30055 2009-10-04 03:08:50 -0700 Bad DOM performance in large SVG files 2010-02-17 11:54:03 -0800 1 1 1 Unclassified WebKit Layout and Rendering 528+ (Nightly build) PC Windows XP RESOLVED FIXED http://paroga.com/webkit/rect.php?c=10000&s=10 P2 Normal --- 32815 33012 1 paroga webkit-unassigned abarth bdakin commit-queue dglazkov emacemac7 eric krit oliver simon.fraser tonikitoo webkit.review.bot zimmermann oldest_to_newest 152143 0 paroga 2009-10-04 03:08:50 -0700 When you click on the 1st rect all rects get yellow, on the 2nd they get blue. When you click on the 3rd the first rect gets lime, on the 4th it gets red. Setting only one rect is as slow as setting all rects. In Firefox 3 you see a big difference between setting all or only one rect. With the c-parameter you can change the count of the rects, with the s-parameter the size. Setting only one rect is much faster in http://paroga.com/webkit/rect.php?c=4 than in http://paroga.com/webkit/rect.php?c=10000. In Firefox there is no difference. 152183 1 paroga 2009-10-04 13:01:58 -0700 Most of the time ist spent in: RenderLayer::paintLayer RenderSVGRoot::paint It seems that the complete SVG will be rendered every time. I tested it with the Qt port where the dirtyRect is always the size of the whole frame. 152701 2 paroga 2009-10-06 06:28:27 -0700 The same page as html (HTML:div insted of SVG:rect): http://paroga.com/webkit/div.php?c=4 http://paroga.com/webkit/div.php?c=10000 When only setting a single rect (click on rect 3 or 4) the HTML version is much faster. When changing the color of all rects the SVG version is faster. 153153 3 eric 2009-10-07 12:47:18 -0700 I'm not sure I see the difference. WebKit only seems to be drawing the SVG rects as their added. Using Quartz Debug, I do not see it repainting the whole SVG every time. Bug 14015 is sorta related (drawing too much). I think we need to see a better example to see the performance difference, if there is one. Your example does not convince me that we're drawing too much. However having written the code, I believe you that we're drawing too much. :) 153291 4 paroga 2009-10-07 22:37:11 -0700 (In reply to comment #3) > I'm not sure I see the difference. WebKit only seems to be drawing the SVG > rects as their added. Using Quartz Debug, I do not see it repainting the whole > SVG every time. Have you tried it wit rect.php?c=10000&s=3? Maybay it there to many rects out of the frame, you can't see a difference. It is not the inital rendering, you can see the difference only when click on rect 1, 2, 3 or 4. I think that the dirtyRegion is calculatet correctly, but nothing cares about it (at least in the Qt port) 153502 5 eric 2009-10-08 13:23:54 -0700 OK, so clicking is important, yes. I didn't realize I had to click on them. I tried these just now in Safari 4.0.3 and SVG was correctly redrawing only one svg rect when toggling the color on one rect. If Qt is re-drawing the whole page, that seems like a qt bug. I did not try on tip of tree, but I kinda doubt this regressed... but maybe it did. 153514 6 paroga 2009-10-08 13:44:40 -0700 (In reply to comment #5) 1) open http://paroga.com/webkit/rect.php?c=4&s=3 2) click on the 1st rect and stop the time until the first rect changes its color 3) click on the 3rd rect and stop the time until the first rect changes its color 4) open http://paroga.com/webkit/rect.php?c=10000&s=3 5) click on the 1st rect and stop the time until the first rect changes its color 6) click on the 3rd rect and stop the time until the first rect changes its color 7) time in step 6 is much bigger than in step 3 Both times should be the same, because only one rect changes it's color. When time in step 2 is much bigger than in step 5 this is ok, because in the second page are much more items to be redrawn. You can see the expected result when you open the page with Firefox. 155490 7 41379 paroga 2009-10-18 09:10:27 -0700 Created attachment 41379 Paint RenderPath only when necessary With this path clicking on the 3rd or 4th rect in http://paroga.com/webkit/rect.php?c=10000&s=3 is about 4 times faster on my computer. (time from click event until the first rect changes it's color) I'm not sure if this is the right place for my changes. 156412 8 eric 2009-10-20 23:17:48 -0700 This looks great! I'd like to test the patch myself first before I r+ this though. 158667 9 41379 oliver 2009-10-28 12:52:31 -0700 Comment on attachment 41379 Paint RenderPath only when necessary Why are you checking for intersection prior to applying the local transform to the paintInfo region? 