize the severity of
this problem since you wrote several times proposing to fix it:

> Date: Sun, 14 Oct 2007 06:10:33 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> Then perhaps the new face dired-warn-writable that I introduced
> yesterday should inherit from `default' on _all_ platforms, not only
> on DOS/Windows.

and

> Date: Mon, 15 Oct 2007 00:20:45 +0200
> From: Eli Zaretskii <eliz@gnu.org>
> It does fix that on platforms where group- and world-writable files
> are the default.  On other platforms, you can customize it, or we can
> make it inherit from `default' on _all_ platforms.

I think it is not yet too late to fix this bug because there is absolutely
no risk of breaking something else with this simple fix.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Cc: emacs-devel@gnu.org
> Date: Thu, 13 Mar 2008 12:37:10 +0200
> 
> >> I propose a patch below for the 22 branch that renames the face
> >> `dired-warn-writable' to more appropriate name `dired-perm-write'
> >> that doesn't associate it with a warning, and changes its parent face
> >> from `font-lock-warning-face' to `font-lock-comment-delimiter-face'
> >> on non-Windows platforms.
> >
> > I object.  I think now is not the time for making behavioral changes
> > in the release branch.  We are too close to the release.
> 
> This is not a behavioral change.

Changes in faces are changes in behavior.

> This is a fix for the grave usability bug
> you introduced to GNU/Linux users.  It seems you realize the severity of
> this problem since you wrote several times proposing to fix it:

That was in October, 6 months ago.  _Then_ I didn't mind, as I don't
mind now that you make this change on the trunk.

> I think it is not yet too late to fix this bug

I think it is too late.

> because there is absolutely no risk of breaking something else with
> this simple fix.

There's no such thing as ``absolutely no risk'' in software changes.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> >> I propose a patch below for the 22 branch that renames the face
>> >> `dired-warn-writable' to more appropriate name `dired-perm-write'
>> >> that doesn't associate it with a warning, and changes its parent face
>> >> from `font-lock-warning-face' to `font-lock-comment-delimiter-face'
>> >> on non-Windows platforms.
>> >
>> > I object.  I think now is not the time for making behavioral changes
>> > in the release branch.  We are too close to the release.
>>
>> This is not a behavioral change.
>
> Changes in faces are changes in behavior.

Faces are not behavior.  Faces are a visual appearance.

>> This is a fix for the grave usability bug
>> you introduced to GNU/Linux users.  It seems you realize the severity of
>> this problem since you wrote several times proposing to fix it:
>
> That was in October, 6 months ago.  _Then_ I didn't mind, as I don't
> mind now that you make this change on the trunk.

OK, installed on the trunk.

>> I think it is not yet too late to fix this bug
>
> I think it is too late.

This can be considered as a bug discovered during the pretest.
What a sense to make pretests if it is not allowed to fix bugs
found in these pretests?!

-- 
Juri Linkov
http://www.jurta.org/emacs/

23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juanma Barranquero]

[I sent this privately by accident]

On Fri, Mar 14, 2008 at 2:07 AM, Juri Linkov <juri@jurta.org> wrote:

>  Faces are not behavior.  Faces are a visual appearance.

Yes. Emacs using a face now, when it used another one yesterday for
the same purpose, is a change in behavior, though.

>  This can be considered as a bug discovered during the pretest.
>  What a sense to make pretests if it is not allowed to fix bugs
>  found in these pretests?!

The severity of the bug must be taken into account so late in the release cycle.

Juanma

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>>  This can be considered as a bug discovered during the pretest.
>>  What a sense to make pretests if it is not allowed to fix bugs
>>  found in these pretests?!
>
> The severity of the bug must be taken into account so late in the
> release cycle.

And?  Do you have an opinion about its severity?  I think it will be
very annoying for the users of Emacs 22.2.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juanma Barranquero]

On Fri, Mar 14, 2008 at 10:30 PM, Juri Linkov <juri@jurta.org> wrote:

>  And?  Do you have an opinion about its severity?

Not really, but bad defaults for faces aren't high priority bugs IMHO
unless they affect pretty fundamental faces (default, mode-line,
things like that).

 Juanma

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>>  And?  Do you have an opinion about its severity?
>
> Not really, but bad defaults for faces aren't high priority bugs IMHO
> unless they affect pretty fundamental faces (default, mode-line,
> things like that).

