extensions mix selection order

While the proposed solution is a feature request, the mixing up of selection order is a real bug that I've run into. jfb--could you provide step-by-step instructions that trigger the bug, starting from a new document? Thanks...

create 2 paths with white fill and black stroke.
place them so that one partially covers the other.
select the one in the back, then the one in the front.

efffect->interpolate.
the preview shows that the new shapes are added frome back to front.

change which ever parameter you want in the effect dialog.
the new shapes are now added from front to back...

you can do the same with pattern-along-path: which shape is considered the
pattern and which the skeleton depends on the selection order; this is
screwed up as soon as you change a parameter.

create 2 paths with white fill and black stroke.
place them so that one partially covers the other.
select the one in the back, then the one in the front.

efffect->interpolate.
the preview shows that the new shapes are added frome back to front.

change which ever parameter you want in the effect dialog.
the new shapes are now added from front to back...

you can do the same with pattern-along-path: which shape is considered the pattern and which the skeleton depends on the selection order; this is screwed up as soon as you change a parameter.

a work around is to use zOrder.instead of selection order. It is less handy, but still makes sens (at least for "interpolate").

I revisited the scripts of pathalongpath, scatter, and interpolate, and updated them to use the z-order.
As a bonus, interpolate now also supports objects like groups, clones, stars...

The files are submitted in https://bugs.launchpad.net/inkscape/+bug/182794 since the changes are also relevant for this bug.

Hi! Here is a patch containing all the changes of interpolate, path-along-path, path-scatter, path-modifier, simple-transform, and rubber-stretch.
(same as https://bugs.launchpad.net/inkscape/+bug/182794)

Could a Dev please take a look at JFB's patch and commit it if looks good?

path along a path, rubber stretch and related facilties are commited, but there are still some problems for interpolate and path-scatter :

In my tests, path-scatter was not visible in the menu, there are some investigation to do for that issue ???

interpolate, we could ask the Interpolate author (and some other user/dev) if he/them thinks that's a benefit.

The fact than interpolate can manage special object, is a good point, and should be committed ???

pathscatter does not showup because of an uppercase S in pathscatter.py
here is the corrected file.
The Makefile.am should also be updated in svn.

Tested it again, with this patchs, the scatter script is now in menu, but when I execute it (selecting two simple paths), only click on live preview return me the following error in an error requester window (Inkscape has received additional data from the script executed. The script did not return an error, but this may indicate the results will not be as expected.):

  File "/tmp/ink_ext_XXXXXX.svgVJAN6T", line 1
    <?xml version="1.0" encoding="UTF-8" standalone="no"?>
    ^
SyntaxError: invalid syntax

For testing, the scripts of JFB are available here :
http://math.univ-lille1.fr/~barraud/Inkscape/pathdeform/

It's good to see the progress being made on this one, however given that it is just medium importance and seems to require more than a simple patch to fix it, this can be delayed to 0.47.

jfb, you now have commit rights, can you please commit these fixes if they are tested enough, and close this bug?

thanks!

Hi!

Oh, I forgot this one... Well, the only thing that is not commited yet is
about the interpolate extension:
I proposed to sort the z-order of the generated objects according the
z-order of the input, but someone said the actual situa

2008/4/14, bbyak <email address hidden>:
>
> jfb, you now have commit rights, can you please commit these fixes if
> they are tested enough, and close this bug?
>
> thanks!
>
>
> --
> extensions mix selection order
> https://bugs.launchpad.net/bugs/182832
> You received this bug notification because you are a direct subscriber
> of the bug.
>

Hi!

Oh, I forgot this one... Well, the only thing that is not commited yet is
about the interpolate extension:
I proposed to sort the z-order of the generated objects according the
z-order of the input, but I am not definitely sure it is a good thing.

However, all these were only quick work around: I maid some scripts
independant of the selection order, but the selection order is still
occasionally reverted...

I see it is marked as "work in progress". I think I messed this up a bit; I
am sorry. It was only a work-around in progress... Someone who knows
extension code better than I do will solve this more efficiently.
I think I should close the bug and reopen a new one? Is this ok?
--jfb.

2008/4/14, bbyak <email address hidden>:
>
> jfb, you now have commit rights, can you please commit these fixes if
> they are tested enough, and close this bug?
>
> thanks!
>
>
> --
> extensions mix selection order
> https://bugs.launchpad.net/bugs/182832
> You received this bug notification because you are a direct subscriber
> of the bug.
>

Is this actually fixed, or are there just workarounds implemented for the affected extensions? I ran into this myself when writing my own extension. Unless it's really fixed I'd like to reopen it (or write a new bug).