[Libreoffice-bugs] [Bug 131758] Consistent polygon tool behavior in Draw.

[bugzilla-daemon at bugs.documentfoundation.org]

https://bugs.documentfoundation.org/show_bug.cgi?id=131758

--- Comment #10 from Jan-Marek Glogowski <glogow at fbihome.de> ---
Hi Armin,

(In reply to Armin Le Grand from comment #5)
> - What is the correct handling here from your POV?
> AFAIK as it was before changes were made in that area. Obviously that
> touched stuff which was not on the radar - what always can happen, esp. in
> areas like these. Turn back, check was was intended to be changed - there
> may be good reasons - and do it in a way without side effects with knowing
> more usages/cases

Sorry, but that doesn't help me. As I already wrote 5.4, 6.2 and 7.0 behave
different w.r.t. the polygon functionality. So I just checked out 3.3.0, which
behaves like 5.4 ... most times.

Sometimes "Ctrl + Alt + left-click" close the polygon and continues with a line
from the last manual point, sometimes the line doesn't appear and you can start
a 2nd, independent polygon. I still think 7.0 is correct and therefore 3.3.0 is
also broken.

> - Would it make sense to split polygon and polyline for more consistent
> behavior?
> 
> No. The one is filled, the other is not. Only consequent thing needed is to
> know the target type (currently in OBJ_TYPE AFAIK, argh, already had that
> better in aw080) and adapt each finished polygon in polyPolygon topology

There is especially the "polygon, filled" tool, so this is already handled
separately by the GUI. I'm talking about GUI functionality, not code
representation, where a polygon can always be closed or not. The latter case is
then a polyline. And I'm just talking about splitting the tooling.

> - Is a PolyPolygon just a group of independent polygons, or do they have to
> share a common point, as 5.4 handling seems to somehow imply, as it continues
> the line?
> 
> Yes. No. :-)
> Why should they share a point? This makes no sense at all - a polyPolygon is
> a bunch of single polygons. Topology defines adapting a FillRule to these
> when visualizing. We just have EvenOdd/NonZero where we use EvenOdd only -
> so it's more or less a xor - for each point # of being inside %1

Great. This is also my understanding. I just wanted to make sure. 

> What difference do you see between creating polygons or polylines? Only diff
> is that one gets outline painted, other get fill painted. Anyways, for our
> attributation you can choose both as you wish. It is better  when the filled
> polygon is closed, but I would have to check if this is currently a criteria
> to offer fill in the attributes - there is also that polygon object type
> (OBJ_*)

>From the UX POV, the "polygon" is not filled per default, but the "polygon,
filled" is. Sure, you can easily change that via the area setting, but the
whole "lines" tool box currently distinguishes between "normal" and filled
shapes, even if the result can be easily switched later by changing the area
attribute.

> That CTRL+ALT (was it just CTRL before?) behaviour is clearly buggy - when
> not using it the line variations (non-filled) are not closed - that's what
> it shopuld be.

It's like this also in 3.3.0. "Ctrl + Alt + left-click" closes the polygon here
on Linux. Just "Ctrl + left-click" does nothing here.

-- 
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/20200402/783f9994/attachment-0001.htm>