I think dired faces are fundamental.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juanma Barranquero]

On Sat, Mar 15, 2008 at 1:17 AM, Juri Linkov <juri@jurta.org> wrote:

>  I think dired faces are fundamental.

I don't use dired often enough to have an opinion.

 Juanma

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Date: Fri, 14 Mar 2008 03:07:10 +0200
> Cc: emacs-devel@gnu.org
> 
> > That was in October, 6 months ago.  _Then_ I didn't mind, as I don't
> > mind now that you make this change on the trunk.
> 
> OK, installed on the trunk.

Thanks.

> >> I think it is not yet too late to fix this bug
> >
> > I think it is too late.
> 
> This can be considered as a bug discovered during the pretest.

It cannot be considered a bug, since the change was deliberate.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Chong Yidong]

Juri Linkov <juri@jurta.org> writes:

> When testing Emacs 22 with `emacs -Q', I see that dired buffers are
> almost unusable by default.  The red bold face for write permissions
> is very distracting that prevents focusing on other parts of the dired
> buffer.  I'm afraid Emacs users will hate us for this if we release
> Emacs 22.2 with this annoyance.

I haven't been paying attention here, so sorry if this has been
discussed elsewhere in the thread.

As far as I can tell, the face is meant to be distracting.  It's
supposed to warn when files are writable by other users.  Why would
this feature make Emacs users hate us?

If you really really dislike (hate?) the warning face, one thing we
could try is to make it red and non-bold on terminals that support
colors.  If that's an acceptable compromise, then we can safely go
ahead and make that change.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> When testing Emacs 22 with `emacs -Q', I see that dired buffers are
>> almost unusable by default.  The red bold face for write permissions
>> is very distracting that prevents focusing on other parts of the dired
>> buffer.  I'm afraid Emacs users will hate us for this if we release
>> Emacs 22.2 with this annoyance.
>
> I haven't been paying attention here, so sorry if this has been
> discussed elsewhere in the thread.
>
> As far as I can tell, the face is meant to be distracting.  It's
> supposed to warn when files are writable by other users.  Why would
> this feature make Emacs users hate us?

I think no face should be distracting.  In very rare cases, a strong
color should indicate an error, so the user has to fix it immediately,
thus removing this error together with the face indication from the view.
In case of group/world write permissions, usually there is no error,
so no action is required from the user to remove write permissions
immediately after seeing them, because in most cases these permissions
are valid.  Often the user even can't change permission, for instance,
when a non-root user browses the /usr hierarchy that has group/world
write permissions on many files.

So file permissions should not be the first place to attract attention
of the user to after visiting a dired buffer.

> If you really really dislike (hate?) the warning face, one thing we
> could try is to make it red and non-bold on terminals that support
> colors.  If that's an acceptable compromise, then we can safely go
> ahead and make that change.

Actually a good face to use is font-lock-comment-delimiter-face
because it is not glaring as font-lock-warning-face.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Cc: Eli Zaretskii <eliz@gnu.org>,  emacs-devel@gnu.org
> Date: Sat, 15 Mar 2008 02:16:54 +0200
> 
> I think no face should be distracting.  In very rare cases, a strong
> color should indicate an error, so the user has to fix it immediately,
> thus removing this error together with the face indication from the view.

Is warning also okay to glare, or just the errors?
font-lock-warning-face has been glaring for ages.

> > If you really really dislike (hate?) the warning face, one thing we
> > could try is to make it red and non-bold on terminals that support
> > colors.  If that's an acceptable compromise, then we can safely go
> > ahead and make that change.
> 
> Actually a good face to use is font-lock-comment-delimiter-face
> because it is not glaring as font-lock-warning-face.

As I wrote elsewhere in this thread, such a change will effectively
reintroduce the problem on text terminals that Martin Rudalics was
trying to fix when he made dired-warn inherit from
font-lock-warning-face.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> I think no face should be distracting.  In very rare cases, a strong
>> color should indicate an error, so the user has to fix it immediately,
>> thus removing this error together with the face indication from the view.
>
> Is warning also okay to glare, or just the errors?
> font-lock-warning-face has been glaring for ages.

