Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 4332 bytes --]

Hi,

I was trying to using the binding for highlight-regexp and I got confused
if the prefix was "C-x w r" or "M-s h r". I "grepping" (actually agging), I
found that below are the ONLY bindings that begin with "C-x w .." and "M-s
h .." (with the exact same number of bindings too).

Can we make the bindings consistent so that it doesn't matter is the user
is using "C-x w" or "M-s h"?

km²~/downloads/:git/emacs/lisp> ag 'define.*C\-xw'
hi-lock.el
277:    (define-key map "\C-xwi" 'hi-lock-find-patterns)
278:    (define-key map "\C-xwl" 'highlight-lines-matching-regexp)
279:    (define-key map "\C-xwp" 'highlight-phrase)
280:    (define-key map "\C-xwh" 'highlight-regexp)
281:    (define-key map "\C-xw." 'highlight-symbol-at-point)
282:    (define-key map "\C-xwr" 'unhighlight-regexp)
283:    (define-key map "\C-xwb" 'hi-lock-write-interactive-patterns)

km²~/downloads/:git/emacs/lisp> ag 'define.*search-map\s+"h'
bindings.el
935:(define-key search-map "hf"   'hi-lock-find-patterns)
932:(define-key search-map "hl"   'highlight-lines-matching-regexp)
931:(define-key search-map "hp"   'highlight-phrase)
930:(define-key search-map "hr"   'highlight-regexp)
933:(define-key search-map "h."   'highlight-symbol-at-point)
934:(define-key search-map "hu"   'unhighlight-regexp)
936:(define-key search-map "hw"   'hi-lock-write-interactive-patterns)

​Some bindings are consistent but some are not. Here are the inconsistent
ones:

hi-lock.el  277: (define-key map        "\C-xwi" 'hi-lock-find-patterns)
bindings.el 935: (define-key search-map     "hf" 'hi-lock-find-patterns)

hi-lock.el  282: (define-key map        "\C-xwr" 'unhighlight-regexp)
bindings.el 934: (define-key search-map     "hu" 'unhighlight-regexp)

hi-lock.el  280: (define-key map        "\C-xwh" 'highlight-regexp)
bindings.el 930: (define-key search-map     "hr" 'highlight-regexp)

hi-lock.el  283: (define-key map        "\C-xwb"
'hi-lock-write-interactive-patterns)
bindings.el 936: (define-key search-map     "hw"
'hi-lock-write-interactive-patterns)

*** Here are my suggested consistent bindings to fix the inconsistent ones:
***

- Currently "C-x w f" is undefined; "f" for find patterns
hi-lock.el  277: (define-key map        "\C-xwf" 'hi-lock-find-patterns) ;
changed
bindings.el 935: (define-key search-map     "hf" 'hi-lock-find-patterns) ;
no change

- Currently "C-x w u" is undefined; "u" for unhighlight
hi-lock.el  282: (define-key map        "\C-xwu" 'unhighlight-regexp) ;
changed
bindings.el 934: (define-key search-map     "hu" 'unhighlight-regexp) ; no
change

- "C-x w r" which was earlier bound to unhighlight-regexp should now be
bound to highlight-regexp; "r" for regexp
hi-lock.el  280: (define-key map        "\C-xwr" 'highlight-regexp) ;
changed
bindings.el 930: (define-key search-map     "hr" 'highlight-regexp) ; no
change

- Currently "C-x w w" is undefined; "w" for write interactive patterns
hi-lock.el  283: (define-key map        "\C-xww"
'hi-lock-write-interactive-patterns) ; changed
bindings.el 936: (define-key search-map     "hw"
'hi-lock-write-interactive-patterns) ; no change

Well.. turns out only 4 bindings need to change in hi-lock.el to make these
consistent.
​
|------------+-------------+---------------------+------------------------------------|
| search-map | hi-lock-map | last key in binding | function
          |
|------------+-------------+---------------------+------------------------------------|
| M-s h      | C-x w       | l                   |
highlight-lines-matching-regexp    |
| M-s h      | C-x w       | p                   | highlight-phrase
          |
| M-s h      | C-x w       | .                   |
highlight-symbol-at-point          |
|------------+-------------+---------------------+------------------------------------|
| M-s h      | C-x w       | f                   | hi-lock-find-patterns
           |
| M-s h      | C-x w       | u                   | unhighlight-regexp
          |
| M-s h      | C-x w       | r                   | highlight-regexp
          |
| M-s h      | C-x w       | w                   |
hi-lock-write-interactive-patterns |
|------------+-------------+---------------------+------------------------------------|

--
Kaushal Modi

[-- Attachment #2: Type: text/html, Size: 11430 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 4769 bytes --]

Adding Juri to the To: list (based on this commit:
http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=e5e4a94293d5a9a157557e53b4fea4e5d280673e
)

On Fri, Jul 10, 2015 at 3:02 PM Kaushal <kaushal.modi@gmail.com> wrote:

