Closed Bug 726324 Opened 14 years ago Closed 14 years ago

Error in SVG Animation with <use> Tag if object not fully visible at the start

Categories

(Core :: SVG, defect)

defect
Not set
normal

Tracking

()

RESOLVED INVALID

People

(Reporter: mbchnr, Unassigned)

Details

(Keywords: regression)

Attachments

(2 files, 1 obsolete file)

Attached file TestSVG.zip (obsolete) —
User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:10.0) Gecko/20100101 Firefox/10.0 Build ID: 20120129021758 Steps to reproduce: I created an SVG animation with an object partly rotated into view, i. e. only partly visible at the start. Actual results: This works fine with a normal object. If the object is replaced with a <use>-tag, referencing a <symbol>, everything not visible at the start will be sliced away. Expected results: Replacing an object with a <use>-tag should not bring up any other result than before. Anyway, it should be possible to have animated objects start outside the visible area.
Component: Untriaged → SVG
OS: Windows 7 → All
Product: Firefox → Core
QA Contact: untriaged → general
Hardware: x86_64 → All
Version: 10 Branch → Trunk
Do you think you could add the testcase as individual attachments that work directly in bugzilla please? It's much easier to work with.
Attachment #596354 - Attachment is obsolete: true
Okay, I have put it all in one file, but the link only opens as text... I understand now that the symptom can be fixed if a larger canvas size is given in the svg tag, hence I am not sure if it's a full bug. It occurs only if the size is defined too small to hold the entire objects, which is true also if no size is given (it defaults to the size of the container, I presume, the div in this case).
Attachment #596404 - Attachment mime type: text/plain → text/html
The transform updates seem to be getting though to the path geometry frames OK and causing UpdateGraphic calls there but nothing is shown my trunk build at all neither the use nor the ordinary rect. Seems to work OK in Firefox 10. Any ideas Jonathan?
Status: UNCONFIRMED → NEW
Ever confirmed: true
reverting to http://hg.mozilla.org/mozilla-central/rev/ea2f8a85e0ed before bug 725903 and bug 725897 landed makes the rectangle on the right display but not the one on the left. Looks like there's two regressions here.
Keywords: regression
The rect in the symbol is not drawn because the dirtyRect (0 0 18000 9000 in app units) does not intersect with the rect in the symbol(10200 10200 3000 3000)
Hi Michael, thanks for the bug report. Just to clarify, are you seeing this problem in Firefox 10, or a beta/aurora/nightly build? I ask because comment 0 says you are using Firefox 10, but I can't reproduce the problem there. (I can reproduce it in beta, aurora and nightly.)
(In reply to Michael from comment #3) > It occurs only if the size is defined too small to hold the entire objects, > which is true also if no size is given (it defaults to the size of the > container, I presume, the div in this case). Actually, no, it defaults to 150px x 300px according to the CSS fallback rules. At least since bug 668163, which is now on beta and aurora. This is responsible for the clipping in the second testcase. To make the SVGs the size of their <div> you need to give them an explicit width/height now.
(In reply to Robert Longson from comment #5) > reverting to http://hg.mozilla.org/mozilla-central/rev/ea2f8a85e0ed before > bug 725903 and bug 725897 landed makes the rectangle on the right display > but not the one on the left. Looks like there's two regressions here. Really? Other than the clipping away of the bottom of the <svg>s in the black bordered <div>s (due to comment 8), I don't see any additional problems at all with builds made at, before and after the changesets of any of my recent changes. In any case, can we have a separate bug for that?
Attached image simpler testcase —
It seems this bug (as originally) report is invalid. To understand why, you need to read the section "If the ‘use’ element references a ‘symbol’ element" here: http://www.w3.org/TR/SVG11/struct.html#SymbolElement The implicit <svg> that the <symbol> gets replaced by is given a default width/height of 100%, so it gets the same dimensions as the outer-<svg>. The problem seems to be that the <rect> is outside/partially outside that area. Animating the <g> to move the position of the <use> it contains does not change this fact.
(In reply to Robert Longson from comment #5) > reverting to http://hg.mozilla.org/mozilla-central/rev/ea2f8a85e0ed before > bug 725903 and bug 725897 landed makes the rectangle on the right display > but not the one on the left. Robert, I'm still very interested to know more about what you're seeing here.
(In reply to Jonathan Watt [:jwatt] from comment #7) > Hi Michael, thanks for the bug report. Just to clarify, are you seeing this > problem in Firefox 10, or a beta/aurora/nightly build? I ask because comment > 0 says you are using Firefox 10, but I can't reproduce the problem there. (I > can reproduce it in beta, aurora and nightly.) Hi Jonathan, oh no I just stumbled in here and don't know anything about betas or such. I'm using the 'normal' 10.0.1 actually and I do see the clipping there.
I think we're talking about different clipping. If you try with the current beta build, you'll see that the bottom of the SVG in the div with the black outline is clipped away. The reason for that is explained in comment 8. https://www.mozilla.org/firefox/beta/
The clipping that I think you're talking about is the correct behavior, and is explained in comment 11.
Ah, I see!
(In reply to Jonathan Watt [:jwatt] from comment #12) > Robert, I'm still very interested to know more about what you're seeing here. Apparently nothing. I've done a clean rebuild and I see what you see now. Sorry for any panic caused.
Is this INVALID then, per comment 14-15?
Yes
Status: NEW → RESOLVED
Closed: 14 years ago
Resolution: --- → INVALID
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: