[Libreoffice-bugs] [Bug 95730] Jumping to block margin is broken after clearing cells

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=95730

--- Comment #3 from ikjt <kalle at goodtimes.fi> ---
The issue is still present in v5.3.2.2.

While testing the current version, I noticed that my original STR contained a
redundant step; clearing the selection is actually not necessary. All that is
needed is to make a selection and then jump to a margin (along a dimension of
the selection).

* * *

Simplified steps to reproduce:

1. On an empty sheet, make a selection from A20 upward to A15.

2. Jump to the top margin (Cmd-Up on a Mac).

Expected results:

The cursor goes to A1.

Actual results:

The cursor goes to A6.

Explanation:

The cursor movement stops 5 rows short of the target because the selection
extends 5 rows towards the target. Conversely, if the selection extends *away*
from the target, the final position overshoots instead of stopping short. It's
as if the movement delta is incorrectly calculated from the edge of the
selection instead of the cursor position.

* * *

There are multiple reports of keyboard navigation bugs that I believe to have
the same root cause. It's pretty clearly just a simple arithmetic bug somewhere
and should be an easy fix once it's located.

Please let me know if there's any way to help get these bugs fixed. I would
gladly have a look at the relevant code if someone pointed out where it
resides. Thanks!

AFAIK, the bug has not been present in any version of OpenOffice, but it
already appears in the very first version of LibreOffice. Thus, it represents a
regression from the shared codebase.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/libreoffice-bugs/attachments/20170502/7d506407/attachment-0001.html>