[scribus] an experiment with lines, and a surprising result

[JLuc]

Le 26/05/2014 18:21, Gregory Pittman a écrit :
> 1. Start with a horizontal line somewhere near the top of the page
> 2. Make a copy of this with multiple duplicate, no X or Y shift
> 3. Rotate the copy 270°
> 4. Now group the two lines
> 5. Make a copy of this with multiple duplicate, no X or Y shift
> 6. Now, select the copy, then flip it horizontally, then flip vertically
> Look at the alignment of the bottom line of this pseudorectangle.
> Where does this come from?

It might be related to the thickness of the line and having to do math with points that are thick points.

Its a general issue : what should be the visible "Xvisible" start point
of a vertical line whose position displays as "Xpp" in the PP,
and whose thickness is "2t" :
Should we have
	Xvis = Xpp ?
	Xvis = Xpp - t ?

First option will lead the line to be on the side of its PP-displayed origin.
2nd option will lead to have the line centered on its PP-displayed origin.

Scribus has choosen the second option,
that AMOF can be felt as leading to better results in general
(but might cause issues too in some cases too)

Check that :
Add a thickness of 10mm at step 1. of your test
and you will see immediately that scribus uses the 2nd option.
Fine.
So what ?
Let's explore further your procedure, now with 10mm thick lines.

You will see in step 4 (before grouping) that when you select the 2 lines,
scribus does a mistake when creating the selection region :
Instead of starting at (X-t , Y-t), the selected region begins at (X-2t, Y-t).
*This is the origin of the issue.*

Fact that computation is correct for Y and uncorrect for X
leads me to think that it could easily be fixed.

4- Grouping the lines keeps that mistake : group has same buggy coordinates as selection.

5 Duplicating the 2 (grouped or not) lines propagate the mistake.
AMOF Setp 5 and 6 behave correctly but since the selection at the begining of step 4
computes a too wide area, the symetry does not lead to expected results.
At the end of the process with thick lines it is very visible.

JL