GNU bug report logs - #69837
29.2; vtable-update-object only works in visible windows

Previous Next

To reply to this bug, email your comments to 69837 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Adam Porter <adam <at> alphapapa.net>
To: bug-gnu-emacs <at> gnu.org
Subject: 29.2; vtable-update-object only works in visible windows
Date: Sat, 16 Mar 2024 22:41:37 -0500

Hi,

I've discovered that `vtable-update-object' only works in buffers that 
are in visible windows.  When trying to update vtables in buffers that 
aren't, the object being updated or replaced fails to be found in the 
cache, apparently because `vtable--cache-key' uses `window-width' in its 
value (so maybe if, when the vtable buffer is not visible, the selected 
window happens to have the same width as the window in which the vtable 
buffer was previously displayed, it will work by chance).

In my case, the vtable is being updated in the background from a timer, 
so its buffer may or may not be visible.  It's preferable to call 
`vtable-update-object' to update two rows in it rather than reverting 
the whole table, because it can be relatively slow to revert the whole 
table when it's large.

I'm not sure of the best way to fix this.  Maybe the vtable's last-used 
window width could be cached and used for the cache key in case its 
buffer is not visible.  I guess that would risk displaying the vtable 
sub-optimally if the next time it's displayed its window width is 
different, but that might be worth it; in that case, the user could 
always revert the table if necessary.

I'm willing to work on a patch for this, but I'd appreciate any 
guidance, since I'm far from an expert on this library, and there are 
many nuances when it comes to buffers, windows, their attributes, etc.

Thanks,
Adam

Message #8 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Adam Porter <adam <at> alphapapa.net>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2;
 vtable-update-object only works in visible windows
Date: Sun, 17 Mar 2024 08:25:46 +0200

> Date: Sat, 16 Mar 2024 22:41:37 -0500
> From: Adam Porter <adam <at> alphapapa.net>
> 
> I've discovered that `vtable-update-object' only works in buffers that 
> are in visible windows.  When trying to update vtables in buffers that 
> aren't, the object being updated or replaced fails to be found in the 
> cache, apparently because `vtable--cache-key' uses `window-width' in its 
> value (so maybe if, when the vtable buffer is not visible, the selected 
> window happens to have the same width as the window in which the vtable 
> buffer was previously displayed, it will work by chance).

Does using with-selected-window help to solve the issue?

If not, can you show a recipe, starting from "emacs -Q", that
reproduces the problem, so we could study it in more detail?

Message #11 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Adam Porter <adam <at> alphapapa.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Sun, 17 Mar 2024 21:05:56 -0500

Hi Eli,

On 3/17/24 01:25, Eli Zaretskii wrote:
>> Date: Sat, 16 Mar 2024 22:41:37 -0500
>> From: Adam Porter <adam <at> alphapapa.net>
>>
>> I've discovered that `vtable-update-object' only works in buffers that
>> are in visible windows.  When trying to update vtables in buffers that
>> aren't, the object being updated or replaced fails to be found in the
>> cache, apparently because `vtable--cache-key' uses `window-width' in its
>> value (so maybe if, when the vtable buffer is not visible, the selected
>> window happens to have the same width as the window in which the vtable
>> buffer was previously displayed, it will work by chance).
> 
> Does using with-selected-window help to solve the issue?

Sometimes, but not reliably, because it doesn't matter whether the 
window is selected (and the buffer might not have a window, anyway). 
I'll try to give a step-by-step explanation:

1. vtable-update-object looks in the vtable's cache to find the cached 
information about the representation of the object being updated or 
replaced.

2. In the process of doing that, it calls vtable--cache-key, which 
returns a cons cell containing the frame-terminal and window-width.

3. So if the selected frame is on the same terminal, and the selected 
window has the same width, as the ones when the vtable was last 
generated, the cache key will match.  This will allow 
vtable-update-object to access the cached values and look for the object 
in them.

But if the terminal is different, or if the selected window's width is 
different, the cache key will be different, so vtable-update-object will 
fail, even when it could potentially succeed.

In my case, the vtable's buffer is (or can be) in a window that is in a 
non-current tab (using tab-bar-mode) in the same frame, so its window is 
not visible.  And so if the selected window in the current tab has the 
same width as the vtable's window, vtable-update-object may work.  But 
if the widths don't match, it will fail.

(I suppose it may also fail if the vtable's buffer's window is visible 
but has changed width since the vtable was generated, but I haven't 
tested that.)

> If not, can you show a recipe, starting from "emacs -Q", that
> reproduces the problem, so we could study it in more detail?

I had hoped to avoid writing that much code to demonstrate it.  But if 
the explanation above doesn't suffice, let me know and I will.

Thanks,
Adam

Message #14 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Adam Porter <adam <at> alphapapa.net>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Mon, 18 Mar 2024 19:05:17 +0200

> Date: Sun, 17 Mar 2024 21:05:56 -0500
> Cc: 69837 <at> debbugs.gnu.org
> From: Adam Porter <adam <at> alphapapa.net>
> 
> Hi Eli,
> 
> On 3/17/24 01:25, Eli Zaretskii wrote:
> >> Date: Sat, 16 Mar 2024 22:41:37 -0500
> >> From: Adam Porter <adam <at> alphapapa.net>
> >>
> >> I've discovered that `vtable-update-object' only works in buffers that
> >> are in visible windows.  When trying to update vtables in buffers that
> >> aren't, the object being updated or replaced fails to be found in the
> >> cache, apparently because `vtable--cache-key' uses `window-width' in its
> >> value (so maybe if, when the vtable buffer is not visible, the selected
> >> window happens to have the same width as the window in which the vtable
> >> buffer was previously displayed, it will work by chance).
> > 
> > Does using with-selected-window help to solve the issue?
> 
> Sometimes, but not reliably, because it doesn't matter whether the 
> window is selected (and the buffer might not have a window, anyway). 
> I'll try to give a step-by-step explanation:

I'm sorry, I'm not familiar with the design of vtable, so I will have
to ask you possibly stupid questions.  Apologies in advance.

> 1. vtable-update-object looks in the vtable's cache to find the cached 
> information about the representation of the object being updated or 
> replaced.

Is this cache used just to speed up something (and so if the object is
not in the cache, Emacs will just work harder to obtain the same
information), or is finding the object in the cache absolutely
necessary for working with the object?

> 2. In the process of doing that, it calls vtable--cache-key, which 
> returns a cons cell containing the frame-terminal and window-width.

If working with an object needs its window-width, does it mean objects
are tightly coupled to their windows?  If so, what happens when the
user or Emacs resizes the window? are the cached objects recomputed to
reflect that?

> 3. So if the selected frame is on the same terminal, and the selected 
> window has the same width, as the ones when the vtable was last 
> generated, the cache key will match.  This will allow 
> vtable-update-object to access the cached values and look for the object 
> in them.

This again seems to imply that an object is tightly coupled to its
window, and changes affecting the window must recompute the cached
value.  Right?

Are these keys (window-width, terminal, etc.) only used to look up the
object, or are they needed for processing the objects as well?

> > If not, can you show a recipe, starting from "emacs -Q", that
> > reproduces the problem, so we could study it in more detail?
> 
> I had hoped to avoid writing that much code to demonstrate it.  But if 
> the explanation above doesn't suffice, let me know and I will.

It doesn't have to be code, it could be a (possibly long) list of
instructions what to type.

Message #17 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Adam Porter <adam <at> alphapapa.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Tue, 19 Mar 2024 20:41:59 -0500

Hi Eli,

On 3/18/24 12:05, Eli Zaretskii wrote:

> I'm sorry, I'm not familiar with the design of vtable, so I will
> have to ask you possibly stupid questions.  Apologies in advance.
I'm not very familiar with it, either, so I will have to give you
possibly stupid answers.  :)  But I will try to give accurate ones.