158686 10 paroga 2009-10-28 13:40:14 -0700 (In reply to comment #9) > Why are you checking for intersection prior to applying the local transform to > the paintInfo region? I only tried to find an easy way to demonstate the performance problem. The paintInfo.context->save(); stuff takes a significant amount of time on my machine. In my optinion there must be a better way to handle "DOM repaints" in general. Always iterateing over _all_ items isn't very cool. Maybe you can confirm this issue and tell me the prefered solution. If you take a look at http://paroga.com/webkit/livedemo.php you can see a (not so good) example what people do with SVG. 158714 11 eric 2009-10-28 14:49:24 -0700 This patch is exactly the right idea. I just need to confirm that it's implemented correctly. I need to make space in my schedule to test your patch locally and check to see if other if statements like this one should be added to SVG renderers. 159130 12 eric 2009-10-29 15:56:25 -0700 Ok, I'm not able to reproduce the slowness you mention about clicking on the 3rd or 4th rect in http://paroga.com/webkit/rect.php?c=10000&s=3. That example is actually really poor, because it requires someone to know that they're supposed to click on a special rect to demonstrate the bug. Much better would be a self-documenting page with buttons one could click on which say what they do. I agree we should probably make a change like this. I just don't think we've justified it very well up until now. The pathalogical case that you're trying to demonstrate is a single group containing a large number of path objects (<rect>, etc) where only one of them changes. In current SVG code, all of them will be asked to paint and then the results will be clipped out. It appears that the clipping is rather cheap in CoreGraphics, but whatever platform you are using is slower. 159242 13 paroga 2009-10-30 01:00:14 -0700 (In reply to comment #12) > That example is actually really poor, because it requires someone to know that > they're supposed to click on a special rect to demonstrate the bug. Much > better would be a self-documenting page with buttons one could click on which > say what they do. I've improved the example. 159407 14 paroga 2009-10-30 12:45:26 -0700 (In reply to comment #12) > Ok, I'm not able to reproduce the slowness you mention about clicking on the > 3rd or 4th rect in http://paroga.com/webkit/rect.php?c=10000&s=3. I can't believe you! ;-) I now compiled r50334 with the Qt-port (4.6 beta) on windows and tested the following patch (with g_checkIntersect true and false): > if (paintInfo.context->paintingDisabled() || style()->visibility() ...) > return; > > + FloatRect boundingBox = repaintRectInLocalCoordinates(); > + if (g_checkIntersect && !boundingBox.intersects(paintInfo.rect)) > + return; > + > paintInfo.context->save(); > paintInfo.context->concatCTM(localToParentTransform()); > > SVGResourceFilter* filter = 0; > > - FloatRect boundingBox = repaintRectInLocalCoordinates(); > if (paintInfo.phase == PaintPhaseForeground) { > PaintInfo savedInfo(paintInfo); When I open http://paroga.com/webkit/rect.php?c=80000&s=3 with g_checkIntersect = true it takes about 2 seconds to change the color of the first element. With g_checkIntersect = false the same action takes about 10 seconds on my machine. Maybe the rendering/clipping isn't as time-consuming with your port? Is it possible that you try it with the Qt-port? When i apply this patch to WebCore it can't be a Qt-specific rendering problem?!? 161786 15 oliver 2009-11-09 14:19:36 -0800 > Maybe the rendering/clipping isn't as time-consuming with your port? Is it > possible that you try it with the Qt-port? When i apply this patch to WebCore > it can't be a Qt-specific rendering problem?!? WebCore uses the platform's graphics library, if that graphics library has bad perf in some cases it can easily cause all sorts of badness. I'm unsure whether the solution should be to get Qt to fix this issue (presumably it's a problem in their graphics library) or fix it in WebCore itself (which may negatively impact performance of other platforms) Maybe this logic should be in GrpahicsContextQt? 