Re: [Gimp-gui] About the "floating selection"

[Jehan <jehan girinstud io>]

On 12/3/20 12:45 AM, Liam R E Quin wrote:
> On Thu, 2020-12-03 at 00:10 +0100, Jehan wrote:
> > - When pasting on a layer mask, still make a floating selection (but
> > probably name it "Floating mask" because the name "Floating
> > *selection*" is weird anyway).
> How much work would it be to support multiple masks, and have paste
> create a new additional mask?
> It'd need
> move up, move down, merge down, delete
> and a UI to choose which one you're editing.
> Is it worth the effort?

It would definitely be nice. And I believe I even saw a report asking 
for this not so long ago. And I'm sure it would be worth the effort for 
whoever has intense usage of masks.

In our case, we do use masks, but not enough for me to put too much 
priority on this. I mean, I might implement if I got very bored some 
day, but don't count too much on it (not for GIMP 3.0 at least). I 
prefer to go the easy way and leave a "floating mask" concept as a 
remnant of the old "floating selection" until someone decides to 
implement it. 😛

> Alternatively, your suggestion is fine, but may be even harder for
> people who are now certain to be unfamiliar with floating layers.
Well someone who starts using mask has most certainly already some good 
usage where they at least read/watch some tutorials, books or classes or 
whatever else with GIMP. It's not a beginner feature. So I don't think 
that this actually counts the same as the current usage of floating 
selection for something as basic as layers.
> Maybe draw a thick green border round a floating selection with an on-
> canvas anchor at lower right???
>
> Getting rid of floating selections as much as possible seems worth it;
> retaining them, with all the code and GUI, for one fairly obscure (but
> awesome) use case seems not as good. E.g. there's an anchor button on
> the bottom of the Layers dock which would only become active when
> pasting into a mask.
>
> How about just getting rid of floating selections altogether, with the
> awareness that we want multiple masks in the future?

We know for a fact that removing features do not go so well and there is 
always someone using it who will come and complain (relevant XKCD: 
https://xkcd.com/1172/). Also if it were something actually useless/bad, 
I may understand, but here it's actually useful (for whoever uses masks 
intensively, and many people do).

Also for years, when we had been asking about floating selections to 
Mitch, he would tell me this use case of pasting to masks, and I never 
really understood what he meant (because, same as Alex, I was like "well 
just merge into the mask if it is selected") until today. So pretty sure 
he won't agree either. And really me too, I don't want now that I got 
it. It's not a good idea to remove preventively before we actually 
implemented the alternative.

In any case the code already exists and it works at least. It won't be 
hard to keep them. Sure I would have preferred to remove it all from a 
maintenance point of view. But it's not like it would be insanely hard 
otherwise.

Jehan

> ankh (Liam)
--
ZeMarmot open animation film
http://film.zemarmot.net
Liberapay: https://liberapay.com/ZeMarmot/
Patreon: https://patreon.com/zemarmot
Tipeee: https://www.tipeee.com/zemarmot