[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
bug#69972: 29.1; Unexpected behavior when scrolling images
From: |
Joseph Turner |
Subject: |
bug#69972: 29.1; Unexpected behavior when scrolling images |
Date: |
Mon, 25 Mar 2024 01:34:40 -0700 |
Joseph Turner <joseph@ushin.org> writes:
> Eli Zaretskii <eliz@gnu.org> writes:
>
>>> Date: Sun, 24 Mar 2024 02:07:16 -0700
>>> From: Joseph Turner via "Bug reports for GNU Emacs,
>>> the Swiss army knife of text editors" <bug-gnu-emacs@gnu.org>
>>>
>>> When the image at point is larger than the current window and there is
>>> no content after the image, interactively scrolling down (with the
>>> scroll-up command) unexpectedly scrolls past the image to blankness.
>>
>> I cannot reproduce this, I think.
>>
>>> Test this by evaluating the following snippet then interactively running
>>> `scroll-up' repeatedly:
>>>
>>> (with-current-buffer (get-buffer-create "*test-scroll-image*")
>>> (erase-buffer)
>>> (insert-image (create-image "splash.png" nil nil
>>> ;; Scale the image more if it doesn't take up the whole window.
>>> :scale 5))
>>> (goto-char (point-max))
>>> (pop-to-buffer (current-buffer)))
>>
>> This recipe doesn't include the call to scroll-up, so I'm unsure how
>> you did that and what you saw. When I try "M-: (scroll-up) RET" or
>> "M-x scroll-up RET", I get several scrolls by window-size, and then
>> "End of buffer" error when I hit the end of the buffer. If this is
>> unexpected, please tell why.
>
> On my machine, I don't get "End of buffer" error.
Correction - I *do* get "End of buffer" error, but only after fully
scrolling past the image (which IMO is unexpected).
> I tested three different ways in the following snippet
>
> (with-current-buffer (get-buffer-create "*test-scroll-image*")
> (erase-buffer)
> (insert-image (create-image "splash.png" nil nil
> ;; Scale the image more if it doesn't take up the whole window.
> :scale 5))
> (goto-char (point-max))
> (pixel-scroll-precision-mode -1) ; Ensure mouse wheel scroll up and down
> works
> (pop-to-buffer (current-buffer)))
>
> 1. "M-: (scroll-up) RET" (repeatedly)
> 2. C-v (repeatedly)
> 3. <wheel-down> (repeatedly)
>
> With all three methods, at first Emacs gradually scrolls the image, but
> then when I reach the bottom of the image, the image disappears entirely
> as Emacs scrolls past it all at once.
>
>>> Even more unexpectedly, when point is before the image, running
>>> `scroll-up' repeatedly eventually scrolls back to the top of the image:
>>>
>>> (with-current-buffer (get-buffer-create "*test-scroll-image*")
>>> (erase-buffer)
>>> (insert-image (create-image "splash.png" nil nil :scale 5))
>>> (goto-char (point-min))
>>> (pop-to-buffer (current-buffer)))
>>
>> I do see this, but why is that a problem? You supposed to use
>> scroll-up-command instead, which handles these marginal cases much
>> better. scroll-up itself is not smart enough to avoid the perceived
>> "scroll back to top", which is actually caused by the fact that we
>> zero out window-vscroll (which is how we handle scrolling past large
>> images).
>
> You're right. This is not a problem in practice. When point is before
> the image, both C-v and <wheel-down> produce the same behavior as above.
>
> Joseph