A warning/error is okay to glare in places that should draw the user's
attention.  But for file attributes it is an annoyance confirmed by
your own opinion:

    From: Eli Zaretskii <eliz@gnu.org>
    Subject: Re: Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning
    Date: Sat, 13 Oct 2007 22:37:07 +0200

    IMO, this whole feature of highlighting attributes is an annoyance,
    and the only reason I could live with it until now is because it
    previously had a not-so-in-my-face color.  Now someone made it glaring
    red, which is too much for me.

You can expect the same angry reaction from other Emacs users after releasing
Emacs 22.2 without fixing this annoyance.

>> > If you really really dislike (hate?) the warning face, one thing we
>> > could try is to make it red and non-bold on terminals that support
>> > colors.  If that's an acceptable compromise, then we can safely go
>> > ahead and make that change.
>>
>> Actually a good face to use is font-lock-comment-delimiter-face
>> because it is not glaring as font-lock-warning-face.
>
> As I wrote elsewhere in this thread, such a change will effectively
> reintroduce the problem on text terminals that Martin Rudalics was
> trying to fix when he made dired-warn inherit from
> font-lock-warning-face.

No, it fixes the original reported problem since font-lock-comment-delimiter-face
defines colors for text terminals.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Juri Linkov <juri@jurta.org>
> Cc: cyd@stupidchicken.com,  emacs-devel@gnu.org
> Date: Sat, 15 Mar 2008 18:20:00 +0200
> 
>     From: Eli Zaretskii <eliz@gnu.org>
>     Subject: Re: Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning
>     Date: Sat, 13 Oct 2007 22:37:07 +0200
> 
>     IMO, this whole feature of highlighting attributes is an annoyance,
>     and the only reason I could live with it until now is because it
>     previously had a not-so-in-my-face color.  Now someone made it glaring
>     red, which is too much for me.

I was talking in the context of MS-Windows, where world-writable is
the default behavior.  Of course, I don't mind that this is turned off
on Posix platforms as well, but I wouldn't press them into it.

> You can expect the same angry reaction from other Emacs users after releasing
> Emacs 22.2 without fixing this annoyance.

This situation is much more rare on Posix platforms.  Anyway, the only
reason I object to this change is because we are too close to a
release.  But it's not my call.

> > As I wrote elsewhere in this thread, such a change will effectively
> > reintroduce the problem on text terminals that Martin Rudalics was
> > trying to fix when he made dired-warn inherit from
> > font-lock-warning-face.
> 
> No, it fixes the original reported problem since font-lock-comment-delimiter-face
> defines colors for text terminals.

Yes, I stand corrected on this one.  Sorry.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Eli Zaretskii]

> From: Chong Yidong <cyd@stupidchicken.com>
> Date: Fri, 14 Mar 2008 18:45:05 -0400
> Cc: Eli Zaretskii <eliz@gnu.org>, emacs-devel@gnu.org
> 
> If you really really dislike (hate?) the warning face, one thing we
> could try is to make it red and non-bold on terminals that support
> colors.  If that's an acceptable compromise, then we can safely go
> ahead and make that change.

It's your call, but at this late stage in the pretest, I'd advise
against _any_ changes that are not strictly necessary for a functional
Emacs.  Face appearance cannot be an obstacle to Emacs functionality,
IMO, because it's all too easy to customize such a face.

Some more data points from the history of this issue:

 . This problem began when Martin Rudalics made the following change:

    2007-10-08  Martin Rudalics  <rudalics@gmx.at>

	    * dired.el (dired-warning): Inherit from font-lock-warning-face to
	    make it show up with eight colors.

 . That change caused _all_ files' mode bits to be highlighted on
   MS-Windows, where the file permission default to world-writable.
   Therefore, I made a subsequent change:

    2007-10-13  Eli Zaretskii  <eliz@gnu.org>

	    * dired.el (dired-warn-writable): New face.
	    (dired-warn-writable-face): New variable.
	    (dired-font-lock-keywords): Use dired-warn-writable-face, instead
	    of dired-warning-face, for group- and world-writable files.
 
   The new face I introduced looks the same as dired-warning did after
   Martin's change, except in the DOS and Windows ports, where it
   looks like the default face.

