[Libreoffice-ux-advise] [Bug 134293] Inconsistent double-click AutoFill behavior

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=134293

--- Comment #13 from Dmitry T. <contradictious.one at gmail.com> ---
(In reply to Joshua Coppersmith from comment #11)
Thanks for trying to convey my point (as i'm, surely, not doing it well enough
myself, due to not being native English speaker), but, to be honest, it feels
that your explanation is not entirely correct or on point.

> Expected behavior would be that a HIDDEN blank cell does not count when
> stopping double-click expansion.

As i mentioned earlier, is it really an expected behavior? Should LO have it's
own (consistent) behavior, or do the same thing Excel does? When i reported
this issue, i didn't consider Excel at all, and based "expected" (by me, that
is) behavior on drag-down fill, but now i think that the way it works in Excel
actually makes more sense. I mean, consider how formulas work - if you copy a
formula that references a row above into a cell right below the hidden row, it
will reference a hidden row above, not the visible row. I think, AutoFill
should act the same.

> However, since a hidden blank cell WILL
> stop double-click expansion, this makes it look like double-click expansion
> is "stopping short" or just not working at all.

That is correct, but i should mention, that empty cell will stop double-click
expansion before ALL hidden rows, not just the hidden blank cell. SO, if, among
10 hidden rows, there are 9 cells with data, and 10th cell is blank, no hidden
cell will be replaced.

> This behavior could be completely disregarded as just the nature of
> double-click expansion, 

It cannot be disregarded yet, as there still exists inconsistent behavior: 
1. If all hidden cells are NOT blank and cells below hidden rows ARE blank,
double-click expansion will suddenly start filling blank cells until it hits a
non-blank cell or the bottom of the sheet (and latter could be performance
heavy), while ignoring neighboring columns.
2. If there is only a single hidden row with a NON-BLANK cell AutoFill does
nothing instead, which makes no sense at all.

> except that when expanding multiple columns at once
> all columns stop short if any column has a blank hidden cell in the
> expansion path.

Now, behavior for multiple columns is also a point of concern (good call), but,
again, i think it should be first agreed on whether LO should have it's own
behavior, or mirror Excel's one. In Excel (and in LO without hidden cells)
behavior is that all cells are filled down as long as all cells are the same
(all blank or non-blank).

Considering all that, summary you gave is kinda moot.

-- 
You are receiving this mail because:
You are on the CC list for the bug.