1293833 - <use>-referenced circle isn't in the expected spot, in mask within <symbol>

Status: RESOLVED FIXED

(Core :: SVG, defect, P3)

Description

[Daniel Holbert [:dholbert]]

Attachment: testcase 1

STR:
 1. Load attached testcase (based on https://codepen.io/thebabydino/pen/YWRVzG/ )
 2. Look at dots on yin-yang.

EXPECTED RESULTS: Dots should be in middle of swooshes' fattest part, to make a nice yin-yang image.

ACTUAL RESULTS: No light-colored dot is visible; and the dark-colored dot is at the bottom of the image instead of over the white swoosh.

Comment 1

[Daniel Holbert [:dholbert]]

Attachment: reference case 1

Here's a reference case, where I've just replaced the <use> circle-references with actual <circle> elements (with cx and cy set, in place of x/y on the <use> element).

Comment 2

[Daniel Holbert [:dholbert]]

I get EXPECTED RESULTS in Chrome as well as Firefox 37.0a1 (2015-01-01).
I get ACTUAL RESULTS in current Nightly.

So, this is a regression. Range:
https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=29b05d283b00&tochange=d7e156a7a0a6
--> regression from bug 1127507

Comment 3

[Andrew Overholt [:overholt]]

Daniel, given jwatt is out right now, can you take a look at this? Do you think we need to hold 49 for a fix here?

Comment 4

[Robert Longson [:longsonr]]

Doesn't this merely put us back to the broken state we were in pre bug 932771?

Comment 5

[Robert Longson [:longsonr]]

I.e. I think we were broken pre bug 932771, then inadvertently fixed (by introducing a different bug) and now were back to broken again.

Comment 6

[Daniel Holbert [:dholbert]]

(In reply to Andrew Overholt [:overholt] (back Aug 31) from comment #3)
> Daniel, given jwatt is out right now, can you take a look at this? Do you
> think we need to hold 49 for a fix here?

Given comment 2 (and Robert's subsequent posts as well), this has been broken since at least Firefox 38, so I don't think it needs to block Firefox 49 in particular, nor does it need urgent attention before jwatt returns (though it'd still be nice to fix).

Comment 7

[Astley Chen (inactive)]

Jonathan, are you currently working on this bug now ?

Comment 8

[Jonathan Watt [:jwatt]]

(In reply to Daniel Holbert [:dholbert] from comment #6)
> (In reply to Andrew Overholt [:overholt] (back Aug 31) from comment #3)
> Given comment 2 (and Robert's subsequent posts as well), this has been
> broken since at least Firefox 38, so I don't think it needs to block Firefox
> 49 in particular, nor does it need urgent attention before jwatt returns
> (though it'd still be nice to fix).

It also seems a bit obscure (<use> linking to <symbol> containing geometry that is <mask>ed by a mask containing a <use> that links to another geometry element). Do we expect much content to be impacted by this, or enough to make it worth spending time investigating?

(In reply to Astley Chen [:astley] (UTC+8) from comment #7)
> Jonathan, are you currently working on this bug now ?

I'm not currently, but could.

Comment 9

[Astley Chen (inactive)]

(In reply to Jonathan Watt [:jwatt] from comment #8)
> I'm not currently, but could.

Thanks. Sounds good to me if you could fix it. :)

Comment 10

[Robert Longson [:longsonr]]

Fixed by patch in bug 1325865