Bug 45976

Summary: In GraphicsContext, drawImage's use of -1 for non-specified widths/heights disallows usage of -1
Product: WebKit Reporter: Matthew Delaney <mdelaney7>
Component: Layout and RenderingAssignee: Matthew Delaney <mdelaney7>
Status: RESOLVED FIXED    
Severity: Normal CC: abarth, commit-queue, dglazkov, eric, kling, krit, mitz, oliver, pdr, simon.fraser, webkit.review.bot
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Mac   
OS: OS X 10.5   
Attachments:
Description Flags
Negative 1 Source and Dest Test Cases
none
Patch none

Matthew Delaney
Reported 2010-09-17 10:57:33 PDT
I noticed that the underlying canvasrenderingcontext2d drawImage and drawImageBuffer calls used the value -1 to in-band signal when the width/height wasn't specified and thus should be interpreted as image.width and image.height respectively. This causes the case where the actual width/height params come in as -1 from the call to context.drawImage to render incorrectly.
Attachments
Negative 1 Source and Dest Test Cases (8.03 KB, application/zip)
2010-09-17 11:39 PDT, Matthew Delaney 		no flags
Details
Patch (14.14 KB, patch)
2010-09-17 12:54 PDT, Matthew Delaney 		no flags
DetailsFormatted DiffDiff
View All Add attachment proposed patch, testcase, etc.
Simon Fraser (smfr)
Comment 1 2010-09-17 11:22:04 PDT
Testcase?
Matthew Delaney
Comment 2 2010-09-17 11:39:27 PDT
Created attachment 67925 [details] Negative 1 Source and Dest Test Cases Two test cases. One showing the problem of having source width and height as negative 1 and the other with the dest width and height as negative one.
Matthew Delaney
Comment 3 2010-09-17 12:54:00 PDT
Created attachment 67940 [details] Patch
Oliver Hunt
Comment 4 2010-09-17 15:32:26 PDT
Comment on attachment 67940 [details] Patch r=me
WebKit Commit Bot
Comment 5 2010-09-18 14:47:13 PDT
Comment on attachment 67940 [details] Patch Clearing flags on attachment: 67940 Committed r67792: <http://trac.webkit.org/changeset/67792>
WebKit Commit Bot
Comment 6 2010-09-18 14:47:18 PDT
All reviewed patches have been landed. Closing bug.
WebKit Review Bot
Comment 7 2010-09-18 16:02:24 PDT
http://trac.webkit.org/changeset/67792 might have broken GTK Linux 64-bit Debug
Dimitri Glazkov (Google)
Comment 8 2010-09-18 16:43:24 PDT
This broke a bunch of SVG tests (pixel results, so not really seen in the waterfall). You can see failures on Chromium Tests build bots that do run pixel tests. For example: http://test-results.appspot.com/dashboards/flakiness_dashboard.html#upstreamWebKit=true&tests=svg%2Ffilters%2FfeTile.svg&showExpectations=true&showLargeExpectations=true It looks like they just stopped drawing.
Dimitri Glazkov (Google)
Comment 9 2010-09-18 16:49:23 PDT
Reverted r67792 for reason: Broke over a hundred SVG tests. Committed r67799: <http://trac.webkit.org/changeset/67799>
Philip Rogers
Comment 10 2013-02-09 19:16:13 PST
This was fixed independently with http://trac.webkit.org/changeset/139911
Note You need to log in before you can comment on or make changes to this bug.