> Hi,
>
> I was trying to using the binding for highlight-regexp and I got confused
> if the prefix was "C-x w r" or "M-s h r". I "grepping" (actually agging), I
> found that below are the ONLY bindings that begin with "C-x w .." and "M-s
> h .." (with the exact same number of bindings too).
>
> Can we make the bindings consistent so that it doesn't matter is the user
> is using "C-x w" or "M-s h"?
>
> km²~/downloads/:git/emacs/lisp> ag 'define.*C\-xw'
> hi-lock.el
> 277:    (define-key map "\C-xwi" 'hi-lock-find-patterns)
> 278:    (define-key map "\C-xwl" 'highlight-lines-matching-regexp)
> 279:    (define-key map "\C-xwp" 'highlight-phrase)
> 280:    (define-key map "\C-xwh" 'highlight-regexp)
> 281:    (define-key map "\C-xw." 'highlight-symbol-at-point)
> 282:    (define-key map "\C-xwr" 'unhighlight-regexp)
> 283:    (define-key map "\C-xwb" 'hi-lock-write-interactive-patterns)
>
> km²~/downloads/:git/emacs/lisp> ag 'define.*search-map\s+"h'
> bindings.el
> 935:(define-key search-map "hf"   'hi-lock-find-patterns)
> 932:(define-key search-map "hl"   'highlight-lines-matching-regexp)
> 931:(define-key search-map "hp"   'highlight-phrase)
> 930:(define-key search-map "hr"   'highlight-regexp)
> 933:(define-key search-map "h."   'highlight-symbol-at-point)
> 934:(define-key search-map "hu"   'unhighlight-regexp)
> 936:(define-key search-map "hw"   'hi-lock-write-interactive-patterns)
>
> ​Some bindings are consistent but some are not. Here are the inconsistent
> ones:
>
> hi-lock.el  277: (define-key map        "\C-xwi" 'hi-lock-find-patterns)
> bindings.el 935: (define-key search-map     "hf" 'hi-lock-find-patterns)
>
> hi-lock.el  282: (define-key map        "\C-xwr" 'unhighlight-regexp)
> bindings.el 934: (define-key search-map     "hu" 'unhighlight-regexp)
>
> hi-lock.el  280: (define-key map        "\C-xwh" 'highlight-regexp)
> bindings.el 930: (define-key search-map     "hr" 'highlight-regexp)
>
> hi-lock.el  283: (define-key map        "\C-xwb"
> 'hi-lock-write-interactive-patterns)
> bindings.el 936: (define-key search-map     "hw"
> 'hi-lock-write-interactive-patterns)
>
> *** Here are my suggested consistent bindings to fix the inconsistent
> ones: ***
>
> - Currently "C-x w f" is undefined; "f" for find patterns
> hi-lock.el  277: (define-key map        "\C-xwf" 'hi-lock-find-patterns) ;
> changed
> bindings.el 935: (define-key search-map     "hf" 'hi-lock-find-patterns) ;
> no change
>
> - Currently "C-x w u" is undefined; "u" for unhighlight
> hi-lock.el  282: (define-key map        "\C-xwu" 'unhighlight-regexp) ;
> changed
> bindings.el 934: (define-key search-map     "hu" 'unhighlight-regexp) ; no
> change
>
> - "C-x w r" which was earlier bound to unhighlight-regexp should now be
> bound to highlight-regexp; "r" for regexp
> hi-lock.el  280: (define-key map        "\C-xwr" 'highlight-regexp) ;
> changed
> bindings.el 930: (define-key search-map     "hr" 'highlight-regexp) ; no
> change
>
> - Currently "C-x w w" is undefined; "w" for write interactive patterns
> hi-lock.el  283: (define-key map        "\C-xww"
> 'hi-lock-write-interactive-patterns) ; changed
> bindings.el 936: (define-key search-map     "hw"
> 'hi-lock-write-interactive-patterns) ; no change
>
> Well.. turns out only 4 bindings need to change in hi-lock.el to make
> these consistent.
> ​
>
> |------------+-------------+---------------------+------------------------------------|
> | search-map | hi-lock-map | last key in binding | function
>             |
>
> |------------+-------------+---------------------+------------------------------------|
> | M-s h      | C-x w       | l                   |
> highlight-lines-matching-regexp    |
> | M-s h      | C-x w       | p                   | highlight-phrase
>             |
> | M-s h      | C-x w       | .                   |
> highlight-symbol-at-point          |
>
> |------------+-------------+---------------------+------------------------------------|
> | M-s h      | C-x w       | f                   | hi-lock-find-patterns
>            |
> | M-s h      | C-x w       | u                   | unhighlight-regexp
>             |
> | M-s h      | C-x w       | r                   | highlight-regexp
>             |
> | M-s h      | C-x w       | w                   |
> hi-lock-write-interactive-patterns |
>
> |------------+-------------+---------------------+------------------------------------|
>
> --
> Kaushal Modi
>

[-- Attachment #2: Type: text/html, Size: 11316 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Juri Linkov]

> Adding Juri to the To: list (based on this commit:
> http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=e5e4a94293d5a9a157557e53b4fea4e5d280673e
> )

I completely support your proposal.  And maybe we could just remove
the duplicate keymap ‘C-x w’ altogether because there has been a need
for a free keymap for other features like window related commands in
http://thread.gmane.org/gmane.emacs.devel/159692/focus=159776 etc.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 595 bytes --]

Thanks for the support. +1 to remove the duplicate keymap.