>> 1. vtable-update-object looks in the vtable's cache to find the
>> cached information about the representation of the object being
>> updated or replaced.
>
> Is this cache used just to speed up something (and so if the object
> is not in the cache, Emacs will just work harder to obtain the same
> information), or is finding the object in the cache absolutely
> necessary for working with the object?
It appears that the caching is integral to the rendering of vtables.  In
`vtable-insert`, a comment explains:

    ;; We maintain a cache per screen/window width, so that we render
    ;; correctly if Emacs is open on two different screens (or the
    ;; user resizes the frame).

Then `vtable--ensure-cache` is called, which returns either an existing
cache or a newly computed one, and its return value is bound and used
when inserting the table's header line and when inserting each object's
representation.  So it seems that the rendering of objects always uses
the cached values.

>> 2. In the process of doing that, it calls vtable--cache-key, which
>> returns a cons cell containing the frame-terminal and
>> window-width.
>
> If working with an object needs its window-width, does it mean
> objects are tightly coupled to their windows?  If so, what happens
> when the user or Emacs resizes the window? are the cached objects
> recomputed to reflect that?
AFAICT, if a vtable's buffer's window is resized, nothing happens
automatically.  But if the vtable is reverted (causing it to be
completely re-rendered), the window's width is taken into account
compared to the width of the table, affecting the vtable's header line,
and possibly the widths of some columns, depending on the table's
specification (see functions `vtable--compute-width` and
`vtable--compute-widths`).  And if an individual object is updated, it
fails if the window width has changed, since the cache key misses.

I suppose the cache serves to sometimes help avoid re-rendering all the
objects in the table, and it uses the window-width as part of the cache
key because the table's buffer could be shown in multiple windows, each
of which could have a different width, requiring a different rendering
of the table (though I can't say I understand exactly how that could
work, since ISTM that rendering it for one window would display the same
rendering in the other window).

>> 3. So if the selected frame is on the same terminal, and the
>> selected window has the same width, as the ones when the vtable was
>> last generated, the cache key will match.  This will allow
>> vtable-update-object to access the cached values and look for the
>> object in them.
>
> This again seems to imply that an object is tightly coupled to its
> window, and changes affecting the window must recompute the cached
> value.  Right?
I think the way it works is roughly like this:

- An object's value(s) affect the width of the columns in its
representation's row.

- A row's columns' widths (or their total width) is compared with the
width of the window to determine how to display the last column (or the
last one visible in the window?).

- So the inverse holds as well: the window's width affects the display
of (at least one of) an object's representation's columns.

So if an object is being re-rendered, and the window's width is the
same, the cache can be used (which also seems to record the line number
at which the object is rendered, which saves from having to iterate
through the table to find the buffer line).  Otherwise, if the window's
width has changed, the cache misses; and since `vtable-update-object`
does not handle this case (it just causes an error), the object is not
updated.  This would require the application to handle the error and
revert the whole table instead.

I would suppose that, in that case, the last-used window width (if it
were known) could be used to update the object, assuming that the width
is either the same or "good enough," to avoid having to re-render the
whole table; if the worst that happened is that a column at the edge of
the table were a bit too wide or narrow compared to the ideal, that
would often be preferable to re-rendering the whole table.  (In my case,
I could have 1,000-2,000 rows in the table, in which case I would prefer
to avoid re-rendering the whole thing; leaving the columns at their
existing width is no problem, especially if some of them are already
past the width of the window--and in my case, the table is always wider
than the window.)

> Are these keys (window-width, terminal, etc.) only used to look up
> the object, or are they needed for processing the objects as well?
The window's width appears to be integral to the rendering of the
objects, both individually and when inserting the whole table at once.
The terminal is used as part of the cache key, probably because a
different display could have a different resolution (DPI), which would
make calculations on one invalid on another (I'm sure you know more
about how that works than I do).

>>> If not, can you show a recipe, starting from "emacs -Q", that
>>> reproduces the problem, so we could study it in more detail?
>>
>> I had hoped to avoid writing that much code to demonstrate it.  But
>> if the explanation above doesn't suffice, let me know and I will.
>
> It doesn't have to be code, it could be a (possibly long) list of
> instructions what to type.
In my case, I'm using vtable in my `listen` package on ELPA.  If I, e.g.
add all the tracks in my music library to a `listen-queue`, which is
rendered with vtable, it's helpful to be able to update just one track's
representation at a time (so that the currently playing track can have
an arrow next to it, and the previously playing one can have its arrow
removed), because rendering the whole table can take a few seconds.
(Smaller tables, e.g. a hundred tracks or so, render very quickly.)

If a concrete demo is still needed, let me know and I can hack something up.

Thanks,
Adam

Message #20 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Adam Porter <adam <at> alphapapa.net>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Thu, 21 Mar 2024 03:36:02 -0500

[Message part 1 (text/plain, inline)]

Hi Eli,

FWIW, I did some hacking on listen.el attempting to further understand 
and work around this problem, and I've found what seems to be a usable 
workaround (for my purposes, anyway).

The attached code defines a macro within which I call 
`listen-queue--vtable-update-object' (which merely incorporates the fix 
to `vtable-update-object' from bug#69664).  As well, I wrap 
`make-vtable' in a function that also sets two buffer-local variables to 
the values of `frame-terminal' and `window-width' at the time of the 
vtable's creation.  The macro locally overrides the functions 
`frame-terminal' and `window-width' to return those saved values.

This allows the cache key to match unconditionally, which allows the 
vtable's objects to be updated even when its buffer is not visible.

In my limited testing, it seems to work fine.  In my estimation, the 
consequences of doing this in the worst case would be that the rows for 
the updated objects might be drawn with some columns at a slightly 
incorrect width, which is easily rectified by reverting the table 
(usually bound to "g").  As well, that worst case (e.g. imagining a 
vtable whose buffer might be initially displayed on one terminal/monitor 
and later on another with different characteristics) would seem to be 
relatively rare (so for my project, it seems like an obviously good 
thing to do).

For Emacs itself, I'm not sure what the best fix would be.  I suppose a 
workaround like this could be implemented as a fallback in case the 
cache key misses; it would seem better to update the object potentially 
sub-optimally than to error and not update it at all.

Another possibility would be to ignore the frame-terminal and 
window-width in the cache key altogether (i.e. so they would always be 
assumed to be the same), but I'm sure that Lars did it this way for a 
reason, so that would seem unwise.

Let me know how you'd like me to proceed.