161979 16 paroga 2009-11-10 01:01:12 -0800 (In reply to comment #15) > WebCore uses the platform's graphics library, if that graphics library has bad > perf in some cases it can easily cause all sorts of badness. I'm unsure > whether the solution should be to get Qt to fix this issue (presumably it's a > problem in their graphics library) or fix it in WebCore itself (which may > negatively impact performance of other platforms) > > Maybe this logic should be in GrpahicsContextQt? Do you think rendering anything ist better than checking if it is necessary? I think WebCore should render as least as possible. In GrpahicsContextQt there won't be any information to check if painting is required. Rely on a "perfect performing" graphics library isn't the right way in opinion. There might be some other ports with slow graphics librarys too. 163779 17 oliver 2009-11-15 15:01:06 -0800 (In reply to comment #16) > (In reply to comment #15) > > WebCore uses the platform's graphics library, if that graphics library has bad > > perf in some cases it can easily cause all sorts of badness. I'm unsure > > whether the solution should be to get Qt to fix this issue (presumably it's a > > problem in their graphics library) or fix it in WebCore itself (which may > > negatively impact performance of other platforms) > > > > Maybe this logic should be in GrpahicsContextQt? > Do you think rendering anything ist better than checking if it is necessary? I > think WebCore should render as least as possible. In GrpahicsContextQt there > won't be any information to check if painting is required. Rely on a "perfect > performing" graphics library isn't the right way in opinion. There might be > some other ports with slow graphics librarys too. The problem is that if the underlying graphics library is already doing the tests it may have a more efficient way of doing so, add to that the fact that the underlying graphics libraries that do perform sensible clipping optimisations will still be doing their own clip tests this seems like a patch that has bad perf impact on good libraries to deal with poor performance in others. I suspect if we really do want webcore to work around what is arguably a perf bug in Qts graphics subsystem the fix should be placed in GraphicsContextQt directly 163828 18 eric 2009-11-15 18:32:46 -0800 I think we can use both approaches, and I think that this approach has some usefulness. It needs to be well tested though. I also agree with Oliver, that if Qt is not properly ignoring drawing commands which are issued outside of the clip region, than that seems like a major bug in Qt. 163831 19 paroga 2009-11-15 18:48:26 -0800 (In reply to comment #17) > The problem is that if the underlying graphics library is already doing the > tests it may have a more efficient way of doing so, add to that the fact that > the underlying graphics libraries that do perform sensible clipping > optimisations will still be doing their own clip tests this seems like a patch > that has bad perf impact on good libraries to deal with poor performance in > others. So this patch has to move to a higher level where "more intelligent" test can be done and the clipping should be fixed in GraphicsContextQt directly. However the following test is much slower(!) in Safari/Chrome than in FF: Click on "Set first lime/red" at http://paroga.com/webkit/rect.php?c=80000&s=2 On my machine: FF ~150ms, Opera ~300ms, Chrome ~2s, Safari ~10s How to get the same performance of FF/Opera in WebCore? Any suggestions? 167320 20 abarth 2009-11-30 12:17:58 -0800 Attachment 41379 passed the style-queue 170456 21 oliver 2009-12-10 10:22:12 -0800 (In reply to comment #19) > (In reply to comment #17) > > The problem is that if the underlying graphics library is already doing the > > tests it may have a more efficient way of doing so, add to that the fact that > > the underlying graphics libraries that do perform sensible clipping > > optimisations will still be doing their own clip tests this seems like a patch > > that has bad perf impact on good libraries to deal with poor performance in > > others. > So this patch has to move to a higher level where "more intelligent" test can > be done and the clipping should be fixed in GraphicsContextQt directly. > > However the following test is much slower(!) in Safari/Chrome than in FF: > Click on "Set first lime/red" at http://paroga.com/webkit/rect.php?c=80000&s=2 > On my machine: FF ~150ms, Opera ~300ms, Chrome ~2s, Safari ~10s > > How to get the same performance of FF/Opera in WebCore? Any suggestions? I don't see this 10s delay you're referring to what platform are you testing on? 170468 22 paroga 2009-12-10 11:12:14 -0800 (In reply to comment #21) > I don't see this 10s delay you're referring to what platform are you testing > on? Safari 4.0.4, Windows XP SP3, IBM T60 2GHz, 3GB 171172 23 41379 oliver 2009-12-12 23:22:35 -0800 Comment on attachment 41379 Paint RenderPath only when necessary I found another testcase where we get pathologically bad perf due to this issue which would imply that there's just too much cost in not clipping prior to calling the graphics library -- my test case just created many many many (~50000 odd) nodes mostly off the screen and this patch represents a huge improvement in that case. Still not sure why i don't see the perf problems you do in your example, but with a testcase that is too slow this shows up as a clear problem in shark. r=me. The patch applies cleanly lets see how the commit bot deals with the changelog. 