On Sat, Jul 11, 2015 at 3:48 PM Juri Linkov <juri@jurta.org> wrote:

> > Adding Juri to the To: list (based on this commit:
> >
> http://git.savannah.gnu.org/cgit/emacs.git/commit/?id=e5e4a94293d5a9a157557e53b4fea4e5d280673e
> > )
>
> I completely support your proposal.  And maybe we could just remove
> the duplicate keymap ‘C-x w’ altogether because there has been a need
> for a free keymap for other features like window related commands in
> http://thread.gmane.org/gmane.emacs.devel/159692/focus=159776 etc.
>

[-- Attachment #2: Type: text/html, Size: 1103 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Vaidheeswaran C]

On Sunday 12 July 2015 01:18 AM, Juri Linkov wrote:

> And maybe we could just remove the duplicate keymap ‘C-x w’
> altogether

When this happens please make the prefix configurable, much like what
outline minor mode does.

Hide Outline Minor Mode Prefix: \x03@
    State : STANDARD.
   Prefix key to use for Outline commands in Outline minor mode. Hide
   The value of this variable is checked as part of loading Outline mode.
   After that, changing the prefix key requires manipulating keymaps.

C-x w is what I am used to and there is NOT an emacs session where I
DO NOT use this binding.  M-s interface and bindings still confuse me.

ps: `r' stands for REMOVING highlights.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 2905 bytes --]

@Vaidheeswaran

I am more interested in having the bindings be consistent. If people are
used to the C-x w prefix and the consensus is to keep it, I'll respect the
democracy.

Here is the reasoning for the proposed changes especially for
`highlight-regexp` and `unhighlight-regexp`.

I understood that currently `highlight-regexp` is bound to "C-x w h" where
`h` stands for Highlight. But then the same was bound to `M-s h r` in the
other prefix map.

I could confuse which prefix had the `r` binding to highlight Regexp and
could end up doing "C-x w r"! (in which case `r` stands for Removing
highlights). Why should we have a key do polar opposite stuff in two
keymaps?

So I surveyed the existing bindings in C-x w keymap:

h -> [h]ighlight-regexp
r -> unhighlight-regexp ([r]remove)
. -> highlight-symbol-at-[point]

On the other hand, in M-s h map:

h -> nil
r -> highlight-[r]egexp
u -> [u]nhighlight-regexp
. -> highlight-symbol-at-[point]

To me, the M-s h map seemed to have more harmony as the h binding wasn't
used up to bind anything because we have a couple of highlighting function
(regexp, symbol at point). Also, even if r for remove makes sense in C-x w
map, remove isn't part of the function name (whereas u for unhighlight is).

And lastly, the h in M-s h also makes sense.. every binding in M-s h has to
do with Highlighting.

M-s h r will Highlight Regexp
M-s h . will Highlight symbol at .
M-s h u will Un-Highlight regexp
.. and so on

I realize that these bindings changes I suggest are trivial. One can argue
that why not just rebind them in your init. But why not make them
consistent so that they are easy to adopt without confusion for anyone them
for the first time? Why should there be an extra effort to remember that r
in C-x w r removes highlight whereas r in M-s h r highlights regexp?

If and when this consistency change is made, existing users can simply
relearn them if they are used to the other way or choose them rebind to the
bindings they are already used to.

--
Kaushal Modi
On Jul 12, 2015 3:07 AM, "Vaidheeswaran C" <
vaidheeswaran.chinnaraju@gmail.com> wrote:

> On Sunday 12 July 2015 01:18 AM, Juri Linkov wrote:
>
> > And maybe we could just remove the duplicate keymap ‘C-x w’
> > altogether
>
> When this happens please make the prefix configurable, much like what
> outline minor mode does.
>
> Hide Outline Minor Mode Prefix:  @
>     State : STANDARD.
>    Prefix key to use for Outline commands in Outline minor mode. Hide
>    The value of this variable is checked as part of loading Outline mode.
>    After that, changing the prefix key requires manipulating keymaps.
>
> C-x w is what I am used to and there is NOT an emacs session where I
> DO NOT use this binding.  M-s interface and bindings still confuse me.
>
> ps: `r' stands for REMOVING highlights.
>
>

[-- Attachment #2: Type: text/html, Size: 3517 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Vaidheeswaran C]

On Sunday 12 July 2015 06:51 PM, Kaushal wrote:
> If and when this consistency change is made, existing users can simply
> relearn them if they are used to the other way or choose them rebind to the
> bindings they are already used to.

I support your proposal.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Vaidheeswaran C]

On Sunday 12 July 2015 06:51 PM, Kaushal wrote:
> If and when this consistency change is made, existing users can simply
> relearn them if they are used to the other way or choose them rebind to the
> bindings they are already used to.

