353575 - gradients, clipPaths and masks inside svg:symbol aren't rendered

Status: RESOLVED FIXED

(Core :: SVG, defect)

Description

[Takeshi Kurosawa :takenspc]

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a1) Gecko/20060918 Minefield/3.0a1
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9a1) Gecko/20060918 Minefield/3.0a1

gradients inside svg:symbol aren't rendered. gradients outside svg:symbol are rendered.

Reproducible: Always

Steps to Reproduce:
1.Open a svg file which has gradient(s) inside svg:symbol aren't rendered.
2.
3.

Actual Results:  
gradient(s) are not rendered.

Expected Results:  
gradient(s) are rendered.

Comment 1

[Takeshi Kurosawa :takenspc]

Attachment: testcase

4 rects shuld have same gradient.

Comment 2

[Ria Klaassen (not reading all bugmail)]

Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/20060920 Minefield/3.0a1
Confirmed.

Comment 3

[Jonathan Watt [:jwatt]]

Why would you put the gradient _inside_ the symbol? The elements in the symbol are probably cloned into a shadow tree, in which case you're increasing memory use (and there are going to be multiple shadow elements with the same ID).

Comment 4

[Takeshi Kurosawa :takenspc]

(In reply to comment #3)
> Why would you put the gradient _inside_ the symbol?

Adobe Illustrator 12.0.1, SVG Export Plug-In exports svg files which include the gradient inside the symbol.

For example, http://spreadfirefox.jp/foxkeh/downloads/parts/tail01.svg
another examples at  http://spreadfirefox.jp/foxkeh/downloads/parts/
# A mascot in those pages is Foxkeh, a mascot of Firefox brought by Mozilla Japan.

Comment 6

[jonathan chetwynd]

#3 jwatt asks "Why would you put the gradient _inside_ the symbol?"

one reason is authoring tool user interface.
beyond hand coding...

ie unless a number of icons are saved as a set at one time, each will have the gradients, then when the icons are agglomerated, there is much duplication:
http://www.openicon.org/icon-ark/freedesktop/gnome/devices.svgz

Comment 7

[adeveria]

For the SVG-edit project, I'm working on arbitrary SVG importing. The idea is to insert a <symbol> element with the data, so multiple copies can be made using <use>. This bug means that import files with gradients will thus appear filled in black when using Firefox.

As a workaround I was going to put the imported <svg> in a <defs>, then have the <use> elements refer to that, but unfortunately Adobe Illustrator incorrectly displays an image like this (it does fine with <symbol>). A bug on their part, certainly, but it forces me to decide whether to support Firefox or Illustrator. 

Considering how <symbol> is very very similar to <svg> elements (in which this bug does not occur), I find it hard to imagine that this bug would be hard to fix.

Comment 9

[Kohei Yoshino]

Just FYI: The Foxkeh assets, including the broken tail SVG files in comment 4, have been moved to
http://www.foxkeh.com/downloads/parts/

Comment 10

[Brian Birtles (:birtles)]

Attachment: Test case with overlapping IDs and hidden groups

A similar problem occurs when we have overlapping IDs and the first occurrence is within a hidden group. This crops up in Santa's Workshop and results in the snowflakes, trees and elves appearing partially or wholly black:

http://ie.microsoft.com/testdrive/Performance/SantasWorkshop/Default.xhtml

The attached test case is a simplified version of what's happening in Santa's Workshop.

I suspect the root cause is the same as this bug. Namely, if a gradient first appears in a context that isn't rendered we seem to refuse to render it from there on.

The fact that Illustrator is exporting this sort of arrangement raises the importance of this bug as does its use in Santa's Workshop.

Comment 11

[Robert Longson [:longsonr]]

The display:none is a different issue, bug 376027 tracks that.

Comment 13

[Kirby]

It has now been a year and eight months since the last entry in this log, and seven years (SEVEN YEARS!) since this bug was first reported and still not been fixed.  WTF??  I use Illustrator.  I create SVG files.  Some of those files have symbols in them.  And the only browser that can't render them with gradients is Firefox.  I don't want a work-around, I don't want a temporary solution: I want to be able to create a symbol with a gradient inside of it and have it render correctly.  After seven years, this is too much to ask of Firefox??  Even an outdated version of Internet Explorer can handle this job.  But not Firefox.  Brilliant.  Seven years.  Wow.  Time to give up and move on to some other browser, I guess.

Comment 14

[Brian Birtles (:birtles)]

Kirby, please keep comments to those which help fix the bug. For progress updates, please follow bug 376027. My previous patch for that bug also fixed this bug (since they have the same root cause) but was unfortunately rejected due to questions raised about the validity of the spec. I have investigated an alternative approach and written it out in detail so that another contributor can implement it. If no one else is available I will try to implement it soon.