171174 24 41379 commit-queue 2009-12-13 01:14:20 -0800 Comment on attachment 41379 Paint RenderPath only when necessary Rejecting patch 41379 from commit-queue. Failed to run "['WebKitTools/Scripts/run-webkit-tests', '--no-launch-safari', '--exit-after-n-failures=1', '--quiet']" exit_code: 1 Running build-dumprendertree Running tests from /Users/eseidel/Projects/CommitQueue/LayoutTests Testing 11771 test cases. svg/batik/filters/filterRegions.svg -> failed Exiting early after 1 failures. 10019 tests run. 313.92s total testing time 10018 test cases (99%) succeeded 1 test case (<1%) had incorrect layout 6 test cases (<1%) had stderr output 171175 25 krit 2009-12-13 01:23:31 -0800 (In reply to comment #23) > (From update of attachment 41379 [details]) > I found another testcase where we get pathologically bad perf due to this issue > which would imply that there's just too much cost in not clipping prior to > calling the graphics library -- my test case just created many many many > (~50000 odd) nodes mostly off the screen and this patch represents a huge > improvement in that case. offscreen: It depends. If the SVG size is also bigger than the screensize, we shouldn't clip to the currently visible area. smfr remind me, that we can include a SVG into a HTML-file and make some 3D transformations (via CSS3) with it. That might not work correctly, if the SVG is clipped. 171222 26 oliver 2009-12-13 12:59:51 -0800 (In reply to comment #25) > (In reply to comment #23) > > (From update of attachment 41379 [details] [details]) > > I found another testcase where we get pathologically bad perf due to this issue > > which would imply that there's just too much cost in not clipping prior to > > calling the graphics library -- my test case just created many many many > > (~50000 odd) nodes mostly off the screen and this patch represents a huge > > improvement in that case. > > offscreen: > It depends. If the SVG size is also bigger than the screensize, we shouldn't > clip to the currently visible area. smfr remind me, that we can include a SVG > into a HTML-file and make some 3D transformations (via CSS3) with it. That > might not work correctly, if the SVG is clipped. We should simply be clipping to the region that is going to actually be drawn to, assuming everything is working correctly that is correct -- if i can apply a 3d transform and that will result in incorrect clipping area then thats a serious bug in the 3d transform paths 173849 27 eric 2009-12-20 19:07:05 -0800 Curious what the status of this patch is? Should this be r-'d now that we know that this patch will at least cause one layout test to fail? 175394 28 41379 eric 2009-12-28 22:16:01 -0800 Comment on attachment 41379 Paint RenderPath only when necessary Marking r- since this causes a test to fail. 175395 29 eric 2009-12-28 22:16:39 -0800 Also the second OOPS in the ChangeLog would have caused the commit to fail. ChangeLogs should mention tests affected by a change. 175561 30 paroga 2009-12-29 08:26:04 -0800 (In reply to comment #28) > (From update of attachment 41379 [details]) > Marking r- since this causes a test to fail. The problem is that an other area of Webkit does something wrong. This patch is ok, but i depends on an "not created bug". Can we create a new bug, make the patch depend on in r+ it again? Who is responsible for the code tested with filterRegions.svg? Maybe he can make a change there to get this patch in? 175589 31 eric 2009-12-29 09:19:56 -0800 Yes, if we're waiting to commit this because a problem exists that we've never actually created a bug for, then definitely we should create a bug. :) Getting an r+ is relatively easy. Especially since this patch already had one. But if we can't land it due to another bug, we should file the other bug and mark this one blocked. 