I support your proposal.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Eli Zaretskii]

> From: Vaidheeswaran C <vaidheeswaran.chinnaraju@gmail.com>
> Date: Sun, 12 Jul 2015 20:18:51 +0530
> 
> On Sunday 12 July 2015 06:51 PM, Kaushal wrote:
> > If and when this consistency change is made, existing users can simply
> > relearn them if they are used to the other way or choose them rebind to the
> > bindings they are already used to.
> 
> I support your proposal.

Could you please make sure your messages are sent/received only once
each?  I see between 2 to 3 identical messages each time you post
something, both here and on help-gnu-emacs.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Vaidheeswaran C]

On Sunday 12 July 2015 08:45 PM, Eli Zaretskii wrote:
> Could you please make sure your messages are sent/received only once
> each?  I see between 2 to 3 identical messages each time you post
> something, both here and on help-gnu-emacs.

Ok.  I am on a flaky connection and sometimes Icedove leaves me confused
on whether or not it did my bidding.

Are you sure it is "each time"...?

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Eli Zaretskii]

> From: Vaidheeswaran C <vaidheeswaran.chinnaraju@gmail.com>
> Date: Sun, 12 Jul 2015 21:03:48 +0530
> 
> On Sunday 12 July 2015 08:45 PM, Eli Zaretskii wrote:
> > Could you please make sure your messages are sent/received only once
> > each?  I see between 2 to 3 identical messages each time you post
> > something, both here and on help-gnu-emacs.
> 
> Ok.  I am on a flaky connection and sometimes Icedove leaves me confused
> on whether or not it did my bidding.
> 
> Are you sure it is "each time"...?

Yes.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 637 bytes --]

In this thread, only one email, the "I support this proposal." one from
Vaidheeswaran was duplicated for me.

--
Kaushal Modi
> From: Vaidheeswaran C <vaidheeswaran.chinnaraju@gmail.com>
> Date: Sun, 12 Jul 2015 21:03:48 +0530
>
> On Sunday 12 July 2015 08:45 PM, Eli Zaretskii wrote:
> > Could you please make sure your messages are sent/received only once
> > each?  I see between 2 to 3 identical messages each time you post
> > something, both here and on help-gnu-emacs.
>
> Ok.  I am on a flaky connection and sometimes Icedove leaves me confused
> on whether or not it did my bidding.
>
> Are you sure it is "each time"...?

Yes.

[-- Attachment #2: Type: text/html, Size: 1006 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Vaidheeswaran C]

On Sunday 12 July 2015 09:55 PM, Eli Zaretskii wrote:
>> From: Vaidheeswaran C <vaidheeswaran.chinnaraju@gmail.com>
>> > Date: Sun, 12 Jul 2015 21:03:48 +0530
>> > 
>> > On Sunday 12 July 2015 08:45 PM, Eli Zaretskii wrote:
>>> > > Could you please make sure your messages are sent/received only once
>>> > > each?  I see between 2 to 3 identical messages each time you post
>>> > > something, both here and on help-gnu-emacs.
>> > 
>> > Ok.  I am on a flaky connection and sometimes Icedove leaves me confused
>> > on whether or not it did my bidding.
>> > 
>> > Are you sure it is "each time"...?
> Yes.

Goodbye to this and other sister lists.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 2649 bytes --]

Hey everyone,

It looks like the consensus is that it would be fine to remove the "C-xw"
as the same commands are also bound in the `search-map' with "M-sh" binding
prefix. The added benefit is that "C-xw" binding can be used for something
else in future.

This is the first time submitting a patch to emacs. So here it is:

From bed743a154eafbda0c716fbec4c97a58bb973a0f Mon Sep 17 00:00:00 2001
From: Kaushal Modi <kaushal.modi@gmail.com>
Date: Tue, 28 Jul 2015 21:38:37 -0400
Subject: [PATCH] Remove the "C-xw" bindings from hi-lock-map

---
 etc/NEWS        |  3 +++
 lisp/hi-lock.el | 11 +----------
 2 files changed, 4 insertions(+), 10 deletions(-)

diff --git a/etc/NEWS b/etc/NEWS
index 5bb7a00..f3a9e9d 100644
--- a/etc/NEWS
+++ b/etc/NEWS
@@ -289,6 +289,9 @@ standards.

 * Changes in Specialized Modes and Packages in Emacs 25.1

+** The C-xw bindings are removed from hi-lock.el as the same commands
+   are bound to the `search-map' in bindings.el
+
 ** Prog mode has some support for multi-mode indentation.
 See `prog-indentation-context' and `prog-widen'.

diff --git a/lisp/hi-lock.el b/lisp/hi-lock.el
index 0255585..d648e30 100644
--- a/lisp/hi-lock.el
+++ b/lisp/hi-lock.el
@@ -272,16 +272,7 @@ a library is being loaded.")
     map)
   "Menu for hi-lock mode.")

-(defvar hi-lock-map
-  (let ((map (make-sparse-keymap "Hi Lock")))
-    (define-key map "\C-xwi" 'hi-lock-find-patterns)
-    (define-key map "\C-xwl" 'highlight-lines-matching-regexp)
-    (define-key map "\C-xwp" 'highlight-phrase)
-    (define-key map "\C-xwh" 'highlight-regexp)
-    (define-key map "\C-xw." 'highlight-symbol-at-point)
-    (define-key map "\C-xwr" 'unhighlight-regexp)
-    (define-key map "\C-xwb" 'hi-lock-write-interactive-patterns)
-    map)
+(defvar hi-lock-map (make-sparse-keymap "Hi Lock")
   "Key map for hi-lock.")

