bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

[Juri Linkov]

>> +(defcustom next-error-find-buffer-function nil
>
> Why make it a defcustom rather than a defvar?

I changed it to defvar in the following patch.

>> +   (when next-error-find-buffer-function
>> +     (funcall next-error-find-buffer-function avoid-current
>> +                                              extra-test-inclusive
>> +                                              extra-test-exclusive))
>
> Could you arrange for the default value of this new *-function var not
> to be nil so we can modify it with add-function?

Then I guess the default function should return nil.  Here is a new patch:

diff --git a/lisp/simple.el b/lisp/simple.el
index 2101cfe..5413b5a 100644
--- a/lisp/simple.el
+++ b/lisp/simple.el
@@ -170,14 +170,12 @@ next-error-buffer-p
 	   (and extra-test-inclusive
 		(funcall extra-test-inclusive))))))
 
-(defcustom next-error-find-buffer-function nil
-  "Function called to find a `next-error' capable buffer."
-  :type '(choice (const :tag "Single next-error capable buffer on selected frame"
-                        next-error-buffer-on-selected-frame)
-                 (const :tag "No default" nil)
-                 (function :tag "Other function"))
-  :group 'next-error
-  :version "27.1")
+(defvar next-error-find-buffer-function 'next-error-find-buffer-function-default
+  "Function to find a `next-error' capable buffer.")
+
+(defun next-error-find-buffer-function-default (&optional avoid-current
+                                                          extra-test-inclusive
+                                                          extra-test-exclusive))
 
 (defun next-error-buffer-on-selected-frame (&optional avoid-current
                                                       extra-test-inclusive
@@ -212,10 +210,9 @@ next-error-find-buffer
 that buffer is rejected."
   (or
    ;; 1. If a customizable function returns a buffer, use it.
-   (when next-error-find-buffer-function
-     (funcall next-error-find-buffer-function avoid-current
-                                              extra-test-inclusive
-                                              extra-test-exclusive))
+   (funcall next-error-find-buffer-function avoid-current
+                                            extra-test-inclusive
+                                            extra-test-exclusive)
    ;; 2. If next-error-last-buffer is an acceptable buffer, use that.
    (if (and next-error-last-buffer
             (next-error-buffer-p next-error-last-buffer avoid-current

bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

[Stefan Monnier]

>> Could you arrange for the default value of this new *-function var not
>> to be nil so we can modify it with add-function?
> Then I guess the default function should return nil.  Here is a new patch:

Fine by me.  Tho, see comments below.


        Stefan


> +(defun next-error-find-buffer-function-default (&optional avoid-current
> +                                                          extra-test-inclusive
> +                                                          extra-test-exclusive))

This looks like a syntax error to me (function without body).  I know
it's accepted (because various parts of the byte-compiler need special
hacks to deal with it), but please always put a body (i.e. an explicit nil).

This said, I don't think you should define this function.  Instead:

> +(defvar next-error-find-buffer-function 'next-error-find-buffer-function-default

should just say

    (defvar next-error-find-buffer-function #'ignore


-- Stefan

bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

[Juri Linkov]

>> +(defun next-error-find-buffer-function-default (&optional avoid-current
>> +                                                          extra-test-inclusive
>> +                                                          extra-test-exclusive))
>
> This looks like a syntax error to me (function without body).  I know
> it's accepted (because various parts of the byte-compiler need special
> hacks to deal with it), but please always put a body (i.e. an explicit nil).

And without body there is no way to add a docstring.

>> +(defvar next-error-find-buffer-function 'next-error-find-buffer-function-default
>
> should just say
>
>     (defvar next-error-find-buffer-function #'ignore

Ah, before arriving to the function without body I mistakenly tried #'identity
(that didn't work) whereas I actually intended to try #'ignore.

But the question that I really don't understand is what to do with the function
next-error-buffer-on-selected-frame that now will have no reference in code,
no use by default.  How the users are supposed to know that it's possible
to put add-function with this function in the init file?  Should we document
this in the docstring of next-error-buffer-on-selected-frame or the docstring
of next-error-find-buffer-function together with examples of using add-function?

bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

[Stefan Monnier]

> Ah, before arriving to the function without body I mistakenly tried #'identity
> (that didn't work) whereas I actually intended to try #'ignore.

Should we provide the K combinator?

> But the question that I really don't understand is what to do with the
> function next-error-buffer-on-selected-frame that now will have no
> reference in code, no use by default.  How the users are supposed to
> know that it's possible to put add-function with this function in the
> init file?

Ah, so that's why you had a defcustom?  Then maybe having it be
a defcustom is a good idea.


        Stefan "to tell you the truth, I haven't bothered to look at the
                code to try to understand what these things are doing,
                so I have no opinion on whether it should be a defcustom
                or not, or if should be a function defcustom rather
                than, say, a boolean defcustom"

bug#20489: [Emacs-diffs] master d48e07a: * lisp/simple.el (next-error-find-buffer-function): New defcustom.

[Juri Linkov]

>> But the question that I really don't understand is what to do with the
>> function next-error-buffer-on-selected-frame that now will have no
>> reference in code, no use by default.  How the users are supposed to
>> know that it's possible to put add-function with this function in the
>> init file?
>
> Ah, so that's why you had a defcustom?  Then maybe having it be
> a defcustom is a good idea.
>
>
>         Stefan "to tell you the truth, I haven't bothered to look at the
>                 code to try to understand what these things are doing,
>                 so I have no opinion on whether it should be a defcustom
>                 or not, or if should be a function defcustom rather
>                 than, say, a boolean defcustom"

I think what we need is to support add-function in defcustom.
Probably we already have all building blocks to create a new
widget type using a list of functions together with PLACE symbols
(e.g. ‘:around’, ‘:before’).  Then saving such customization should
apply add-function on that list of functions in the customized order.