Re: emacs-27 027da65: Fix display of minibuffer prompt in ido.el

Re: emacs-27 027da65: Fix display of minibuffer prompt in ido.el

[Dmitry Gutov]

Hi Eli,

On 12.02.2020 21:41, Eli Zaretskii wrote:
> +@item minibuffer-message
> +@kindex minibuffer-message @r{(text property)}
> +This text property tells where to display temporary messages in an
> +active minibuffer.

TBH, I'm wary of documenting this variable in the manual, since we might 
prefer to do away with it later, if a sufficiently stable fix for the 
overlay-based solution is found.

Re: emacs-27 027da65: Fix display of minibuffer prompt in ido.el

[Stefan Monnier]

>> +@item minibuffer-message
>> +@kindex minibuffer-message @r{(text property)}
>> +This text property tells where to display temporary messages in an
>> +active minibuffer.
>
> TBH, I'm wary of documenting this variable in the manual, since we might
> prefer to do away with it later, if a sufficiently stable fix for the 
> overlay-based solution is found.

I tend to agree.  This also suggests it should be renamed to
`minibuffer--message`.


        Stefan

Re: emacs-27 027da65: Fix display of minibuffer prompt in ido.el

[Eli Zaretskii]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org,  Eli Zaretskii <eliz@gnu.org>
> Date: Thu, 13 Feb 2020 18:39:40 -0500
> 
> >> +@item minibuffer-message
> >> +@kindex minibuffer-message @r{(text property)}
> >> +This text property tells where to display temporary messages in an
> >> +active minibuffer.
> >
> > TBH, I'm wary of documenting this variable in the manual, since we might
> > prefer to do away with it later, if a sufficiently stable fix for the 
> > overlay-based solution is found.
> 
> I tend to agree.

I don't, at least not for the reasons described by Dmitry.  Display of
overlay strings is quite messy, especially at EOB; it already causes
quite a few complex and sometimes inelegant code fragments in the
display engine, in particular related to cursor positioning.  So
anything that allows us to avoid using overlay strings, and simply
insert text into a buffer, is a win from my POV.

And using a special text property to guide display (and other
important features) is a technique we use elsewhere in Emacs.
Granted, we don't always document such text properties, but in this
case, we are talking about an area where quite a few external packages
tend to offer all kinds of add-ons and improvements, so I thought it
was important to document it, so that they could play well with the
new set-message-function feature.

That said, if you (or someone else) have other good reasons we would
like not to publish this property, which are stronger than the
considerations that led me to document it, I'm all ears.

Thanks.

Re: emacs-27 027da65: Fix display of minibuffer prompt in ido.el

[Dmitry Gutov]

On 14.02.2020 10:08, Eli Zaretskii wrote:
> That said, if you (or someone else) have other good reasons we would
> like not to publish this property, which are stronger than the
> considerations that led me to document it, I'm all ears.

I don't think there are many external packages left to worry about, 
especially if we take care of Ivy, from a recent bug report (which might 
be entirely unrelated to minibuffer-message).

But continuing to support the new text property only takes a few lines, 
so it's not the end of the world either.