 ;; Visible Functions
-- 
1.9.2


On Sun, Jul 19, 2015 at 4:00 AM Vaidheeswaran C <
vaidheeswaran.chinnaraju@gmail.com> wrote:

> On Sunday 12 July 2015 09:55 PM, Eli Zaretskii wrote:
> >> From: Vaidheeswaran C <vaidheeswaran.chinnaraju@gmail.com>
> >> > Date: Sun, 12 Jul 2015 21:03:48 +0530
> >> >
> >> > On Sunday 12 July 2015 08:45 PM, Eli Zaretskii wrote:
> >>> > > Could you please make sure your messages are sent/received only
> once
> >>> > > each?  I see between 2 to 3 identical messages each time you post
> >>> > > something, both here and on help-gnu-emacs.
> >> >
> >> > Ok.  I am on a flaky connection and sometimes Icedove leaves me
> confused
> >> > on whether or not it did my bidding.
> >> >
> >> > Are you sure it is "each time"...?
> > Yes.
>
> Goodbye to this and other sister lists.
>
>
>
>

[-- Attachment #2: Type: text/html, Size: 4048 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Juri Linkov]

> It looks like the consensus is that it would be fine to remove the "C-xw"
> as the same commands are also bound in the `search-map' with "M-sh" binding
> prefix. The added benefit is that "C-xw" binding can be used for something
> else in future.

Do we have a 2-step deprecation process where the first step
is to warn the users about the planned changes by issuing
a message like “use the new key instead” like we do with
‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Stefan Monnier]

> Do we have a 2-step deprecation process where the first step
> is to warn the users about the planned changes by issuing
> a message like “use the new key instead” like we do with
> ‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?

No, for key-bindings we just make the change, since we don't pay nearly
as much attention to breaking "user compatibility" (which can be fixed
with a define-key in ~/.emacs).
OTOH if the change ends up breaking third-party packages it might be
more problematic.


        Stefan

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 1327 bytes --]

Juri,

I thought quite some bit on how we can add in a warning for key binding
change. But as Stefan says, it is probably not worth it.

First of, we would need to create a wrapper function or a wrapper macro to
create wrapper functions for all the commands that we want to phase out
from being bound with the "C-x w" prefix map. The wrapper function will
call the wrapped function and then issue a warning/message in the echo area
saying that that function is already bound to "M-s h" prefix and will no
longer be bound to "C-x w" prefix in future. After that, in a future emacs
version, the final step to unbind those "C-x w" bindings will remain.

On Thu, Jul 30, 2015 at 6:48 PM Stefan Monnier <monnier@iro.umontreal.ca>
wrote:

> > Do we have a 2-step deprecation process where the first step
> > is to warn the users about the planned changes by issuing
> > a message like “use the new key instead” like we do with
> > ‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?
>
> No, for key-bindings we just make the change, since we don't pay nearly
> as much attention to breaking "user compatibility" (which can be fixed
> with a define-key in ~/.emacs).
> OTOH if the change ends up breaking third-party packages it might be
> more problematic.
>
>
>         Stefan
>

[-- Attachment #2: Type: text/html, Size: 1701 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 1479 bytes --]

@Juri So what did you think?

On Thu, Jul 30, 2015 at 9:32 PM Kaushal <kaushal.modi@gmail.com> wrote:

> Juri,
>
> I thought quite some bit on how we can add in a warning for key binding
> change. But as Stefan says, it is probably not worth it.
>
> First of, we would need to create a wrapper function or a wrapper macro to
> create wrapper functions for all the commands that we want to phase out
> from being bound with the "C-x w" prefix map. The wrapper function will
> call the wrapped function and then issue a warning/message in the echo area
> saying that that function is already bound to "M-s h" prefix and will no
> longer be bound to "C-x w" prefix in future. After that, in a future emacs
> version, the final step to unbind those "C-x w" bindings will remain.
>
> On Thu, Jul 30, 2015 at 6:48 PM Stefan Monnier <monnier@iro.umontreal.ca>
> wrote:
>
>> > Do we have a 2-step deprecation process where the first step
>> > is to warn the users about the planned changes by issuing
>> > a message like “use the new key instead” like we do with
>> > ‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?
>>
>> No, for key-bindings we just make the change, since we don't pay nearly
>> as much attention to breaking "user compatibility" (which can be fixed
>> with a define-key in ~/.emacs).
>> OTOH if the change ends up breaking third-party packages it might be
>> more problematic.
>>
>>
>>         Stefan
>>
>

[-- Attachment #2: Type: text/html, Size: 2073 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Juri Linkov]

> @Juri So what did you think?

Stefan says, it is not worth it, if it is not breaking third-party packages.

