bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[E Sabof]

[-- Attachment #1: Type: text/plain, Size: 184 bytes --]

In emacs -Q

(window-line-height)
C-o
C-x C-e

Results in nil.

Interestingly enough, I can specify line numbers (or not) and it will work,
but only as long as (= (point) (point-max))

[-- Attachment #2: Type: text/html, Size: 311 bytes --]

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[Eli Zaretskii]

> Date: Sun, 16 Jun 2013 16:24:43 +0100
> From: E Sabof <esabof@gmail.com>
> 
> In emacs -Q
> 
> (window-line-height)
> C-o
> C-x C-e
> 
> Results in nil.

That's a documented limitation.  From the doc string:

  Return nil if window display is not up-to-date.  In that case, use
  `pos-visible-in-window-p' to obtain the information.

> Interestingly enough, I can specify line numbers (or not) and it will work,
> but only as long as (= (point) (point-max))

You cannot easily predict when the window display will be up-to-date
by the time this function runs.

Closing as not-a-bug.

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[E Sabof]

[-- Attachment #1: Type: text/plain, Size: 823 bytes --]

On Sun, Jun 16, 2013 at 5:35 PM, Eli Zaretskii <eliz@gnu.org> wrote:

> > Date: Sun, 16 Jun 2013 16:24:43 +0100
> > From: E Sabof <esabof@gmail.com>
> >
> > In emacs -Q
> >
> > (window-line-height)
> > C-o
> > C-x C-e
> >
> > Results in nil.
>
> That's a documented limitation.  From the doc string:
>
>   Return nil if window display is not up-to-date.  In that case, use
>   `pos-visible-in-window-p' to obtain the information.
>
> > Interestingly enough, I can specify line numbers (or not) and it will
> work,
> > but only as long as (= (point) (point-max))
>
> You cannot easily predict when the window display will be up-to-date
> by the time this function runs.
>
> Closing as not-a-bug.
>

I have also tried

(progn (pos-visible-in-window-p)
       (window-line-height))

but it didn't seem to make any difference.

[-- Attachment #2: Type: text/html, Size: 1553 bytes --]

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[Eli Zaretskii]

> Date: Sun, 16 Jun 2013 17:44:07 +0100
> From: E Sabof <esabof@gmail.com>
> Cc: 14636-done@debbugs.gnu.org
> 
> > That's a documented limitation.  From the doc string:
> >
> >   Return nil if window display is not up-to-date.  In that case, use
> >   `pos-visible-in-window-p' to obtain the information.
> >
> > > Interestingly enough, I can specify line numbers (or not) and it will
> > work,
> > > but only as long as (= (point) (point-max))
> >
> > You cannot easily predict when the window display will be up-to-date
> > by the time this function runs.
> >
> > Closing as not-a-bug.
> >
> 
> I have also tried
> 
> (progn (pos-visible-in-window-p)
>        (window-line-height))
> 
> but it didn't seem to make any difference.

Why should it?  Calling pos-visible-in-window-p does not guarantee
that the window will be up-to-date, it just gives you the same
information without requiring that the display be up-to-date.  Its
disadvantage is that it is slower.

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[E Sabof]

[-- Attachment #1: Type: text/plain, Size: 884 bytes --]

  > I have also tried

> >
> > (progn (pos-visible-in-window-p)
> >        (window-line-height))
> >
> > but it didn't seem to make any difference.
>
> Why should it?  Calling pos-visible-in-window-p does not guarantee
> that the window will be up-to-date, it just gives you the same
> information without requiring that the display be up-to-date.  Its
> disadvantage is that it is slower.
>

You might be confusing posn-at-point (which returns the height of the
character amoung other things), with pos-visible-in-window-p, which returns
t or nil.

In any case, the height of the character can be different from the height
of the line, so they are not equivalent. So either a) the documentation is
incomplete and misleading, or b) there is a bug in the implementation
(which I think is more likely, given the arbitrary circumstances under
which this function does and doesn't work).

[-- Attachment #2: Type: text/html, Size: 1449 bytes --]

bug#14636: 24.3.50.1; window-line-height doesn't seem to work, unless the point is at the end-of-buffer

[Eli Zaretskii]

> Date: Sun, 16 Jun 2013 18:19:06 +0100
> From: E Sabof <esabof@gmail.com>
> Cc: 14636@debbugs.gnu.org
> 
> You might be confusing posn-at-point (which returns the height of the
> character amoung other things), with pos-visible-in-window-p, which returns
> t or nil.

I'm not confusing them.  And pos-visible-in-window-p can return more
than just nil or t, see its doc string.

> In any case, the height of the character can be different from the height
> of the line, so they are not equivalent.

They are not equivalent, in the sense that a single call to
pos-visible-in-window-p will not do what window-line-height does.  You
need to call pos-visible-in-window-p twice, for the line you are
interested in and the one after it, to get the same information.

> So either a) the documentation is incomplete and misleading, or b)
> there is a bug in the implementation (which I think is more likely,
> given the arbitrary circumstances under which this function does and
> doesn't work).

There's no bug.  It looks arbitrary to you because whether the window
display is up-to-date when window-line-height runs is impossible to
predict.  Take a look at the source of window-line-height, and you
will see it.