Thanks,
Adam

[example.el (text/x-emacs-lisp, attachment)]

Message #23 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Eli Zaretskii <eliz <at> gnu.org>
To: Adam Porter <adam <at> alphapapa.net>
Cc: 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Sat, 06 Apr 2024 14:22:53 +0300

> Date: Thu, 21 Mar 2024 03:36:02 -0500
> Cc: 69837 <at> debbugs.gnu.org
> From: Adam Porter <adam <at> alphapapa.net>
> 
> FWIW, I did some hacking on listen.el attempting to further understand 
> and work around this problem, and I've found what seems to be a usable 
> workaround (for my purposes, anyway).
> 
> The attached code defines a macro within which I call 
> `listen-queue--vtable-update-object' (which merely incorporates the fix 
> to `vtable-update-object' from bug#69664).  As well, I wrap 
> `make-vtable' in a function that also sets two buffer-local variables to 
> the values of `frame-terminal' and `window-width' at the time of the 
> vtable's creation.  The macro locally overrides the functions 
> `frame-terminal' and `window-width' to return those saved values.
> 
> This allows the cache key to match unconditionally, which allows the 
> vtable's objects to be updated even when its buffer is not visible.
> 
> In my limited testing, it seems to work fine.  In my estimation, the 
> consequences of doing this in the worst case would be that the rows for 
> the updated objects might be drawn with some columns at a slightly 
> incorrect width, which is easily rectified by reverting the table 
> (usually bound to "g").  As well, that worst case (e.g. imagining a 
> vtable whose buffer might be initially displayed on one terminal/monitor 
> and later on another with different characteristics) would seem to be 
> relatively rare (so for my project, it seems like an obviously good 
> thing to do).
> 
> For Emacs itself, I'm not sure what the best fix would be.  I suppose a 
> workaround like this could be implemented as a fallback in case the 
> cache key misses; it would seem better to update the object potentially 
> sub-optimally than to error and not update it at all.
> 
> Another possibility would be to ignore the frame-terminal and 
> window-width in the cache key altogether (i.e. so they would always be 
> assumed to be the same), but I'm sure that Lars did it this way for a 
> reason, so that would seem unwise.
> 
> Let me know how you'd like me to proceed.

I think you should install your workaround.  I don't see how it could
be worse than what we have now.

P.S. And sorry for a long silence.

Message #26 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Eli Zaretskii <eliz <at> gnu.org>
Cc: Adam Porter <adam <at> alphapapa.net>, 69837 <at> debbugs.gnu.org
Subject: Re: bug#69837: 29.2;
 vtable-update-object only works in visible windows
Date: Fri, 28 Feb 2025 19:09:29 -0800

Eli Zaretskii <eliz <at> gnu.org> writes:

>> Date: Thu, 21 Mar 2024 03:36:02 -0500
>> Cc: 69837 <at> debbugs.gnu.org
>> From: Adam Porter <adam <at> alphapapa.net>
>>
>> FWIW, I did some hacking on listen.el attempting to further understand
>> and work around this problem, and I've found what seems to be a usable
>> workaround (for my purposes, anyway).
>>
>> The attached code defines a macro within which I call
>> `listen-queue--vtable-update-object' (which merely incorporates the fix
>> to `vtable-update-object' from bug#69664).  As well, I wrap
>> `make-vtable' in a function that also sets two buffer-local variables to
>> the values of `frame-terminal' and `window-width' at the time of the
>> vtable's creation.  The macro locally overrides the functions
>> `frame-terminal' and `window-width' to return those saved values.
>>
>> This allows the cache key to match unconditionally, which allows the
>> vtable's objects to be updated even when its buffer is not visible.
>>
>> In my limited testing, it seems to work fine.  In my estimation, the
>> consequences of doing this in the worst case would be that the rows for
>> the updated objects might be drawn with some columns at a slightly
>> incorrect width, which is easily rectified by reverting the table
>> (usually bound to "g").  As well, that worst case (e.g. imagining a
>> vtable whose buffer might be initially displayed on one terminal/monitor
>> and later on another with different characteristics) would seem to be
>> relatively rare (so for my project, it seems like an obviously good
>> thing to do).
>>
>> For Emacs itself, I'm not sure what the best fix would be.  I suppose a
>> workaround like this could be implemented as a fallback in case the
>> cache key misses; it would seem better to update the object potentially
>> sub-optimally than to error and not update it at all.
>>
>> Another possibility would be to ignore the frame-terminal and
>> window-width in the cache key altogether (i.e. so they would always be
>> assumed to be the same), but I'm sure that Lars did it this way for a
>> reason, so that would seem unwise.
>>
>> Let me know how you'd like me to proceed.
>
> I think you should install your workaround.  I don't see how it could
> be worse than what we have now.
>
> P.S. And sorry for a long silence.

Adam, could you please install the workaround?  Or maybe you did
already?

Message #31 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Spencer Baugh <sbaugh <at> janestreet.com>
To: Stefan Kangas <stefankangas <at> gmail.com>
Cc: Adam Porter <adam <at> alphapapa.net>, 69837 <at> debbugs.gnu.org,
 Eli Zaretskii <eliz <at> gnu.org>
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Thu, 19 Jun 2025 17:36:36 -0400

[Message part 1 (text/plain, inline)]

Stefan Kangas <stefankangas <at> gmail.com> writes:

> Eli Zaretskii <eliz <at> gnu.org> writes:
>
>>> Date: Thu, 21 Mar 2024 03:36:02 -0500
>>> Cc: 69837 <at> debbugs.gnu.org
>>> From: Adam Porter <adam <at> alphapapa.net>
>>>
>>> FWIW, I did some hacking on listen.el attempting to further understand
>>> and work around this problem, and I've found what seems to be a usable
>>> workaround (for my purposes, anyway).
>>>
>>> The attached code defines a macro within which I call
>>> `listen-queue--vtable-update-object' (which merely incorporates the fix
>>> to `vtable-update-object' from bug#69664).  As well, I wrap
>>> `make-vtable' in a function that also sets two buffer-local variables to
>>> the values of `frame-terminal' and `window-width' at the time of the
>>> vtable's creation.  The macro locally overrides the functions
>>> `frame-terminal' and `window-width' to return those saved values.
>>>
>>> This allows the cache key to match unconditionally, which allows the
>>> vtable's objects to be updated even when its buffer is not visible.
>>>
>>> In my limited testing, it seems to work fine.  In my estimation, the
>>> consequences of doing this in the worst case would be that the rows for
>>> the updated objects might be drawn with some columns at a slightly
>>> incorrect width, which is easily rectified by reverting the table
>>> (usually bound to "g").  As well, that worst case (e.g. imagining a
>>> vtable whose buffer might be initially displayed on one terminal/monitor
>>> and later on another with different characteristics) would seem to be
>>> relatively rare (so for my project, it seems like an obviously good
>>> thing to do).
>>>
>>> For Emacs itself, I'm not sure what the best fix would be.  I suppose a
>>> workaround like this could be implemented as a fallback in case the
>>> cache key misses; it would seem better to update the object potentially
>>> sub-optimally than to error and not update it at all.
>>>
>>> Another possibility would be to ignore the frame-terminal and
>>> window-width in the cache key altogether (i.e. so they would always be
>>> assumed to be the same), but I'm sure that Lars did it this way for a
>>> reason, so that would seem unwise.
>>>
>>> Let me know how you'd like me to proceed.
>>
>> I think you should install your workaround.  I don't see how it could
>> be worse than what we have now.
>>
>> P.S. And sorry for a long silence.
>
> Adam, could you please install the workaround?  Or maybe you did
> already?

I ran into this same problem.  I think Adam's workaround is on the right
track, but the issue is present in a number of other functions in
vtable.el, not just vtable-update-object.

The root of the issue is that when we're interacting with the text of an
inserted vtable, we should use the "cache" object that was last used to
insert that vtable, rather than one based on the current selected window
and terminal.  Otherwise we will experience various inconsistencies,
e.g. inserting a new line that has different column widths from the rest
of the vtable text.

My attached patch solves this by saving the "cache" object in a text
property on the vtable text, and updating all the relevant vtable
functions to access the cache via that text property rather than by
computing a cache key from the current selected window/terminal.

Adam, could you check that this solves the problem for your use case?

[0001-Fix-implicit-usage-of-the-current-window-width-in-vt.patch (text/x-patch, inline)]

From cdde849bf78c20dbc3f5c3abba0e86171c8de188 Mon Sep 17 00:00:00 2001
From: Spencer Baugh <sbaugh <at> janestreet.com>
Date: Thu, 19 Jun 2025 17:20:12 -0400
Subject: [PATCH] Fix implicit usage of the current window-width in vtable.el

Previously, many functions in vtable.el called vtable--cache,
which computed vtable--cache-key based on the current selected
window and frame; this could cause vtable functions to fail or
misbehave if they were not called from the selected window and
frame that vtable-insert was last called in.

