bug#9224: Scrolling bugs and word wrapping

bug#9224: Scrolling bugs and word wrapping

[Antoine Levitt]

Hi,

(setq-default word-wrap t)
C-x C-f scrolling-fail.el
position point on "b"
C-p, watch point going to beginning of line instead of previous line

This started quite recently, and fails even with (setq
bidi-display-reordering nil)

bug#9224: Scrolling bugs and word wrapping

[Antoine Levitt]

[-- Attachment #1: Type: text/plain, Size: 307 bytes --]

02/08/11 20:01, Antoine Levitt
> Hi,
>
> (setq-default word-wrap t)
> C-x C-f scrolling-fail.el
> position point on "b"
> C-p, watch point going to beginning of line instead of previous line
>
> This started quite recently, and fails even with (setq
> bidi-display-reordering nil)

Once more, with feeling:

[-- Attachment #2: scrolling_fail.el --]
[-- Type: application/emacs-lisp, Size: 11 bytes --]

bug#9224: Scrolling bugs and word wrapping

[Eli Zaretskii]

> From: Antoine Levitt <antoine.levitt@gmail.com>
> Date: Tue, 02 Aug 2011 20:01:33 +0200
> 
> (setq-default word-wrap t)
> C-x C-f scrolling-fail.el
> position point on "b"
> C-p, watch point going to beginning of line instead of previous line
> 
> This started quite recently, and fails even with (setq
> bidi-display-reordering nil)

Could you please bisect to find the offending commit?  It's probably
my fault, even though it fails without bidi.

TIA

bug#9224: Scrolling bugs and word wrapping

[Antoine Levitt]

02/08/11 21:05, Eli Zaretskii
>> From: Antoine Levitt <antoine.levitt@gmail.com>
>> Date: Tue, 02 Aug 2011 20:01:33 +0200
>> 
>> (setq-default word-wrap t)
>> C-x C-f scrolling-fail.el
>> position point on "b"
>> C-p, watch point going to beginning of line instead of previous line
>> 
>> This started quite recently, and fails even with (setq
>> bidi-display-reordering nil)
>
> Could you please bisect to find the offending commit?  It's probably
> my fault, even though it fails without bidi.
>
> TIA

Turns out I'm a moron. It only fails with bidi on. That's my fault for
using setq and not setq-default.

bug#9224: Scrolling bugs and word wrapping

[Antoine Levitt]

02/08/11 22:34, Antoine Levitt
> 02/08/11 21:05, Eli Zaretskii
>>> From: Antoine Levitt <antoine.levitt@gmail.com>
>>> Date: Tue, 02 Aug 2011 20:01:33 +0200
>>> 
>>> (setq-default word-wrap t)
>>> C-x C-f scrolling-fail.el
>>> position point on "b"
>>> C-p, watch point going to beginning of line instead of previous line
>>> 
>>> This started quite recently, and fails even with (setq
>>> bidi-display-reordering nil)
>>
>> Could you please bisect to find the offending commit?  It's probably
>> my fault, even though it fails without bidi.
>>
>> TIA
>
> Turns out I'm a moron. It only fails with bidi on. That's my fault for
> using setq and not setq-default.

Fixed now, thanks!

The memory bug also seems fixed, I'll report back if I see any increase.

bug#9224: Scrolling bugs and word wrapping

[Eli Zaretskii]

> From: Antoine Levitt <antoine.levitt@gmail.com>
> Date: Fri, 05 Aug 2011 16:05:45 +0200
> 
> Fixed now, thanks!

It must have been fixed indirectly, because I didn't know I fixed this
bug as well.  But I confirm it now works correctly.  Oh, well, need a
bit of luck sometimes.

I'm closing this bug, then.

> The memory bug also seems fixed, I'll report back if I see any increase.

Yes, please do.  And thanks for all your help.

bug#9224: Scrolling bugs and word wrapping

[Antoine Levitt]

05/08/11 16:30, Eli Zaretskii
>> From: Antoine Levitt <antoine.levitt@gmail.com>
>> Date: Fri, 05 Aug 2011 16:05:45 +0200
>> 
>> Fixed now, thanks!
>
> It must have been fixed indirectly, because I didn't know I fixed this
> bug as well.  But I confirm it now works correctly.  Oh, well, need a
> bit of luck sometimes.
>
> I'm closing this bug, then.

Er, may have declared victory too soon on this one. Just had the exact
same thing happen to me with latest rev. At the end of
erc-scroll-to-bottom from trunk, put point like this:

                (sit-for 0)))))))X)

and C-p. It's not specific to this code either, pretty much any code
with nontrivial indentation and parentheses will do. It actually also
happens on this very message, with point on the X.

Sorry ;-)