GIMP Version: 2.8.18
Operating System: Windows
GIMP Experience: Basic Level

List any relevant plug-ins or scripts:
my filter Mosaic_Lyle_Style

List any ERROR messages you received:
no error, simply wrong transparency values

---

After a while, stimulated by other Gimp chatters, I wanted to re-visit that filter an create an outcome.
I got a very strange outcome where the pieces of the mosaic did not appear correctly.
I started debugging the filter stopping at a point where the texture for the pieces was prepared, and I noticed that the "points" created by stroking (the filter uses a created brush to stroke using points and not lines) had different alpha values: some 32, some 60 or so, other close to 255.
In the PDB it is written that the gimp-edit-stroke and correlated ...stroke-vectors is influenced (only) by the foreground colour, which was set to (0,0,0,255), thus pure black w/o transparency.
Any idea?
(for those who like to enter into the code, the code is available under Gimp Scripts Plug-ins)

BTW, to succeed in creating a mosaic look, I had to add a trick to scan all pixels of the texture layer and reset the transparency to 255.
Did any other user of my filter have similar problems? If so, I have to post an amended version, but before I would like to understand where the cause of this strange Alpha values lies

Link to plugin code?

http://gimpscripts.com/2016/02/diego-mosaic-lyle-style/
download the update

I've always got good results with this filter, Dinasset. Is this last filter you posted, the one you have mended?

I guess I discovered the cause.
In setting the characteristics of the brush, I didn't set among the others the "hardness" (likely because I was duplicating "2. Hardness 100" std brush and I wrongly supposed that the duplication did it for me).
And it could be that -for some previous online adjustment for other tests- I had modified the "opacity" of the active brush.
So I presume that the newly created brush this morning assumed the active opacity.
And this explain also why up to now I didn't have that problem (very likely I keep the default opacity to 100 in 99.9% of my plays).

I guess I discovered the cause.
In setting the characteristics of the brush, I didn't set among the others the "hardness" (likely because I was duplicating "2. Hardness 100" std brush and I wrongly supposed that the duplication did it for me).
And it could be that -for some previous online adjustment for other tests- I had modified the "opacity" of the active brush.
So I presume that the newly created brush this morning assumed the active opacity.
And this explain also why up to now I didn't have that problem (very likely I keep the default opacity to 100 in 99.9% of my plays).

Issabella, no it's the official one, for Ofnuts.
Thanks for your input.
Now please double-check my hypothesis just written above.
Set opacity of your Paintbrush to something like 30, instead of the std 100.
After that rerun the filter and see whether you get wrong results.
If so, my guess is correct.

I see it alright. I set the brush at opacity 30%

you set opacity at 30 BEFORE running the filter?

Yes I did.

so, that behaviour is only on my PC...

thanks Issabella for your tests

double-checked me too:
restored the official version (w/o hardness)
called the filter having std opacity at 100: OK as expected
called the filter having opacity set at 30: still OK! what did happen this morning??? (and I tried more than once, there must be some other environment variable in the play)

We all can continue testing it with different opacities and if strange behaviour, tell you. No problem for now

Set opacity to 30% and ran the filter default settings.

thanks Rod.
What happened to me Yesterday morning is still a "mystery" for me...