Hi Friends,
I had trouble once before applying a gradient to a group of circles. I could make it work just fine with squares, but there seemed to be some problem with circles. And I actually reported the problem at Launchpad, but it didn't seem to attrack much attention .

So here I am again, trying to apply a gradient to a group of circles, and having trouble. I want to give you an SVG file to examine, but it's too big to upload to the forum, and I'm having trouble with my usual SVG host. So I can't do that at the moment, but hopefully soom.

I do have a screen shot, to show my Inkscape window with all the various settings.

So for the time being, does anyone know about a bug with applying a gradient to a group of circles (btw, I did search Launchpad, but can't even find my own bug with a search -- or any others on this topic). OR does anyone see anything wrong with the setting I'm showing in my Inkscape window?

Thanks for your help -- I'll have the SVG uploaded asap.
All best

brynn wrote:(btw, I did search Launchpad, but can't even find my own bug with a search -- or any others on this topic)

Go to your Launchpad page, click on Bugs > Advanced search, select all status options and unselected 'hide duplicates' > Search: you get a list of all bugs you ever reported and commented ;-)

brynn wrote:I had trouble once before applying a gradient to a group of circles. I could make it work just fine with squares, but there seemed to be some problem with circles.

I tried to reproduce this (with 0.46 and 0.47pre) and had the same difficulties applying a gradient to a group of circles or stars - but only if I made them with the 'Create Tiled Clones…' tool. OTOH - 'draw a circle, create a cloud by stamping (drag + <SPACE>), unset the stroke of all circles, group them and apply gradient to the stroke of the group' works for me.

Here's a workaround that lets me apply the gradient to the group, not the individual circles:

1. create original circle, convert it to a path
2. use the 'Create Tiled Clones…' to create the array of circles you need
3. select all clones and unlink them, (re)move the unneeded original circle
4. keep all circles (which are really paths with 4 nodes) selected, open 'Fill and Stroke'
5. remove 'Fill', unset 'Stroke' (the question mark button on the far right side, after the pattern icon)
6. group the (still) selected circles
7. apply gradient to the stroke of the group, adjust stroke width as needed

There seems to an underlying issue with 'Inkscape shapes' (circle, star), clones and inherited styles in groups - if I find a related bug report I will add a link here.

hth, ~suv

[edit date="Wed Sep 30, 2009"]
removed detailed links to user specific launchpad pages - my apologies to brynn for adding them in the first place
[/edit]

I haven't investigated this issue myself, but if it only happens with clones, it's probably not a bug. Clones always display the fill and stroke properties of the original shape unless the original's are set to Unset. In other words, if you change a fill or stroke property on a clone it will ignore that unless you set the original's corresponding property (fill or stroke) to Unset. That's how clones work. It makes them cumbersome to work with for styling purposes, but that's how it is.

Thanks ~suv
Yes, I did create my circles originally with Tiled Clones. I don't remember if I converted to Path before tiling...probably not. So I will repeat all my steps, using a circular path rather than a circle for the original tile.

Thanks for your efforts to reproduce my problem, despite lack of SVG file to examine (still troubleshooting host issue).

Go to your Launchpad page, click on Bugs > Advanced search, select all status options and unselected 'hide duplicates' > Search: you get a list of all bugs you ever reported and commented

Oh, well yes, I do know how to locate my own bugs; I actually still have my original email with the link. I was more commenting on the difficutly I have searching at Launchpad. Having filed a bug on the inability to apply a gradient to a group of circles, and hoping to find other bugs related to it, I had expected a search for 'gradient, group, circles' would find it....but no...

Oops, kelan, you posted while I was typing. Thanks!
I do know that if you're going to use the Color tab of the Tiled Clones dialog, you have to set the fill and stroke of the original to Unset. But I thought that once I broke the link of the clones, they should behave like regular circles, as before cloning. Doesn't that make sense?

Thanks again kelan and ~suv for your comments; I really appreciate your efforts.
All best

brynn wrote:I was more commenting on the difficutly I have searching at Launchpad. Having filed a bug on the inability to apply a gradient to a group of circles, and hoping to find other bugs related to it, I had expected a search for 'gradient, group, circles' would find it....but no...

confirmed, the launchpad user interface doesn't offer a premium search experience ;-)

Oftentimes I resort to Google with the search restricted to the bug tracker, like this: (your comments are part of the first bug report listed in the current search result... ;-)

Any success with the gradient on grouped circles yet?

I haven't found a bug report that points to identical issues with Inkscape shapes and gradients as stroke attributes, but keep actively looking whenever I dig into a bug search (yes, with google - I'm not not that impressed by launchpad's interface ;-)

It could have its cause in the special way (or workaround with regard to the SVG specs) that Inkscape stores the 'shape' information and its transformations as sodipodi attributes. I'm still trying to create a reduced case that demonstrates the problem with a minimal number of elements and that highlights the different gradient rendering between groups of (unlinked tiled cloned) paths or rectangles and groups of (unlinked tiled cloned) shapes like circle, star or spiral.

~suv

It seems to be working fine, ~suv. Once I used the circular path rather than the circle shape, I can apply the gradient as I wanted.

I'm still trying to create a reduced case that demonstrates the problem with a minimal number of elements and that highlights the different gradient rendering between groups of (unlinked tiled cloned) paths or rectangles and groups of (unlinked tiled cloned) shapes like circle, star or spiral.

Why? Is it just finding the time to do it? Because I wouldn't mind making up a sample file with tiny arrays of like 8 to 10 tiled clones of paths and objects of each shape for you. Although....you wouldn't necessarily need to use tiled clones to illustrate the problem, since it occurs in circle shapes that have not been tiled and cloned.

Or I guess you maybe have a better idea of what is needed? But if I can be of any help, please feel free to let me know

brynn wrote:Why?

I like the "detective" work - that's how I learn about Inkscape ;-) and I want to get as close to what reproducibly causes the problem as possible (i.e. a minimal case).

brynn wrote:Is it just finding the time to do it?

partly time, partly time management or getting lost… I am not well organized.

brynn wrote:Because I wouldn't mind making up a sample file with tiny arrays of like 8 to 10 tiled clones of paths and objects of each shape for you.

Go ahead - it was not my intention to take this off of you ;-)

brynn wrote:Although....you wouldn't necessarily need to use tiled clones to illustrate the problem, since it occurs in circle shapes that have not been tiled and cloned.

See - that worked for me when I previously searched for a solution, but if it doesn't work for you then my assumption was wrong. OTOH I have seen inconsistent results depending on whether 'Object to Path' was applied before or after cloning and unlinking - I wanted to re-test this out of mere curiosity…

Go ahead - it was not my intention to take this off of you

Oh no, that's not even the issue. I would very much appreciate if you could make a better report that I did back then. And I'm sure that you can. I was just offering to help, to save you some time by creating the sample image document. I don't know, maybe it's better for one person to do the whole thing? In that case, I think you're much better qualified. But, if it wouldn't affect the integrity of the report, for me to make the example image, and you to write the report, then I'd be glad to take that chore off your hands

On 2nd thought though, when you say "minimal case", you're probably better equipped. I do know what minimal case means, but I think your detective work is the most important part, and depending on what you learn, it's going to affect what needs to be in the example image. So for the sake of the integrity of the report, I think it's better left in your capable hands. Just know that if there is anything I could do to help, I would be glad to

…getting closer? ;-)

Nice!
So you found what's causing the problem -- good work!