Now, the vtable cache is stored with the text of the vtable, so
that functions which need to interact with some vtable text can
do so reliably without having to use the same selected window
and frame.

Also, vtable-update-object has always required TABLE to be
present at point in the current buffer; now its docstring states
this.

* lisp/emacs-lisp/vtable.el (vtable--current-cache)
(vtable--cache-widths, vtable--cache-lines): Add.
(vtable-insert): Save cache in 'vtable-cache.
(vtable--ensure-cache, vtable--recompute-cache): Inline into
vtable-insert.
(vtable--widths, vtable--cache): Delete.
(vtable-update-object): Use vtable--current-cache and update
docstring.  (bug#69837)
(vtable-remove-object, vtable-insert-object): Use
vtable--current-cache and save cache in 'vtable-cache.
(vtable--sort, vtable--alter-column-width)
(vtable-previous-column, vtable-next-column): Use
vtable--current-cache.
---
 lisp/emacs-lisp/vtable.el | 111 ++++++++++++++++++++------------------
 1 file changed, 59 insertions(+), 52 deletions(-)

diff --git a/lisp/emacs-lisp/vtable.el b/lisp/emacs-lisp/vtable.el
index 00785113edb..991570b9799 100644
--- a/lisp/emacs-lisp/vtable.el
+++ b/lisp/emacs-lisp/vtable.el
@@ -282,10 +282,9 @@ vtable-update-object
   "Update OBJECT's representation in TABLE.
 If OLD-OBJECT is non-nil, replace OLD-OBJECT with OBJECT and display it.
 In either case, if the existing object is not found in the table (being
-compared with `equal'), signal an error.  Note a limitation: if TABLE's
-buffer is not in a visible window, or if its window has changed width
-since it was updated, updating the TABLE is not possible, and an error
-is signaled."
+compared with `equal'), signal an error.
+
+TABLE must be at point in the current buffer."
   (unless old-object
     (setq old-object object))
   (let* ((objects (vtable-objects table))
@@ -302,14 +301,13 @@ vtable-update-object
       (unless objects
         (error "Can't find the old object"))
       (setcar (cdr objects) object))
-    ;; Then update the cache...
-    ;; FIXME: If the table's buffer has no visible window, or if its
-    ;; width has changed since the table was updated, the cache key will
-    ;; not match and the object can't be updated.  (Bug #69837).
-    (if-let* ((line-number (seq-position (car (vtable--cache table)) old-object
-                                         (lambda (a b)
-                                           (equal (car a) b))))
-              (line (elt (car (vtable--cache table)) line-number)))
+    ;; Then update the rendered vtable in the current buffer.
+    (if-let* ((cache (vtable--current-cache))
+             (line-number (seq-position (vtable--cache-lines cache)
+                                        old-object
+                                        (lambda (a b)
+                                          (equal (car a) b))))
+             (line (elt (vtable--cache-widths cache) line-number)))
         (progn
           (setcar line object)
           (setcdr line (vtable--compute-cached-line table object))
@@ -320,10 +318,11 @@ vtable-update-object
                   (start (point)))
               (delete-line)
               (vtable--insert-line table line line-number
-                                   (nth 1 (vtable--cache table))
+                                   (vtable--cache-lines cache)
                                    (vtable--spacer table))
               (add-text-properties start (point) (list 'keymap keymap
-                                                       'vtable table))))
+                                                       'vtable table
+                                                       'vtable-cache cache))))
           ;; We may have inserted a non-numerical value into a previously
           ;; all-numerical table, so recompute.
           (vtable--recompute-numerical table (cdr line)))
@@ -335,11 +334,12 @@ vtable-remove-object
   ;; First remove from the objects.
   (setf (vtable-objects table) (delq object (vtable-objects table)))
   ;; Then adjust the cache and display.
-  (let ((cache (vtable--cache table))
-        (inhibit-read-only t))
-    (setcar cache (delq (assq object (car cache)) (car cache)))
-    (save-excursion
-      (vtable-goto-table table)
+  (save-excursion
+    (vtable-goto-table table)
+    (let ((cache (vtable--current-cache))
+          (inhibit-read-only t))
+      (setcar cache (delq (assq object (vtable--cache-lines cache))
+                          (vtable--cache-lines cache)))
       (when (vtable-goto-object object)
         (delete-line)))))
 
@@ -400,7 +400,7 @@ vtable-insert-object
     ;; Then adjust the cache and display.
     (save-excursion
       (vtable-goto-table table)
-      (let* ((cache (vtable--cache table))
+      (let* ((cache (vtable--current-cache))
              (inhibit-read-only t)
              (keymap (get-text-property (point) 'keymap))
              (ellipsis (if (vtable-ellipsis table)
@@ -408,13 +408,14 @@ vtable-insert-object
                                        'face (vtable-face table))
                          ""))
              (ellipsis-width (string-pixel-width ellipsis))
-             (elem (if location  ; This binding mirrors the binding of `pos' above.
+             (lines (vtable--cache-lines cache))
+             (elem (if location ; This binding mirrors the binding of `pos' above.
                        (if (integerp location)
-                           (nth location (car cache))
-                         (or (assq location (car cache))
-                             (and before (caar cache))))
-                     (if before (caar cache))))
-             (pos (memq elem (car cache)))
+                           (nth location lines)
+                         (or (assq location lines)
+                             (and before (car lines))))
+                     (if before (car lines))))
+             (pos (memq elem lines))
              (line (cons object (vtable--compute-cached-line table object))))
         (if (or before
                 (and pos (integerp location)))
@@ -433,13 +434,13 @@ vtable-insert-object
                     (forward-line 1)  ; Insert *after*.
                   (vtable-end-of-table)))
             ;; Otherwise, append the object.
-            (setcar cache (nconc (car cache) (list line)))
+            (setcar cache (nconc (vtable--cache-lines cache) (list line)))
             (vtable-end-of-table)))
         (let ((start (point)))
           ;; FIXME: We have to adjust colors in lines below this if we
           ;; have :row-colors.
           (vtable--insert-line table line 0
-                               (nth 1 cache) (vtable--spacer table)
+                               (vtable--cache-widths cache) (vtable--spacer table)
                                ellipsis ellipsis-width)
           (add-text-properties start (point) (list 'keymap keymap
                                                    'vtable table)))
@@ -512,15 +513,11 @@ vtable--compute-columns
 (defun vtable--spacer (table)
   (vtable--compute-width table (vtable-separator-width table)))
 
-(defun vtable--recompute-cache (table)
-  (let* ((data (vtable--compute-cache table))
-         (widths (vtable--compute-widths table data)))
-    (setf (gethash (vtable--cache-key) (slot-value table '-cache))
-          (list data widths))))
+(defun vtable--cache-widths (cache)
+  (nth 1 cache))
 
-(defun vtable--ensure-cache (table)
-  (or (vtable--cache table)
-      (vtable--recompute-cache table)))
+(defun vtable--cache-lines (cache)
+  (car cache))
 