> On Thu, Jul 30, 2015 at 9:32 PM Kaushal <kaushal.modi@gmail.com> wrote:
>> I thought quite some bit on how we can add in a warning for key binding
>> change. But as Stefan says, it is probably not worth it.
>>
>> First of, we would need to create a wrapper function or a wrapper macro to
>> create wrapper functions for all the commands that we want to phase out
>> from being bound with the "C-x w" prefix map. The wrapper function will
>> call the wrapped function and then issue a warning/message in the echo area
>> saying that that function is already bound to "M-s h" prefix and will no
>> longer be bound to "C-x w" prefix in future. After that, in a future emacs
>> version, the final step to unbind those "C-x w" bindings will remain.
>>
>> On Thu, Jul 30, 2015 at 6:48 PM Stefan Monnier <monnier@iro.umontreal.ca>
>> wrote:
>>
>>> > Do we have a 2-step deprecation process where the first step
>>> > is to warn the users about the planned changes by issuing
>>> > a message like “use the new key instead” like we do with
>>> > ‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?
>>>
>>> No, for key-bindings we just make the change, since we don't pay nearly
>>> as much attention to breaking "user compatibility" (which can be fixed
>>> with a define-key in ~/.emacs).
>>> OTOH if the change ends up breaking third-party packages it might be
>>> more problematic.
>>>
>>>
>>>         Stefan

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal]

[-- Attachment #1: Type: text/plain, Size: 1980 bytes --]

Great! Then should the patch be applied? I haven't heard from anyone that
this change will break a 3rd party package.

I don't have commit rights. So either you, Stefan or someone else having
the commit rights will have to apply the patch.

Thanks.

On Wed, Aug 19, 2015 at 6:28 PM Juri Linkov <juri@linkov.net> wrote:

> > @Juri So what did you think?
>
> Stefan says, it is not worth it, if it is not breaking third-party
> packages.
>
> > On Thu, Jul 30, 2015 at 9:32 PM Kaushal <kaushal.modi@gmail.com> wrote:
> >> I thought quite some bit on how we can add in a warning for key binding
> >> change. But as Stefan says, it is probably not worth it.
> >>
> >> First of, we would need to create a wrapper function or a wrapper macro
> to
> >> create wrapper functions for all the commands that we want to phase out
> >> from being bound with the "C-x w" prefix map. The wrapper function will
> >> call the wrapped function and then issue a warning/message in the echo
> area
> >> saying that that function is already bound to "M-s h" prefix and will no
> >> longer be bound to "C-x w" prefix in future. After that, in a future
> emacs
> >> version, the final step to unbind those "C-x w" bindings will remain.
> >>
> >> On Thu, Jul 30, 2015 at 6:48 PM Stefan Monnier <
> monnier@iro.umontreal.ca>
> >> wrote:
> >>
> >>> > Do we have a 2-step deprecation process where the first step
> >>> > is to warn the users about the planned changes by issuing
> >>> > a message like “use the new key instead” like we do with
> >>> > ‘define-obsolete-function-alias’ and ‘make-obsolete-variable’?
> >>>
> >>> No, for key-bindings we just make the change, since we don't pay nearly
> >>> as much attention to breaking "user compatibility" (which can be fixed
> >>> with a define-key in ~/.emacs).
> >>> OTOH if the change ends up breaking third-party packages it might be
> >>> more problematic.
> >>>
> >>>
> >>>         Stefan
>

[-- Attachment #2: Type: text/html, Size: 2722 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal Modi]

[-- Attachment #1: Type: text/plain, Size: 1535 bytes --]

Hi Juri,

Thanks for mentioning that this bindings topic still needs to be resolved
in the other debbugs thread (
http://debbugs.gnu.org/cgi/bugreport.cgi?bug=22520 ).

Here are my reasons to remove the "C-x w" map:
- The bindings are inconsistent between the "M-s h" and "C-s w" maps.
- In addition to confusion, the bindings are also duplicated in two prefix
maps.
- Also if we remove this prefix map today, we will have lesser number of
disappointed users than when and if it is removed in the future.

Here is the piece from earlier in this thread where I compared the
inconsistent bindings (the line numbers can be ignored as they might not be
the same today):

hi-lock.el 277: (define-key map "\C-xwi" 'hi-lock-find-patterns)
bindings.el 935: (define-key search-map "hf" 'hi-lock-find-patterns)

hi-lock.el 282: (define-key map "\C-xwr" 'unhighlight-regexp)
bindings.el 934: (define-key search-map "hu" 'unhighlight-regexp)

hi-lock.el 280: (define-key map "\C-xwh" 'highlight-regexp)
bindings.el 930: (define-key search-map "hr" 'highlight-regexp)

hi-lock.el 283: (define-key map "\C-xwb"
'hi-lock-write-interactive-patterns)
bindings.el 936: (define-key search-map "hw"
'hi-lock-write-interactive-patterns)

The source of this discussion is because I kept getting confused if I need
to do "M-s h r" or "C-x w r" to highlight a regexp. So after a point, I
bound C-x w to nil to solve the problem.

That made me think that this topic should also be brought up in this thread
and have this fixed directly in the source.

