bug#52314: Set message functions

bug#52314: Set message functions

[Juri Linkov]

[-- Attachment #1: Type: text/plain, Size: 860 bytes --]

Version: 29.0.50
Severity: wishlist
Tags: patch

To address several requests, there is a patch for Emacs 29 that supports:

1. inhibiting messages selectively like discussed in bug#42865, bug#44629;

2. multi-line accumulated messages like discussed on emacs-devel
   under subject "Intelligent stacking of messages in the echo area";

3. combining all them plus minibuffer messages into a pipeline:
   first inhibit-message can filter out some messages, then the second
   function can accumulate 10 old messages into a multi-line message,
   then the third function will display them in the active minibuffer.

By default, 'set-message-functions' will be '(set-minibuffer-message)'
with the current behavior, but can be customized to
'(inhibit-message set-multi-message set-minibuffer-message)'
to implement the hook-like list of functions described above:


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: set-message-functions.patch --]
[-- Type: text/x-diff, Size: 2595 bytes --]

diff --git a/lisp/minibuffer.el b/lisp/minibuffer.el
index 0a5fb72774..3eadae88db 100644
--- a/lisp/minibuffer.el
+++ b/lisp/minibuffer.el
@@ -852,7 +852,51 @@ set-minibuffer-message
         ;; was handled specially by this function.
         t))))
 
-(setq set-message-function 'set-minibuffer-message)
+(setq set-message-function 'set-message-functions)
+
+(defcustom set-message-functions '(set-minibuffer-message)
+  "List of functions to handle display of echo-area messages.
+Each function is called with one argument that is the text of a message.
+If a function returns nil, a previous message string is given to the
+next function in the list, and if the last function returns nil, the
+last message string is displayed in the echo area.
+If a function returns a string, the returned string is given to the
+next function in the list, and if the last function returns a string,
+it's displayed in the echo area.
+If a function returns any other non-nil value, no more functions are
+called from the list, and no message will be displayed in the echo area."
+  :type '(choice (const :tag "No special message handling" nil)
+                 (repeat
+                  (choice (function-item :tag "Inhibit some messages"
+                                         inhibit-message)
+                          (function-item :tag "Accumulate messages"
+                                         set-multi-message)
+                          (function-item :tag "Handle minibuffer"
+                                         set-minibuffer-message)
+                          (function :tag "Custom function"))))
+  :version "29.1")
+
+(defun set-message-functions (message)
+  (run-hook-wrapped 'set-message-functions
+                    (lambda (fun)
+                      (when (stringp message)
+                        (let ((ret (funcall fun message)))
+                          (when ret (setq message ret))))
+                      nil))
+  message)
+
+(defcustom inhibit-message-regexp nil
+  "Regexp to inhibit messages by the function `inhibit-message'."
+  :type '(choice (const :tag "Don't inhibit messages" nil)
+                 (regexp :tag "Inhibit messages that match regexp"))
+  :version "29.1")
+
+(defun inhibit-message (message)
+  "Don't display MESSAGE when it matches the regexp `inhibit-message-regexp'.
+This function is intended to be added to `set-message-functions'."
+  (or (and (stringp inhibit-message-regexp)
+           (string-match-p inhibit-message-regexp message))
+      message))
 
 (defun clear-minibuffer-message ()
   "Clear minibuffer message.

bug#52314: Set message functions

[Stefan Kangas]

Juri Linkov <juri@linkov.net> writes:

> To address several requests, there is a patch for Emacs 29 that supports:
[snip]

Cool!

> +(defcustom inhibit-message-regexp nil
> +  "Regexp to inhibit messages by the function `inhibit-message'."
> +  :type '(choice (const :tag "Don't inhibit messages" nil)
> +                 (regexp :tag "Inhibit messages that match regexp"))
> +  :version "29.1")

How about making this optionally support a list as well?  That would
make it slightly easier to customize with `M-x customize', I think,
especially once you start racking up many ignored messages.

bug#52314: Set message functions

[Juri Linkov]

>> +(defcustom inhibit-message-regexp nil
>> +  "Regexp to inhibit messages by the function `inhibit-message'."
>> +  :type '(choice (const :tag "Don't inhibit messages" nil)
>> +                 (regexp :tag "Inhibit messages that match regexp"))
>> +  :version "29.1")
>
> How about making this optionally support a list as well?  That would
> make it slightly easier to customize with `M-x customize', I think,
> especially once you start racking up many ignored messages.