 (defun vtable-insert (table)
   (let* ((spacer (vtable--spacer table))
@@ -533,7 +530,12 @@ vtable-insert
          ;; We maintain a cache per screen/window width, so that we render
          ;; correctly if Emacs is open on two different screens (or the
          ;; user resizes the frame).
-         (widths (nth 1 (vtable--ensure-cache table))))
+         (cache (or (gethash (vtable--cache-key) (slot-value table '-cache))
+                    (let* ((data (vtable--compute-cache table))
+                           (widths (vtable--compute-widths table data)))
+                      (setf (gethash (vtable--cache-key) (slot-value table '-cache))
+                            (list data widths)))))
+         (widths (vtable--cache-widths cache)))
     ;; Don't insert any header or header line if the user hasn't
     ;; specified the columns.
     (when (slot-value table '-has-column-spec)
@@ -546,18 +548,20 @@ vtable-insert
         (add-text-properties start (point)
                              (list 'keymap vtable-header-line-map
                                    'rear-nonsticky t
-                                   'vtable table))
+                                   'vtable table
+                                   'vtable-cache cache))
         (setq start (point))))
-    (vtable--sort table)
+    (vtable--sort table cache)
     ;; Insert the data.
     (let ((line-number 0))
-      (dolist (line (car (vtable--cache table)))
+      (dolist (line (vtable--cache-lines cache))
         (vtable--insert-line table line line-number widths spacer
                              ellipsis ellipsis-width)
         (setq line-number (1+ line-number))))
     (add-text-properties start (point)
                          (list 'rear-nonsticky t
-                               'vtable table))
+                               'vtable table
+                               'vtable-cache cache))
     (goto-char start)))
 
 (defun vtable--insert-line (table line line-number widths spacer
@@ -659,16 +663,22 @@ vtable--insert-line
 (defun vtable--cache-key ()
   (cons (frame-terminal) (window-width)))
 
-(defun vtable--cache (table)
-  (gethash (vtable--cache-key) (slot-value table '-cache)))
+(defun vtable--current-cache ()
+  "Return the current cache for the table at point.
+
+In `vtable-insert', the lines and widths of the vtable text are computed
+based on the current selected frame and window and stored in a cache.
+Subsequent interaction with the text of the vtable should use that cache
+via this function rather than by calling `vtable--cache-key' to look up
+the cache."
+  (get-text-property (point) 'vtable-cache))
 
 (defun vtable--clear-cache (table)
   (setf (gethash (vtable--cache-key) (slot-value table '-cache)) nil))
 
-(defun vtable--sort (table)
+(defun vtable--sort (table cache)
   (pcase-dolist (`(,index . ,direction) (vtable-sort-by table))
-    (let ((cache (vtable--cache table))
-          (numerical (vtable-column--numerical
+    (let ((numerical (vtable-column--numerical
                       (elt (vtable-columns table) index)))
           (numcomp (if (eq direction 'descend)
                        #'> #'<))
@@ -971,9 +981,6 @@ vtable-revert
     (when column
       (vtable-goto-column column))))
 
-(defun vtable--widths (table)
-  (nth 1 (vtable--ensure-cache table)))
-
 ;;; Commands.
 
 (defvar-keymap vtable-header-mode-map
@@ -998,7 +1005,7 @@ vtable-narrow-current-column
                                 (- (* (vtable--char-width table) (or n 1))))))
 
 (defun vtable--alter-column-width (table column delta)
-  (let ((widths (vtable--widths table)))
+  (let ((widths (vtable--cache-widths (vtable--current-cache))))
     (setf (aref widths column)
           (max (* (vtable--char-width table) 2)
                (+ (aref widths column) delta)))
@@ -1020,14 +1027,14 @@ vtable-previous-column
   (interactive)
   (vtable-goto-column
    (max 0 (1- (or (vtable-current-column)
-                  (length (vtable--widths (vtable-current-table))))))))
+                  (length (vtable--cache-widths (vtable--current-cache))))))))
 
 (defun vtable-next-column ()
   "Go to the next column."
   (interactive)
   (when (vtable-current-column)
     (vtable-goto-column
-     (min (1- (length (vtable--widths (vtable-current-table))))
+     (min (1- (length (vtable--cache-widths (vtable--current-cache))))
           (1+ (vtable-current-column))))))
 
 (defun vtable-revert-command ()
-- 
2.39.3

Message #34 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Stéphane Marks <shipmints <at> gmail.com>
To: Spencer Baugh <sbaugh <at> janestreet.com>
Cc: Adam Porter <adam <at> alphapapa.net>, 69837 <at> debbugs.gnu.org,
 Eli Zaretskii <eliz <at> gnu.org>, Stefan Kangas <stefankangas <at> gmail.com>
Subject: Re: bug#69837: 29.2;
 vtable-update-object only works in visible windows
Date: Thu, 19 Jun 2025 17:41:03 -0400

[Message part 1 (text/plain, inline)]

On Thu, Jun 19, 2025 at 5:37 PM Spencer Baugh via Bug reports for GNU
Emacs, the Swiss army knife of text editors <bug-gnu-emacs <at> gnu.org> wrote:

> Stefan Kangas <stefankangas <at> gmail.com> writes:
>
> > Eli Zaretskii <eliz <at> gnu.org> writes:
> >
> >>> Date: Thu, 21 Mar 2024 03:36:02 -0500
> >>> Cc: 69837 <at> debbugs.gnu.org
> >>> From: Adam Porter <adam <at> alphapapa.net>
> >>>
> >>> FWIW, I did some hacking on listen.el attempting to further understand
> >>> and work around this problem, and I've found what seems to be a usable
> >>> workaround (for my purposes, anyway).
> >>>
> >>> The attached code defines a macro within which I call
> >>> `listen-queue--vtable-update-object' (which merely incorporates the fix
> >>> to `vtable-update-object' from bug#69664).  As well, I wrap
> >>> `make-vtable' in a function that also sets two buffer-local variables
> to
> >>> the values of `frame-terminal' and `window-width' at the time of the
> >>> vtable's creation.  The macro locally overrides the functions
> >>> `frame-terminal' and `window-width' to return those saved values.
> >>>
> >>> This allows the cache key to match unconditionally, which allows the
> >>> vtable's objects to be updated even when its buffer is not visible.
> >>>
> >>> In my limited testing, it seems to work fine.  In my estimation, the
> >>> consequences of doing this in the worst case would be that the rows for
> >>> the updated objects might be drawn with some columns at a slightly
> >>> incorrect width, which is easily rectified by reverting the table
> >>> (usually bound to "g").  As well, that worst case (e.g. imagining a
> >>> vtable whose buffer might be initially displayed on one
> terminal/monitor
> >>> and later on another with different characteristics) would seem to be
> >>> relatively rare (so for my project, it seems like an obviously good
> >>> thing to do).
> >>>
> >>> For Emacs itself, I'm not sure what the best fix would be.  I suppose a
> >>> workaround like this could be implemented as a fallback in case the
> >>> cache key misses; it would seem better to update the object potentially
> >>> sub-optimally than to error and not update it at all.
> >>>
> >>> Another possibility would be to ignore the frame-terminal and
> >>> window-width in the cache key altogether (i.e. so they would always be
> >>> assumed to be the same), but I'm sure that Lars did it this way for a
> >>> reason, so that would seem unwise.
> >>>
> >>> Let me know how you'd like me to proceed.
> >>
> >> I think you should install your workaround.  I don't see how it could
> >> be worse than what we have now.
> >>
> >> P.S. And sorry for a long silence.
> >
> > Adam, could you please install the workaround?  Or maybe you did
> > already?
>
> I ran into this same problem.  I think Adam's workaround is on the right
> track, but the issue is present in a number of other functions in
> vtable.el, not just vtable-update-object.
>
> The root of the issue is that when we're interacting with the text of an
> inserted vtable, we should use the "cache" object that was last used to
> insert that vtable, rather than one based on the current selected window
> and terminal.  Otherwise we will experience various inconsistencies,
> e.g. inserting a new line that has different column widths from the rest
> of the vtable text.
>
> My attached patch solves this by saving the "cache" object in a text
> property on the vtable text, and updating all the relevant vtable
> functions to access the cache via that text property rather than by
> computing a cache key from the current selected window/terminal.
>
> Adam, could you check that this solves the problem for your use case?
>

I have a very big patch coming for vtable under
https://debbugs.gnu.org/cgi/bugreport.cgi?bug=78843 which addresses this
issue, and a ton of other bugs (and features).  That's coming in the next
day or so, if you can wait?

