In Emacs commit f596f0db82c0b1ff3fe8e8f1d8b07d2fe7504ab6, from Nov 2021, the function `indent-according-to-mode' was given an optional inhibit-widen argument. That argument being passed causes orgalist's advice to fail, as the lambda doesn't accept any additional arguments. One way to fix it would be like that: (unless (advice-member-p 'orgalist-fix-bug:31361 'indent-according-to-mode) (advice-add 'indent-according-to-mode :around (lambda (old &optional inhibit-widen) "Workaround bug#31361." (or (orgalist--indent-line) (let ((indent-line-function (advice--cd*r indent-line-function))) (funcall old inhibit-widen)))) '((name . orgalist-fix-bug:31361))))) Or I suppose a more future-proof approach might be to use a &rest and then `apply' instead of `funcall'. Eric
Eric Abrahamsen <eric@ericabrahamsen.net> writes:
> In Emacs commit f596f0db82c0b1ff3fe8e8f1d8b07d2fe7504ab6, from Nov 2021,
> the function `indent-according-to-mode' was given an optional
> inhibit-widen argument. That argument being passed causes orgalist's
> advice to fail, as the lambda doesn't accept any additional arguments.
> One way to fix it would be like that:
>
> (unless (advice-member-p 'orgalist-fix-bug:31361 'indent-according-to-mode)
> (advice-add 'indent-according-to-mode
> :around (lambda (old &optional inhibit-widen)
> "Workaround bug#31361."
> (or (orgalist--indent-line)
> (let ((indent-line-function
> (advice--cd*r indent-line-function)))
> (funcall old inhibit-widen))))
> '((name . orgalist-fix-bug:31361)))))
>
> Or I suppose a more future-proof approach might be to use a &rest and
> then `apply' instead of `funcall'.
>
A better solution would probably be to fix this without using
add-advice. While advice can be a useful escape hatch, it really is best
avoided, especially given that it doesn't always play nice with lexical
binding. I note this one is also calling an undocumented internal
function.
Tim Cross <theophilusx@gmail.com> writes:
> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>
>> In Emacs commit f596f0db82c0b1ff3fe8e8f1d8b07d2fe7504ab6, from Nov 2021,
>> the function `indent-according-to-mode' was given an optional
>> inhibit-widen argument. That argument being passed causes orgalist's
>> advice to fail, as the lambda doesn't accept any additional arguments.
>> One way to fix it would be like that:
>>
>> (unless (advice-member-p 'orgalist-fix-bug:31361 'indent-according-to-mode)
>> (advice-add 'indent-according-to-mode
>> :around (lambda (old &optional inhibit-widen)
>> "Workaround bug#31361."
>> (or (orgalist--indent-line)
>> (let ((indent-line-function
>> (advice--cd*r indent-line-function)))
>> (funcall old inhibit-widen))))
>> '((name . orgalist-fix-bug:31361)))))
>>
>> Or I suppose a more future-proof approach might be to use a &rest and
>> then `apply' instead of `funcall'.
>>
>
> A better solution would probably be to fix this without using
> add-advice. While advice can be a useful escape hatch, it really is best
> avoided, especially given that it doesn't always play nice with lexical
> binding. I note this one is also calling an undocumented internal
> function.
In principle I quite agree! But orgalist is basically built on top of
add-function/advice-add, so I didn't think that was an option. And
also assumed that, given what Orgalist is trying to do, there isn't a
cleaner solution right now.
Eric Abrahamsen <eric@ericabrahamsen.net> writes:
> Tim Cross <theophilusx@gmail.com> writes:
>
>> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>>
>>> In Emacs commit f596f0db82c0b1ff3fe8e8f1d8b07d2fe7504ab6, from Nov 2021,
>>> the function `indent-according-to-mode' was given an optional
>>> inhibit-widen argument. That argument being passed causes orgalist's
>>> advice to fail, as the lambda doesn't accept any additional arguments.
>>> One way to fix it would be like that:
>>>
>>> (unless (advice-member-p 'orgalist-fix-bug:31361 'indent-according-to-mode)
>>> (advice-add 'indent-according-to-mode
>>> :around (lambda (old &optional inhibit-widen)
>>> "Workaround bug#31361."
>>> (or (orgalist--indent-line)
>>> (let ((indent-line-function
>>> (advice--cd*r indent-line-function)))
>>> (funcall old inhibit-widen))))
>>> '((name . orgalist-fix-bug:31361)))))
>>>
>>> Or I suppose a more future-proof approach might be to use a &rest and
>>> then `apply' instead of `funcall'.
>>>
>>
>> A better solution would probably be to fix this without using
>> add-advice. While advice can be a useful escape hatch, it really is best
>> avoided, especially given that it doesn't always play nice with lexical
>> binding. I note this one is also calling an undocumented internal
>> function.
>
> In principle I quite agree! But orgalist is basically built on top of
> add-function/advice-add, so I didn't think that was an option. And
> also assumed that, given what Orgalist is trying to do, there isn't a
> cleaner solution right now.
What do you think, Nicolas? Can I patch this to use &rest and `apply'?