Kaushal

[-- Attachment #2: Type: text/html, Size: 1960 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Juri Linkov]

> Here are my reasons to remove the "C-x w" map:
> - The bindings are inconsistent between the "M-s h" and "C-s w" maps.
> - In addition to confusion, the bindings are also duplicated in two prefix
> maps.
> - Also if we remove this prefix map today, we will have lesser number of
> disappointed users than when and if it is removed in the future.
>
> Here is the piece from earlier in this thread where I compared the
> inconsistent bindings (the line numbers can be ignored as they might not be
> the same today):
>
> hi-lock.el 277: (define-key map "\C-xwi" 'hi-lock-find-patterns)
> bindings.el 935: (define-key search-map "hf" 'hi-lock-find-patterns)
>
> hi-lock.el 282: (define-key map "\C-xwr" 'unhighlight-regexp)
> bindings.el 934: (define-key search-map "hu" 'unhighlight-regexp)
>
> hi-lock.el 280: (define-key map "\C-xwh" 'highlight-regexp)
> bindings.el 930: (define-key search-map "hr" 'highlight-regexp)
>
> hi-lock.el 283: (define-key map "\C-xwb"
> 'hi-lock-write-interactive-patterns)
> bindings.el 936: (define-key search-map "hw"
> 'hi-lock-write-interactive-patterns)
>
> The source of this discussion is because I kept getting confused if I need
> to do "M-s h r" or "C-x w r" to highlight a regexp. So after a point, I
> bound C-x w to nil to solve the problem.
>
> That made me think that this topic should also be brought up in this thread
> and have this fixed directly in the source.

Maybe conducting a poll will help to see how many users will be disappointed.

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[John Wiegley]

>>>>> Juri Linkov <juri@linkov.net> writes:

>> The source of this discussion is because I kept getting confused if I need
>> to do "M-s h r" or "C-x w r" to highlight a regexp. So after a point, I
>> bound C-x w to nil to solve the problem.
>> 
>> That made me think that this topic should also be brought up in this thread
>> and have this fixed directly in the source.

> Maybe conducting a poll will help to see how many users will be
> disappointed.

How long have we had both?

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Kaushal Modi]

[-- Attachment #1: Type: text/plain, Size: 645 bytes --]

On Thu, Feb 4, 2016 at 8:48 PM John Wiegley <jwiegley@gmail.com> wrote:

> How long have we had both?


The "C-x w" bindings have been since the origin! of hi-lock.el on
2000-08-02, added by Gerd Moellmann (
http://git.savannah.gnu.org/cgit/emacs.git/commit/?h=emacs-25&id=abb2db1cf3b6505afedcafb51e6f8dca36585b9c
 ).

The "M-s h" bindings came later on 2008-06-25, added by Juri (
http://git.savannah.gnu.org/cgit/emacs.git/commit/?h=emacs-25&id=56d62ee1fba23e69035e5146a12323cdfb60ea9b
 ).

In another thread on emacs-devel, I learnt that Alan Mackenzie is using the
"C-x w" bindings. So I don't know .. what would be a good way to poll this?

[-- Attachment #2: Type: text/html, Size: 1288 bytes --]

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[John Wiegley]

>>>>> Kaushal Modi <kaushal.modi@gmail.com> writes:

> In another thread on emacs-devel, I learnt that Alan Mackenzie is using the
> "C-x w" bindings. So I don't know .. what would be a good way to poll this?

I have a mind to deprecate the C-x w binding then, for removal in a future
version. M-s h makes more sense, and you say, is always available.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[Alan Mackenzie]

Hello, John.

On Sun, Feb 07, 2016 at 12:57:26PM -0500, John Wiegley wrote:
> >>>>> Kaushal Modi <kaushal.modi@gmail.com> writes:

> > In another thread on emacs-devel, I learnt that Alan Mackenzie is using the
> > "C-x w" bindings. So I don't know .. what would be a good way to poll this?

> I have a mind to deprecate the C-x w binding then, for removal in a future
> version. M-s h makes more sense, and you say, is always available.

Just for the record, I'd have no great objection to the C-x w bindings
being deprecated, then removed.  It would be a little inconvenient
having to learn new bindings, but no more than a little.

> -- 
> John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
> http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: Suggestion to have highlight related bindings consistent between search-map and hi-lock-map

[David Koppelman]

The change is fine with me, the consistency makes sense. (I don't
use the default hi-lock bindings anyway.)