Comment 15

[Kirby]

(In reply to Brian Birtles (:birtles) from comment #14)
> Kirby, please keep comments to those which help fix the bug. For progress
> updates, please follow bug 376027. My previous patch for that bug also fixed
> this bug (since they have the same root cause) but was unfortunately
> rejected due to questions raised about the validity of the spec. I have
> investigated an alternative approach and written it out in detail so that
> another contributor can implement it. If no one else is available I will try
> to implement it soon.

Brian: I am sorry my previous message came out sounding so angry and frustrated.  I wish you the best of luck in finding a solution to this problem— know that I am continuing to monitor it with fingers-crossed.

Comment 16

[andy]

What "Kirby" said above ^^^ I agree with that.

I registered just to post this comment.

Comment 17

[guilherme]

I have a similar issue, described here http://stackoverflow.com/questions/32211408/the-svg-use-tag-not-rendering-gradients-firefox, this would be very nice if resolved.

Comment 18

[Robert Longson [:longsonr]]

That's a different issue and invalid to boot as I've explained in my answer there.

Comment 25

[Hunter Jennings]

(In reply to Kirby from comment #15)

(In reply to Brian Birtles (:birtles) from comment #14)

Kirby, please keep comments to those which help fix the bug. For progress
updates, please follow bug 376027. My previous patch for that bug also fixed
this bug (since they have the same root cause) but was unfortunately
rejected due to questions raised about the validity of the spec. I have
investigated an alternative approach and written it out in detail so that
another contributor can implement it. If no one else is available I will try
to implement it soon.

Brian: I am sorry my previous message came out sounding so angry and
frustrated. I wish you the best of luck in finding a solution to this
problem— know that I am continuing to monitor it with fingers-crossed.

Hey Kirby, I was experiencing this exact issue...in 2020. Good news is that there seems to be a workaround, here's a code pen that some created which illustrates the technique: https://codepen.io/RwwL/pen/GgaGgV.

Long story short, you need to define the clip paths outside of the symbol, once you've done that you should be able to reference that symbol with the use element. Not ideal, but definitely let me still use symbols for my icon sprites! Hopefully this is helpful for your use case

Comment 26

[Sergey]

Update on the bug reproducing:

This SVG file will be proper rendered in the last version of Chrome, IE (exept Firefox):

<svg id="custom-shape" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 50 50">
        <symbol viewBox="0 0 50 50" id="sun">
            <defs>
                <radialGradient id="grad1" >
                    <stop stop-color="#EFDC75" offset="0%" />
                    <stop stop-color="#B6A01A" offset="100%" />
                  </radialGradient>
            </defs>
            <circle cx="25" cy="25" r="25" fill="url(#grad1)" />
        </symbol>
        <use xlink:href="#sun" x="0" y="0" width="50" />
</svg>

---

For this moment the solution is to use defs outside a symbol:

<svg id="custom-shape" xmlns="http://www.w3.org/2000/svg" viewBox="0 0 50 50">
            <defs>
                <radialGradient id="grad1" >
                    <stop stop-color="#EFDC75" offset="0%" />
                    <stop stop-color="#B6A01A" offset="100%" />
                  </radialGradient>
            </defs>
        <symbol viewBox="0 0 50 50" id="sun">
            <circle cx="25" cy="25" r="25" fill="url(#grad1)" />
        </symbol>
        <use xlink:href="#sun" x="0" y="0" width="50" />
</svg>

---

For dynamically created SVG I add symbols with their design incapsulated. For Firefox I need adittionally manage common <defs>. I think it is not convinient.

Please, pay attention on this bug.

Comment 27

[Emilio Cobos Álvarez (:emilio)]

I think this should be easy to fix, actually.

Comment 28

[Emilio Cobos Álvarez (:emilio)]

Attachment: Bug 353575 - Allow IDTracker to look up elements in <svg:use> shadow trees. r=smaug

Other UAs allow this, and it seems in the general consensus of
https://github.com/w3c/webcomponents/issues/179.

This matches WebKit's behavior. Blink, for some reason shows red on the
test-case, probably because they're not doing quite this, but they
manage to render masks inside the display: none symbol element or such.

Comment 29

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/8efff752175e
Allow IDTracker to look up elements in <svg:use> shadow trees. r=smaug

Comment 30

[Stefan Hindli [:stefan_hindli]]

https://hg.mozilla.org/mozilla-central/rev/8efff752175e