   IOW, I didn't change the visual appearance of the mode bits on
   Posix platforms introduced by Martin's change, I just fixed a long
   standing problem whereby Emacs would highlight mode bits that are
   normal on DOS/Windows filesystems.

Now, Martin did his change for a reason, see the log entry above.  It
looks like the change on the trunk by Juri simply reversed Martin's
change (modulo face renaming), which means we are now back at the
problem whereby world-writable files will not stand out on 8-color
text terminals.  At the very least, we should take care of that
original problem on the trunk.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

> It's your call, but at this late stage in the pretest, I'd advise
> against _any_ changes that are not strictly necessary for a functional
> Emacs.  Face appearance cannot be an obstacle to Emacs functionality,
> IMO, because it's all too easy to customize such a face.

But this will require for most users of dired to customize this face
immediately after seeing it, and always run Emacs with .emacs that
contains this customization.  This will be annoying for them.

> Now, Martin did his change for a reason, see the log entry above.  It
> looks like the change on the trunk by Juri simply reversed Martin's
> change (modulo face renaming), which means we are now back at the
> problem whereby world-writable files will not stand out on 8-color
> text terminals.  At the very least, we should take care of that
> original problem on the trunk.

No, your analysis is not correct.

This problem began with the following bug report:
http://lists.gnu.org/archive/html/bug-gnu-emacs/2007-08/msg00212.html

    Date: Thu, 30 Aug 2007 10:53:38 +1000
    To: bug-gnu-emacs@gnu.org
    From: trentbuck@gmail.com
    Subject: dired-warning face not colored on 8-color ttys

    The `dired-warning' face inherits from font-lock-comment-face.  Since
    Emacs 22, this face is no longer red on eight-color terminals, but
    rather is the same colour as normal text.  This effectively means that
    the user cannot see warnings.

    I submit that dired-warning should inherit from
    font-lock-comment-delimiter-face.  Probably this change should be made
    to a number of other faces that inherit from font-lock-comment-face.

My change fixed this original problem by inheriting from
font-lock-comment-delimiter-face instead of font-lock-comment-face.
So we can say that the original problem is fixed now on the trunk.

During the discussion about fixing this problem, it was proposed to try
inheriting from font-lock-warning-face.  This change was unnecessary to
fix the original problem, and was accompanied by numerous doubts:
http://lists.gnu.org/archive/html/emacs-devel/2007-09/msg00436.html
http://lists.gnu.org/archive/html/emacs-devel/2007-10/msg00756.html

After this experimental change was installed, it caused complaints:
http://lists.gnu.org/archive/html/emacs-devel/2007-10/msg00724.html

Now we have a situation the very opposite to what it has to be:
the trunk contains a fix for the reported bug, and the 22 branch
contains an experimental incompatible change.

I think the most reasonable and safest thing to do now on the
branch is to install the original bug fix that changes the
parent of the dired-warning face from font-lock-comment-face
to font-lock-comment-delimiter-face.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Stefan Monnier]

> I think the most reasonable and safest thing to do now on the
> branch is to install the original bug fix that changes the
> parent of the dired-warning face from font-lock-comment-face
> to font-lock-comment-delimiter-face.

That sounds reasonable after all,


        Stefan

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[martin rudalics]

>>I think the most reasonable and safest thing to do now on the
>>branch is to install the original bug fix that changes the
>>parent of the dired-warning face from font-lock-comment-face
>>to font-lock-comment-delimiter-face.
> 
> 
> That sounds reasonable after all,

By all means let's do that.

Re: 23.0.50; Middle w in of permissions in dired-mode is red and bold: dired-warning

[Juri Linkov]

>> I think the most reasonable and safest thing to do now on the
>> branch is to install the original bug fix that changes the
>> parent of the dired-warning face from font-lock-comment-face
>> to font-lock-comment-delimiter-face.
>
> That sounds reasonable after all,

Done.

-- 
Juri Linkov
http://www.jurta.org/emacs/

Re: 23.0.50; Middle ``wZZ in of permissions in dired-mode is red and bold: dired-warning

[Stefan Monnier]

> This will not help, as `font-lock-maximum-decoration' defaults to t
> since Emacs 20.1.

AFAIK it defaulted to t ever since it was created (could be 20.1 for all
I know).


        Stefan