380338 - Transformations affect path construction

Status: RESOLVED INVALID

(Core :: Graphics: Canvas2D, defect)

Description

[Hixie (not reading bugmail)]

STEPS TO REPRODUCE
 1. Create a path.
 2. Set a transformation.
 3. Fill the path.

ACTUAL RESULTS
 Transformation is ignored.

EXPECTED RESULTS
 Transformation should have an affect.


STEPS TO REPRODUCE
 1. Set a transformation.
 2. Create a path.
 3. Undo the transformation.
 4. Fill the path.

ACTUAL RESULTS
 Transformation is applied.

EXPECTED RESULTS
 Transformation should have no affect.

Comment 1

[Hixie (not reading bugmail)]

TESTCASES
   http://www.hixie.ch/tests/adhoc/html/canvas/007.html
   http://www.hixie.ch/tests/adhoc/html/canvas/008.html
   http://www.hixie.ch/tests/adhoc/html/canvas/020.html
   http://www.hixie.ch/tests/adhoc/html/canvas/032.html

(Some of these may also expose other bugs.)

Comment 2

[Philip Taylor]

I believe bug 338640 is relevant here.

Comment 3

[Takeshi Kurosawa :takenspc]

Attachment: Possible fix

When the path construct, reset the transform
# current transform doesn't affect path construction.

Before the path fill/stroke/clip, copy the path from context, then add context the path.
# current transform affect path construction.

After the path fill/stroke/clip, reset the transform, then add context the path again.
# current transform doesn't affect path construction.

Comment 4

[Takeshi Kurosawa :takenspc]

Attachment: Patch

Patch with correct |diff| options and some comments.

Comment 6

[Philip Taylor]

I think this is the most significant canvas interoperability issue now, and believe it should be resolved (one way or another) before Firefox 3.

Opera implements what the spec says - transformations are applied at fill/stroke/etc time, not at path construction time.

Firefox and Safari don't - transformations are applied at path construction time. (See also <http://bugs.webkit.org/show_bug.cgi?id=13669>.)

It would be great if FF3, Opera 9.5 and Safari 3 could all agree and work the same in this case.

Comment 7

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

I'm pretty sure we had a discussion on the whatwg list about this and decided that what fx and safari are doing is correct (the spec is very wrong; applying the transform at draw time isn't very useful), but I'll post to the whatwg mailing list to verify that before WONTFIX'ing this.

Comment 8

[Hixie (not reading bugmail)]

No decision was made that I can recall. The last I remember discussing this, I was leaving it up to the browser vendors to come to a decision. (Their are arguments in favour of both techniques, the spec/Opera and Webkit/Safari.)

Comment 9

[:Gavin Sharp [email: [e-mail]]]

Renominating given comment 8.

Comment 10

[Vladimir Vukicevic [:vlad] [:vladv] (needinfo me, slow to respond)]

Plusing this but only as a reminder to send mail to the whatwg list; I am 95% certain that our behaviour won't change.

Comment 11

[Hixie (not reading bugmail)]

Since you won't change, I've changed the spec instead... I guess Opera will have to change to match.

Comment 12

[Takeshi Kurosawa :takenspc]

The spec has been changed.
http://html5.org/tools/web-apps-tracker?from=1157&to=1158
http://www.w3.org/TR/2008/WD-html5-20080122/#transformations0

-> INVALID

Comment 13

[Takeshi Kurosawa :takenspc]

Comment on attachment 271053 [details] [diff] [review]
Patch

cancel the review request: this bug was marked as invalid.