-Stéphane

[Message part 2 (text/html, inline)]

Message #37 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Spencer Baugh <sbaugh <at> janestreet.com>
To: Stéphane Marks <shipmints <at> gmail.com>
Cc: Adam Porter <adam <at> alphapapa.net>, 69837 <at> debbugs.gnu.org,
 Eli Zaretskii <eliz <at> gnu.org>, Stefan Kangas <stefankangas <at> gmail.com>
Subject: Re: bug#69837: 29.2; vtable-update-object only works in visible
 windows
Date: Thu, 19 Jun 2025 18:02:59 -0400

[Message part 1 (text/plain, inline)]

Stéphane Marks <shipmints <at> gmail.com> writes:
> On Thu, Jun 19, 2025 at 5:37 PM Spencer Baugh via Bug reports for GNU Emacs, the Swiss army knife of text editors
> <bug-gnu-emacs <at> gnu.org> wrote:
>
>  Stefan Kangas <stefankangas <at> gmail.com> writes:
>
>  > Eli Zaretskii <eliz <at> gnu.org> writes:
>  >
>  >>> Date: Thu, 21 Mar 2024 03:36:02 -0500
>  >>> Cc: 69837 <at> debbugs.gnu.org
>  >>> From: Adam Porter <adam <at> alphapapa.net>
>  >>>
>  >>> FWIW, I did some hacking on listen.el attempting to further understand
>  >>> and work around this problem, and I've found what seems to be a usable
>  >>> workaround (for my purposes, anyway).
>  >>>
>  >>> The attached code defines a macro within which I call
>  >>> `listen-queue--vtable-update-object' (which merely incorporates the fix
>  >>> to `vtable-update-object' from bug#69664).  As well, I wrap
>  >>> `make-vtable' in a function that also sets two buffer-local variables to
>  >>> the values of `frame-terminal' and `window-width' at the time of the
>  >>> vtable's creation.  The macro locally overrides the functions
>  >>> `frame-terminal' and `window-width' to return those saved values.
>  >>>
>  >>> This allows the cache key to match unconditionally, which allows the
>  >>> vtable's objects to be updated even when its buffer is not visible.
>  >>>
>  >>> In my limited testing, it seems to work fine.  In my estimation, the
>  >>> consequences of doing this in the worst case would be that the rows for
>  >>> the updated objects might be drawn with some columns at a slightly
>  >>> incorrect width, which is easily rectified by reverting the table
>  >>> (usually bound to "g").  As well, that worst case (e.g. imagining a
>  >>> vtable whose buffer might be initially displayed on one terminal/monitor
>  >>> and later on another with different characteristics) would seem to be
>  >>> relatively rare (so for my project, it seems like an obviously good
>  >>> thing to do).
>  >>>
>  >>> For Emacs itself, I'm not sure what the best fix would be.  I suppose a
>  >>> workaround like this could be implemented as a fallback in case the
>  >>> cache key misses; it would seem better to update the object potentially
>  >>> sub-optimally than to error and not update it at all.
>  >>>
>  >>> Another possibility would be to ignore the frame-terminal and
>  >>> window-width in the cache key altogether (i.e. so they would always be
>  >>> assumed to be the same), but I'm sure that Lars did it this way for a
>  >>> reason, so that would seem unwise.
>  >>>
>  >>> Let me know how you'd like me to proceed.
>  >>
>  >> I think you should install your workaround.  I don't see how it could
>  >> be worse than what we have now.
>  >>
>  >> P.S. And sorry for a long silence.
>  >
>  > Adam, could you please install the workaround?  Or maybe you did
>  > already?
>
>  I ran into this same problem.  I think Adam's workaround is on the right
>  track, but the issue is present in a number of other functions in
>  vtable.el, not just vtable-update-object.
>
>  The root of the issue is that when we're interacting with the text of an
>  inserted vtable, we should use the "cache" object that was last used to
>  insert that vtable, rather than one based on the current selected window
>  and terminal.  Otherwise we will experience various inconsistencies,
>  e.g. inserting a new line that has different column widths from the rest
>  of the vtable text.
>
>  My attached patch solves this by saving the "cache" object in a text
>  property on the vtable text, and updating all the relevant vtable
>  functions to access the cache via that text property rather than by
>  computing a cache key from the current selected window/terminal.
>
>  Adam, could you check that this solves the problem for your use case?
>
> I have a very big patch coming for vtable under https://debbugs.gnu.org/cgi/bugreport.cgi?bug=78843 which addresses this issue,
> and a ton of other bugs (and features).  That's coming in the next day or so, if you can wait?

Certainly I can wait.  Cc me when you send the patch.

Though if your patch is very large, maybe this relatively small patch
should land first.

(Also, I hope that you're adding tests in your very large patch?)

BTW, an updated version of my patch is attached, just fixing a silly bug
that was in the first version.

[0001-Fix-implicit-usage-of-the-current-window-width-in-vt.patch (text/x-patch, inline)]

From 5607bb42202e050edc5e33b8fa9c1c96d65d8746 Mon Sep 17 00:00:00 2001
From: Spencer Baugh <sbaugh <at> janestreet.com>
Date: Thu, 19 Jun 2025 17:20:12 -0400
Subject: [PATCH] Fix implicit usage of the current window-width in vtable.el

Previously, many functions in vtable.el called vtable--cache,
which computed vtable--cache-key based on the current selected
window and frame; this could cause vtable functions to fail or
misbehave if they were not called from the selected window and
frame that vtable-insert was last called in.

Now, the vtable cache is stored with the text of the vtable, so
that functions which need to interact with some vtable text can
do so reliably without having to use the same selected window
and frame.

Also, vtable-update-object has always required TABLE to be
present at point in the current buffer; now its docstring states
this.

* lisp/emacs-lisp/vtable.el (vtable--current-cache)
(vtable--cache-widths, vtable--cache-lines): Add.
(vtable-insert): Save cache in 'vtable-cache.
(vtable--ensure-cache, vtable--recompute-cache): Inline into
vtable-insert.
(vtable--widths, vtable--cache): Delete.
(vtable-update-object): Use vtable--current-cache and update
docstring.  (bug#69837)
(vtable-remove-object, vtable-insert-object): Use
vtable--current-cache and save cache in 'vtable-cache.
(vtable--sort, vtable--alter-column-width)
(vtable-previous-column, vtable-next-column): Use
vtable--current-cache.
---
 lisp/emacs-lisp/vtable.el | 111 ++++++++++++++++++++------------------
 1 file changed, 59 insertions(+), 52 deletions(-)