David


Kaushal <kaushal.modi@gmail.com> writes:

> Hi,
> I was trying to using the binding for highlight-regexp and I got confused
> if the prefix was "C-x w r" or "M-s h r". I "grepping" (actually agging),
> I found that below are the ONLY bindings that begin with "C-x w .." and
> "M-s h .." (with the exact same number of bindings too).
> Can we make the bindings consistent so that it doesn't matter is the user
> is using "C-x w" or "M-s h"?
> kmA^2~/downloads/:git/emacs/lisp> ag 'define.*C\-xw'
> hi-lock.el
> 277: A  A (define-key map "\C-xwi" 'hi-lock-find-patterns)
> 278: A  A (define-key map "\C-xwl" 'highlight-lines-matching-regexp)
> 279: A  A (define-key map "\C-xwp" 'highlight-phrase)
> 280: A  A (define-key map "\C-xwh" 'highlight-regexp)
> 281: A  A (define-key map "\C-xw." 'highlight-symbol-at-point)
> 282: A  A (define-key map "\C-xwr" 'unhighlight-regexp)
> 283: A  A (define-key map "\C-xwb" 'hi-lock-write-interactive-patterns)
> kmA^2~/downloads/:git/emacs/lisp> ag 'define.*search-map\s+"h'
> bindings.el
> 935:(define-key search-map "hf" A  'hi-lock-find-patterns)
> 932:(define-key search-map "hl" A  'highlight-lines-matching-regexp)
> 931:(define-key search-map "hp" A  'highlight-phrase)
> 930:(define-key search-map "hr" A  'highlight-regexp)
> 933:(define-key search-map "h." A  'highlight-symbol-at-point)
> 934:(define-key search-map "hu" A  'unhighlight-regexp)
> 936:(define-key search-map "hw" A  'hi-lock-write-interactive-patterns)
> a**Some bindings are consistent but some are not. Here are the
> inconsistent ones:
> hi-lock.el A 277: (define-key map A  A  A  A "\C-xwi"
> 'hi-lock-find-patterns)
> bindings.el 935: (define-key search-map A  A  "hf" 'hi-lock-find-patterns)
> hi-lock.el A 282: (define-key map A  A  A  A "\C-xwr" 'unhighlight-regexp)
> bindings.el 934: (define-key search-map A  A  "hu" 'unhighlight-regexp)
> hi-lock.el A 280: (define-key map A  A  A  A "\C-xwh" 'highlight-regexp)
> bindings.el 930: (define-key search-map A  A  "hr" 'highlight-regexp)
> hi-lock.el A 283: (define-key map A  A  A  A "\C-xwb"
> 'hi-lock-write-interactive-patterns)
> bindings.el 936: (define-key search-map A  A  "hw"
> 'hi-lock-write-interactive-patterns)
> *** Here are my suggested consistent bindings to fix the inconsistent
> ones: ***
> - Currently "C-x w f" is undefined; "f" for find patterns
> hi-lock.el A 277: (define-key map A  A  A  A "\C-xwf"
> 'hi-lock-find-patterns) ; changed
> bindings.el 935: (define-key search-map A  A  "hf" 'hi-lock-find-patterns)
> ; no change
> - Currently "C-x w u" is undefined; "u" for unhighlight
> hi-lock.el A 282: (define-key map A  A  A  A "\C-xwu" 'unhighlight-regexp)
> ; changed
> bindings.el 934: (define-key search-map A  A  "hu" 'unhighlight-regexp) ;
> no change
> - "C-x w r" which was earlier bound to unhighlight-regexp should now be
> bound to highlight-regexp; "r" for regexp
> hi-lock.el A 280: (define-key map A  A  A  A "\C-xwr" 'highlight-regexp) ;
> changed
> bindings.el 930: (define-key search-map A  A  "hr" 'highlight-regexp) ; no
> change
> - Currently "C-x w w" is undefined; "w" for write interactive patterns
> hi-lock.el A 283: (define-key map A  A  A  A "\C-xww"
> 'hi-lock-write-interactive-patterns) ; changed
> bindings.el 936: (define-key search-map A  A  "hw"
> 'hi-lock-write-interactive-patterns) ; no change
> Well.. turns out only 4 bindings need to change in hi-lock.el to make
> these consistent.
> a**
> |------------+-------------+---------------------+------------------------------------|
> | search-map | hi-lock-map | last key in binding | function A  A  A  A  A 
> A  A  A  A  A  A  A  A  |
> |------------+-------------+---------------------+------------------------------------|
> | M-s h A  A  A | C-x w A  A  A  | l A  A  A  A  A  A  A  A  A  |
> highlight-lines-matching-regexp A  A |
> | M-s h A  A  A | C-x w A  A  A  | p A  A  A  A  A  A  A  A  A  |
> highlight-phrase A  A  A  A  A  A  A  A  A  |
> | M-s h A  A  A | C-x w A  A  A  | . A  A  A  A  A  A  A  A  A  |
> highlight-symbol-at-point A  A  A  A  A |
> |------------+-------------+---------------------+------------------------------------|
> | M-s h A  A  A | C-x w A  A  A  | f A  A  A  A  A  A  A  A  A  |
> hi-lock-find-patterns A  A  A  A  A  A  A |
> | M-s h A  A  A | C-x w A  A  A  | u A  A  A  A  A  A  A  A  A  |
> unhighlight-regexp A  A  A  A  A  A  A  A  |
> | M-s h A  A  A | C-x w A  A  A  | r A  A  A  A  A  A  A  A  A  |
> highlight-regexp A  A  A  A  A  A  A  A  A  |
> | M-s h A  A  A | C-x w A  A  A  | w A  A  A  A  A  A  A  A  A  |
> hi-lock-write-interactive-patterns |
> |------------+-------------+---------------------+------------------------------------|
> --
> Kaushal Modi