RE: Request for pointers and advice: displaying several buffers inside a single window

[Ihor Radchenko]

> But as for your question about seeing only some (or
> all) zones, and hiding the text between them: That's
> indeed possible, but it's not about narrowing.

Got it.

> It's instead about making a set of zones (or its
> complement) _invisible_.  For this you also need
> library `isearch-prop.el'.  If you load that library
> then you can use these keys on prefix key C-x n M-=:

I missed it when reading commentary. Probably because I was mostly
looking at key bindings part. I guess that it would be helpful to add
these key bindings to **Keys** section of the commentary, similarly to
highlight.el-related bindings.

> I can see that use case.  But I'd encourage
> people to think beyond Org mode use cases for
> the kind of thing being discussed.  Being able
> to have, in effect, an indirect buffer that
> refers to multiple buffers is _much_ more
> general than any Org mode uses.

I agree that the discussed concept can be used far beyond org-mode
scope. (Otherwise, there would be little reason to discuss adding it to
Emacs core.) I just shared my personal use case (in addition to the
provided links).

> Anyway, good luck with your project.  I hope
> it will ultimately be general enough to help
> with _all_ of the various uses people have
> envisioned for acting on areas of different
> buffers in the same Emacs window / indirect
> buffer.  I wouldn't like to see something that
> is limited to something like Org mode, or is
> limited to use with multiple major modes.  I'd
> like to see something very general and flexible.

FYI, my pre-alpha implementation of synchronised text is in github
(https://github.com/yantar92/mirror-text). If you (or anyone else in the
thread) is interested, feedback is welcome.


Drew Adams <address@hidden> writes:

>> > To be clear, there's no need to cycle among zones.
>> > No need to see only one at a time. You can see all
>> > of them all of the time, as well as see the other,
>> > non-zone text, of any and all buffers.
>> >
>> > What zones.el does not offer is showing the text
>> > of more than one buffer in the same Emacs window.
>> 
>> I was not able to achieve this. Do you mean that zones
>> located more than one screen away in the same buffer
>> can be shown all together resulting in a narrowing with
>> multiple zones shown one after other without text
>> between them?
>
> No, not exactly; not a multi-narrowing.
>
> I meant only that you need not access a zone or
> make it visible only by narrowing to it.  Without
> narrowing, zones are still defined, and you can use
> them and act on them in an infinite number of ways.
> Zones are just defined areas of a buffer.
>
> But as for your question about seeing only some (or
> all) zones, and hiding the text between them: That's
> indeed possible, but it's not about narrowing.
>
> It's instead about making a set of zones (or its
> complement) _invisible_.  For this you also need
> library `isearch-prop.el'.  If you load that library
> then you can use these keys on prefix key C-x n M-=:
>
> v - isearchp-toggle-anti-zones-invisible
> V - isearchp-toggle-zones-invisible
> ~ - isearchp-toggle-complementing-domain
> d - isearchp-toggle-dimming-outside-search-area
>
> The first of these toggles hiding (making invisible)
> the complement of (the union of) the current set of
> zones, that is, the anti-zones.
>
> With a prefix arg it also toggles visibility of the
> zones themselves the other way (e.g. makes them
> visible when it makes the anti-zones invisible, and
> vice versa).
>
> Invisibility, here, is the usual Emacs invisibility
> of text.
>
> So for example, if you have a bunch of zones in a
> given buffer, and you use `C-x n M-= v', then all
> of the text outside those zones (the anti-zones)
> is made invisible (disappears).  You see the zones
> right next to each other, with no intervening text.  Repeating `C-x n M-= v' 
> shows the anti-zones again.
>
>> > Maybe consider this feedback as just letting you
>> > know that I, at least, don't quite understand
>> > what you're trying to do (or why).
>> >
>> > Emacs doesn't let you use the same window for
>> > multiple buffers, as far as I know.
>> 
>> I am trying to suggest something for displaying
>> text from multiple buffers in a single window.
>> My idea is to modify Emacs buffer internals
>> to achieve this.
>
> I understand that.
>
>> > Emacs doesn't let you use the same window for
>> > multiple buffers, as far as I know.
>> >
>> > You can finagle ways to show text from multiple
>> > buffers in the same window, e.g. by copying it.
>> > And if you do that, and you then want to edit
>> > the copies, then, yes, you'll need to then sync
>> > up the original buffers with your edits.
>> 
>> > I wonder what your reason is for wanting that?
>> > That "why" might help explain your request.
>> 
>> I think the reasons were discussed in ... and
>> popped up several times in internet...
>
> Yes, I've seen those.
>
> FWIW, I agree that being able to do arbitrary
> editing (e.g. search-&-replace) in such a
> context would be useful.
>
> That we're talking about a single window here
> in effect means we're talking about having a
> window that shows a buffer that is like an
> indirect buffer that refers to parts of
> multiple buffers.  One way to think of it
> could be as an extension of the notion of
> indirect buffer.
>
> zones.el doesn't help with this.  It does let
> you do such things for zones in the _same_
> buffer.  And it does let you work on sets of
> zones across multiple buffers.  But it doesn't
> let you work on the latter in the same window
> (i.e., the same ~indirect buffer for multiple
> buffers).
>
>> For me, the reasons are mostly related to org
>> mode. For example, it would be cool to have
>> the same org heading in multiple places (and be
>> able to edit the heading from any of those places).
>
> I can see that use case.  But I'd encourage
> people to think beyond Org mode use cases for
> the kind of thing being discussed.  Being able
> to have, in effect, an indirect buffer that
> refers to multiple buffers is _much_ more
> general than any Org mode uses.
>
> (I'm saying "indirect buffer" here, but I
> know that indirect buffers currently are
> limited.  They are in some ways too tightly
> related to their base buffers.)
>
> In a way, Org mode tries to give you some
> similar behavior, by delimiting buffer areas
> using plain-text tags (similar to what XML
> tags do).
>
> The approach taken by zones.el is better in
> this regard, I think.  It defines zones only
> by buffer and positions (which can be markers
> or Lisp-readable markers).  A zone is like an
> overlay, but it has an identifier, it can be
> Lisp-readable and persistent, and it can be
> buffer-independent (used in different buffers).
>
> Anyway, good luck with your project.  I hope
> it will ultimately be general enough to help
> with _all_ of the various uses people have
> envisioned for acting on areas of different
> buffers in the same Emacs window / indirect
> buffer.  I wouldn't like to see something that
> is limited to something like Org mode, or is
> limited to use with multiple major modes.  I'd
> like to see something very general and flexible.

-- 
Ihor Radchenko,
PhD,
Center for Advancing Materials Performance from the Nanoscale (CAMP-nano)
State Key Laboratory for Mechanical Behavior of Materials, Xi'an Jiaotong 
University, Xi'an, China
Email: address@hidden, address@hidden