diff --git a/lisp/emacs-lisp/vtable.el b/lisp/emacs-lisp/vtable.el
index 00785113edb..d6be6d361f7 100644
--- a/lisp/emacs-lisp/vtable.el
+++ b/lisp/emacs-lisp/vtable.el
@@ -282,10 +282,9 @@ vtable-update-object
   "Update OBJECT's representation in TABLE.
 If OLD-OBJECT is non-nil, replace OLD-OBJECT with OBJECT and display it.
 In either case, if the existing object is not found in the table (being
-compared with `equal'), signal an error.  Note a limitation: if TABLE's
-buffer is not in a visible window, or if its window has changed width
-since it was updated, updating the TABLE is not possible, and an error
-is signaled."
+compared with `equal'), signal an error.
+
+TABLE must be at point in the current buffer."
   (unless old-object
     (setq old-object object))
   (let* ((objects (vtable-objects table))
@@ -302,14 +301,13 @@ vtable-update-object
       (unless objects
         (error "Can't find the old object"))
       (setcar (cdr objects) object))
-    ;; Then update the cache...
-    ;; FIXME: If the table's buffer has no visible window, or if its
-    ;; width has changed since the table was updated, the cache key will
-    ;; not match and the object can't be updated.  (Bug #69837).
-    (if-let* ((line-number (seq-position (car (vtable--cache table)) old-object
-                                         (lambda (a b)
-                                           (equal (car a) b))))
-              (line (elt (car (vtable--cache table)) line-number)))
+    ;; Then update the rendered vtable in the current buffer.
+    (if-let* ((cache (vtable--current-cache))
+             (line-number (seq-position (vtable--cache-lines cache)
+                                        old-object
+                                        (lambda (a b)
+                                          (equal (car a) b))))
+             (line (elt (vtable--cache-lines cache) line-number)))
         (progn
           (setcar line object)
           (setcdr line (vtable--compute-cached-line table object))
@@ -320,10 +318,11 @@ vtable-update-object
                   (start (point)))
               (delete-line)
               (vtable--insert-line table line line-number
-                                   (nth 1 (vtable--cache table))
+                                   (vtable--cache-widths cache)
                                    (vtable--spacer table))
               (add-text-properties start (point) (list 'keymap keymap
-                                                       'vtable table))))
+                                                       'vtable table
+                                                       'vtable-cache cache))))
           ;; We may have inserted a non-numerical value into a previously
           ;; all-numerical table, so recompute.
           (vtable--recompute-numerical table (cdr line)))
@@ -335,11 +334,12 @@ vtable-remove-object
   ;; First remove from the objects.
   (setf (vtable-objects table) (delq object (vtable-objects table)))
   ;; Then adjust the cache and display.
-  (let ((cache (vtable--cache table))
-        (inhibit-read-only t))
-    (setcar cache (delq (assq object (car cache)) (car cache)))
-    (save-excursion
-      (vtable-goto-table table)
+  (save-excursion
+    (vtable-goto-table table)
+    (let ((cache (vtable--current-cache))
+          (inhibit-read-only t))
+      (setcar cache (delq (assq object (vtable--cache-lines cache))
+                          (vtable--cache-lines cache)))
       (when (vtable-goto-object object)
         (delete-line)))))
 
@@ -400,7 +400,7 @@ vtable-insert-object
     ;; Then adjust the cache and display.
     (save-excursion
       (vtable-goto-table table)
-      (let* ((cache (vtable--cache table))
+      (let* ((cache (vtable--current-cache))
              (inhibit-read-only t)
              (keymap (get-text-property (point) 'keymap))
              (ellipsis (if (vtable-ellipsis table)
@@ -408,13 +408,14 @@ vtable-insert-object
                                        'face (vtable-face table))
                          ""))
              (ellipsis-width (string-pixel-width ellipsis))
-             (elem (if location  ; This binding mirrors the binding of `pos' above.
+             (lines (vtable--cache-lines cache))
+             (elem (if location ; This binding mirrors the binding of `pos' above.
                        (if (integerp location)
-                           (nth location (car cache))
-                         (or (assq location (car cache))
-                             (and before (caar cache))))
-                     (if before (caar cache))))
-             (pos (memq elem (car cache)))
+                           (nth location lines)
+                         (or (assq location lines)
+                             (and before (car lines))))
+                     (if before (car lines))))
+             (pos (memq elem lines))
              (line (cons object (vtable--compute-cached-line table object))))
         (if (or before
                 (and pos (integerp location)))
@@ -433,13 +434,13 @@ vtable-insert-object
                     (forward-line 1)  ; Insert *after*.
                   (vtable-end-of-table)))
             ;; Otherwise, append the object.
-            (setcar cache (nconc (car cache) (list line)))
+            (setcar cache (nconc (vtable--cache-lines cache) (list line)))
             (vtable-end-of-table)))
         (let ((start (point)))
           ;; FIXME: We have to adjust colors in lines below this if we
           ;; have :row-colors.
           (vtable--insert-line table line 0
-                               (nth 1 cache) (vtable--spacer table)
+                               (vtable--cache-widths cache) (vtable--spacer table)
                                ellipsis ellipsis-width)
           (add-text-properties start (point) (list 'keymap keymap
                                                    'vtable table)))
@@ -512,15 +513,11 @@ vtable--compute-columns
 (defun vtable--spacer (table)
   (vtable--compute-width table (vtable-separator-width table)))
 
-(defun vtable--recompute-cache (table)
-  (let* ((data (vtable--compute-cache table))
-         (widths (vtable--compute-widths table data)))
-    (setf (gethash (vtable--cache-key) (slot-value table '-cache))
-          (list data widths))))
+(defun vtable--cache-widths (cache)
+  (nth 1 cache))
 
-(defun vtable--ensure-cache (table)
-  (or (vtable--cache table)
-      (vtable--recompute-cache table)))
+(defun vtable--cache-lines (cache)
+  (car cache))
 
 (defun vtable-insert (table)
   (let* ((spacer (vtable--spacer table))
@@ -533,7 +530,12 @@ vtable-insert
          ;; We maintain a cache per screen/window width, so that we render
          ;; correctly if Emacs is open on two different screens (or the
          ;; user resizes the frame).
-         (widths (nth 1 (vtable--ensure-cache table))))
+         (cache (or (gethash (vtable--cache-key) (slot-value table '-cache))
+                    (let* ((data (vtable--compute-cache table))
+                           (widths (vtable--compute-widths table data)))
+                      (setf (gethash (vtable--cache-key) (slot-value table '-cache))
+                            (list data widths)))))
+         (widths (vtable--cache-widths cache)))
     ;; Don't insert any header or header line if the user hasn't
     ;; specified the columns.
     (when (slot-value table '-has-column-spec)
@@ -546,18 +548,20 @@ vtable-insert
         (add-text-properties start (point)
                              (list 'keymap vtable-header-line-map
                                    'rear-nonsticky t
-                                   'vtable table))
+                                   'vtable table
+                                   'vtable-cache cache))
         (setq start (point))))
-    (vtable--sort table)
+    (vtable--sort table cache)
     ;; Insert the data.
     (let ((line-number 0))
-      (dolist (line (car (vtable--cache table)))
+      (dolist (line (vtable--cache-lines cache))
         (vtable--insert-line table line line-number widths spacer
                              ellipsis ellipsis-width)
         (setq line-number (1+ line-number))))
     (add-text-properties start (point)
                          (list 'rear-nonsticky t
-                               'vtable table))
+                               'vtable table
+                               'vtable-cache cache))
     (goto-char start)))
 
 (defun vtable--insert-line (table line line-number widths spacer
@@ -659,16 +663,22 @@ vtable--insert-line
 (defun vtable--cache-key ()
   (cons (frame-terminal) (window-width)))
 
-(defun vtable--cache (table)
-  (gethash (vtable--cache-key) (slot-value table '-cache)))
+(defun vtable--current-cache ()
+  "Return the current cache for the table at point.
+
+In `vtable-insert', the lines and widths of the vtable text are computed
+based on the current selected frame and window and stored in a cache.
+Subsequent interaction with the text of the vtable should use that cache
+via this function rather than by calling `vtable--cache-key' to look up
+the cache."
+  (get-text-property (point) 'vtable-cache))
 
 (defun vtable--clear-cache (table)
   (setf (gethash (vtable--cache-key) (slot-value table '-cache)) nil))
 
-(defun vtable--sort (table)
+(defun vtable--sort (table cache)
   (pcase-dolist (`(,index . ,direction) (vtable-sort-by table))
-    (let ((cache (vtable--cache table))
-          (numerical (vtable-column--numerical
+    (let ((numerical (vtable-column--numerical
                       (elt (vtable-columns table) index)))
           (numcomp (if (eq direction 'descend)
                        #'> #'<))
@@ -971,9 +981,6 @@ vtable-revert
     (when column
       (vtable-goto-column column))))
 
-(defun vtable--widths (table)
-  (nth 1 (vtable--ensure-cache table)))
-
 ;;; Commands.
 