Thanks, good idea, will add it in the next version of the patch.
Also this will help to add more regexps easier with e.g.:

  (add-to-list 'inhibit-message-regexps "^Mark set$")

bug#52314: Set message functions

[Lars Ingebrigtsen]

Juri Linkov <juri@linkov.net> writes:

> To address several requests, there is a patch for Emacs 29 that supports:
>
> 1. inhibiting messages selectively like discussed in bug#42865, bug#44629;
>
> 2. multi-line accumulated messages like discussed on emacs-devel
>    under subject "Intelligent stacking of messages in the echo area";
>
> 3. combining all them plus minibuffer messages into a pipeline:
>    first inhibit-message can filter out some messages, then the second
>    function can accumulate 10 old messages into a multi-line message,
>    then the third function will display them in the active minibuffer.
>
> By default, 'set-message-functions' will be '(set-minibuffer-message)'
> with the current behavior, but can be customized to
> '(inhibit-message set-multi-message set-minibuffer-message)'
> to implement the hook-like list of functions described above:

Hm...  the patch only had the defcustom and some helper functions, so
was this just to get feedback on the interface?

If so, I think it looks good -- but I'm a bit vague about what
set-multi-message would look like.

bug#52314: Set message functions

[Stefan Kangas]

tags 52314 - patch
thanks

Juri Linkov <juri@linkov.net> writes:

>>> +(defcustom inhibit-message-regexp nil
>>> +  "Regexp to inhibit messages by the function `inhibit-message'."
>>> +  :type '(choice (const :tag "Don't inhibit messages" nil)
>>> +                 (regexp :tag "Inhibit messages that match regexp"))
>>> +  :version "29.1")
>>
>> How about making this optionally support a list as well?  That would
>> make it slightly easier to customize with `M-x customize', I think,
>> especially once you start racking up many ignored messages.
>
> Thanks, good idea, will add it in the next version of the patch.
> Also this will help to add more regexps easier with e.g.:
>
>   (add-to-list 'inhibit-message-regexps "^Mark set$")

I'm removing the patch tag for now, as it seems like this is not yet
ready for installing on master.

bug#52314: Set message functions

[Juri Linkov]

>>>> +(defcustom inhibit-message-regexp nil
>>>> +  "Regexp to inhibit messages by the function `inhibit-message'."
>>>> +  :type '(choice (const :tag "Don't inhibit messages" nil)
>>>> +                 (regexp :tag "Inhibit messages that match regexp"))
>>>> +  :version "29.1")
>>>
>>> How about making this optionally support a list as well?  That would
>>> make it slightly easier to customize with `M-x customize', I think,
>>> especially once you start racking up many ignored messages.
>>
>> Thanks, good idea, will add it in the next version of the patch.
>> Also this will help to add more regexps easier with e.g.:
>>
>>   (add-to-list 'inhibit-message-regexps "^Mark set$")
>
> I'm removing the patch tag for now, as it seems like this is not yet
> ready for installing on master.

Why not?  I have been testing this patch for 1.5 years, and
everything works fine.  So it's ready for installing on master.

bug#52314: Set message functions

[Stefan Kangas]

tags 52314 + patch
thanks

Juri Linkov <juri@linkov.net> writes:

>> I'm removing the patch tag for now, as it seems like this is not yet
>> ready for installing on master.
>
> Why not?  I have been testing this patch for 1.5 years, and
> everything works fine.  So it's ready for installing on master.

I thought there was more to do?  If I misunderstood, even better.

Then you can consider this a reminder to please install this.  :-)

bug#52314: Set message functions

[Juri Linkov]

close 52314 29.0.50
thanks

>>> I'm removing the patch tag for now, as it seems like this is not yet
>>> ready for installing on master.
>>
>> Why not?  I have been testing this patch for 1.5 years, and
>> everything works fine.  So it's ready for installing on master.
>
> I thought there was more to do?  If I misunderstood, even better.
>
> Then you can consider this a reminder to please install this.  :-)

Sorry, only now I noticed that I forgot to send a newer patch
that implemented your suggestion to rename inhibit-message-regexp
to inhibit-message-regexps.  I have tested it for a long time,
so now it's pushed in the commit 9d5fc2c7eb.