 (defvar-keymap vtable-header-mode-map
@@ -998,7 +1005,7 @@ vtable-narrow-current-column
                                 (- (* (vtable--char-width table) (or n 1))))))
 
 (defun vtable--alter-column-width (table column delta)
-  (let ((widths (vtable--widths table)))
+  (let ((widths (vtable--cache-widths (vtable--current-cache))))
     (setf (aref widths column)
           (max (* (vtable--char-width table) 2)
                (+ (aref widths column) delta)))
@@ -1020,14 +1027,14 @@ vtable-previous-column
   (interactive)
   (vtable-goto-column
    (max 0 (1- (or (vtable-current-column)
-                  (length (vtable--widths (vtable-current-table))))))))
+                  (length (vtable--cache-widths (vtable--current-cache))))))))
 
 (defun vtable-next-column ()
   "Go to the next column."
   (interactive)
   (when (vtable-current-column)
     (vtable-goto-column
-     (min (1- (length (vtable--widths (vtable-current-table))))
+     (min (1- (length (vtable--cache-widths (vtable--current-cache))))
           (1+ (vtable-current-column))))))
 
 (defun vtable-revert-command ()
-- 
2.39.3

Message #40 received at 69837 <at> debbugs.gnu.org (full text, mbox):

From: Stéphane Marks <shipmints <at> gmail.com>
To: Spencer Baugh <sbaugh <at> janestreet.com>
Cc: Adam Porter <adam <at> alphapapa.net>, 69837 <at> debbugs.gnu.org,
 Eli Zaretskii <eliz <at> gnu.org>, Stefan Kangas <stefankangas <at> gmail.com>
Subject: Re: bug#69837: 29.2;
 vtable-update-object only works in visible windows
Date: Thu, 19 Jun 2025 18:06:40 -0400

[Message part 1 (text/plain, inline)]

On Thu, Jun 19, 2025 at 6:02 PM Spencer Baugh <sbaugh <at> janestreet.com> wrote:

> Stéphane Marks <shipmints <at> gmail.com> writes:
> > On Thu, Jun 19, 2025 at 5:37 PM Spencer Baugh via Bug reports for GNU
> Emacs, the Swiss army knife of text editors
> > <bug-gnu-emacs <at> gnu.org> wrote:
> >
> >  Stefan Kangas <stefankangas <at> gmail.com> writes:
> >
> >  > Eli Zaretskii <eliz <at> gnu.org> writes:
> >  >
> >  >>> Date: Thu, 21 Mar 2024 03:36:02 -0500
> >  >>> Cc: 69837 <at> debbugs.gnu.org
> >  >>> From: Adam Porter <adam <at> alphapapa.net>
> >  >>>
> >  >>> FWIW, I did some hacking on listen.el attempting to further
> understand
> >  >>> and work around this problem, and I've found what seems to be a
> usable
> >  >>> workaround (for my purposes, anyway).
> >  >>>
> >  >>> The attached code defines a macro within which I call
> >  >>> `listen-queue--vtable-update-object' (which merely incorporates the
> fix
> >  >>> to `vtable-update-object' from bug#69664).  As well, I wrap
> >  >>> `make-vtable' in a function that also sets two buffer-local
> variables to
> >  >>> the values of `frame-terminal' and `window-width' at the time of the
> >  >>> vtable's creation.  The macro locally overrides the functions
> >  >>> `frame-terminal' and `window-width' to return those saved values.
> >  >>>
> >  >>> This allows the cache key to match unconditionally, which allows the
> >  >>> vtable's objects to be updated even when its buffer is not visible.
> >  >>>
> >  >>> In my limited testing, it seems to work fine.  In my estimation, the
> >  >>> consequences of doing this in the worst case would be that the rows
> for
> >  >>> the updated objects might be drawn with some columns at a slightly
> >  >>> incorrect width, which is easily rectified by reverting the table
> >  >>> (usually bound to "g").  As well, that worst case (e.g. imagining a
> >  >>> vtable whose buffer might be initially displayed on one
> terminal/monitor
> >  >>> and later on another with different characteristics) would seem to
> be
> >  >>> relatively rare (so for my project, it seems like an obviously good
> >  >>> thing to do).
> >  >>>
> >  >>> For Emacs itself, I'm not sure what the best fix would be.  I
> suppose a
> >  >>> workaround like this could be implemented as a fallback in case the
> >  >>> cache key misses; it would seem better to update the object
> potentially
> >  >>> sub-optimally than to error and not update it at all.
> >  >>>
> >  >>> Another possibility would be to ignore the frame-terminal and
> >  >>> window-width in the cache key altogether (i.e. so they would always
> be
> >  >>> assumed to be the same), but I'm sure that Lars did it this way for
> a
> >  >>> reason, so that would seem unwise.
> >  >>>
> >  >>> Let me know how you'd like me to proceed.
> >  >>
> >  >> I think you should install your workaround.  I don't see how it could
> >  >> be worse than what we have now.
> >  >>
> >  >> P.S. And sorry for a long silence.
> >  >
> >  > Adam, could you please install the workaround?  Or maybe you did
> >  > already?
> >
> >  I ran into this same problem.  I think Adam's workaround is on the right
> >  track, but the issue is present in a number of other functions in
> >  vtable.el, not just vtable-update-object.
> >
> >  The root of the issue is that when we're interacting with the text of an
> >  inserted vtable, we should use the "cache" object that was last used to
> >  insert that vtable, rather than one based on the current selected window
> >  and terminal.  Otherwise we will experience various inconsistencies,
> >  e.g. inserting a new line that has different column widths from the rest
> >  of the vtable text.
> >
> >  My attached patch solves this by saving the "cache" object in a text
> >  property on the vtable text, and updating all the relevant vtable
> >  functions to access the cache via that text property rather than by
> >  computing a cache key from the current selected window/terminal.
> >
> >  Adam, could you check that this solves the problem for your use case?
> >
> > I have a very big patch coming for vtable under
> https://debbugs.gnu.org/cgi/bugreport.cgi?bug=78843 which addresses this
> issue,
> > and a ton of other bugs (and features).  That's coming in the next day
> or so, if you can wait?
>
> Certainly I can wait.  Cc me when you send the patch.
>
> Though if your patch is very large, maybe this relatively small patch
> should land first.
>
> (Also, I hope that you're adding tests in your very large patch?)
>
> BTW, an updated version of my patch is attached, just fixing a silly bug
> that was in the first version.
>

Your patch doesn't address several other intertwined issues that my bigger
patch addresses.  I've been testing it for ages now and with other vtable
users.  I can send a new vtable.el and vtable.texi to you off line if you'd
like a look before I finish writing the long changelog.

-Stéphane

[Message part 2 (text/html, inline)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.