175833 32 krit 2009-12-29 23:52:58 -0800 Can we move this fix to SVGRenderBase::prepareToRenderSVGContent?? This would stop rendering earlier and can improve the rendering time on other parts like images too. 175858 33 paroga 2009-12-30 02:19:59 -0800 (In reply to comment #31) > Yes, if we're waiting to commit this because a problem exists that we've never > actually created a bug for, then definitely we should create a bug. :) Can we r+ and cq+ it again since bug 32815 is fixed now? (In reply to comment #32) > Can we move this fix to SVGRenderBase::prepareToRenderSVGContent?? > > This would stop rendering earlier and can improve the rendering time on other > parts like images too. I think we should commit this change and maybe another too. This patch is not in the SVG part so it might improve also some other parts. 175862 34 krit 2009-12-30 03:12:06 -0800 > (In reply to comment #32) > (In reply to comment #32) > > Can we move this fix to SVGRenderBase::prepareToRenderSVGContent?? > > > > This would stop rendering earlier and can improve the rendering time on other > > parts like images too. > I think we should commit this change and maybe another too. This patch is not > in the SVG part so it might improve also some other parts. RenderPath is SVG only. 175872 35 paroga 2009-12-30 03:57:10 -0800 (In reply to comment #34) > RenderPath is SVG only. I meant that maybe another componet will use RenderPath in the future, but thats not a real reason. Moving it to prepareToRenderSVGContent won't prevent _unnecessary_ paintInfo.context->save(), concatCTM(...), restore() and so on. It should be done as early as possible! Iterating over _all_ items, even if only one pixel changed, is already not very performant. (checking which items need a repaint might be even more cpu consuming) 175874 36 krit 2009-12-30 04:22:13 -0800 (In reply to comment #35) > (In reply to comment #34) > > RenderPath is SVG only. > I meant that maybe another componet will use RenderPath in the future, but > thats not a real reason. > > Moving it to prepareToRenderSVGContent won't prevent _unnecessary_ > paintInfo.context->save(), concatCTM(...), restore() and so on. > It should be done as early as possible! Iterating over _all_ items, even if > only one pixel changed, is already not very performant. (checking which items > need a repaint might be even more cpu consuming) Can you please upload a new patch with a link to the bugreport Changelog and the title of the bug? There shouldn't be Oops'es in a Changelog. Please write why it is not possible to test the speed improvement and I think we can commit it. 175888 37 45658 paroga 2009-12-30 05:47:41 -0800 Created attachment 45658 Paint RenderPath only when necessary 175890 38 webkit.review.bot 2009-12-30 05:49:17 -0800 style-queue ran check-webkit-style on attachment 45658 without any errors. 175922 39 45658 krit 2009-12-30 08:01:52 -0800 Comment on attachment 45658 Paint RenderPath only when necessary LGTM. Thanks for the patch. 175923 40 45658 commit-queue 2009-12-30 08:10:48 -0800 Comment on attachment 45658 Paint RenderPath only when necessary Rejecting patch 45658 from commit-queue. Failed to run "['WebKitTools/Scripts/run-webkit-tests', '--no-launch-safari', '--exit-after-n-failures=1', '--quiet']" exit_code: 1 Running build-dumprendertree Running tests from /Users/eseidel/Projects/CommitQueueSVN/LayoutTests Testing 11854 test cases. svg/custom/marker-overflow-clip.svg -> failed Exiting early after 1 failures. 10306 tests run. 313.90s total testing time 10305 test cases (99%) succeeded 1 test case (<1%) had incorrect layout 7 test cases (<1%) had stderr output Full output: http://webkit-commit-queue.appspot.com/results/152351 175931 41 krit 2009-12-30 08:56:48 -0800 (In reply to comment #40) > (From update of attachment 45658 [details]) > Rejecting patch 45658 from commit-queue. > > Failed to run "['WebKitTools/Scripts/run-webkit-tests', '--no-launch-safari', > '--exit-after-n-failures=1', '--quiet']" exit_code: 1 > Running build-dumprendertree > Running tests from /Users/eseidel/Projects/CommitQueueSVN/LayoutTests > Testing 11854 test cases. > svg/custom/marker-overflow-clip.svg -> failed > > Exiting early after 1 failures. 10306 tests run. > 313.90s total testing time > > 10305 test cases (99%) succeeded > 1 test case (<1%) had incorrect layout > 7 test cases (<1%) had stderr output > > Full output: http://webkit-commit-queue.appspot.com/results/152351 Ahhh sorry. There is another bug about markers and boundingBoxes: bug 33012 . Without knowing the actual result with your patch, I guess that this fixes the problem. 176502 42 45658 eric 2010-01-03 18:46:18 -0800 Comment on attachment 45658 Paint RenderPath only when necessary Marking r- since this seems to cause a test to fail? Correct me if I'm wrong Dirk. 176546 43 krit 2010-01-03 23:58:23 -0800 (In reply to comment #42) > (From update of attachment 45658 [details]) > Marking r- since this seems to cause a test to fail? Correct me if I'm wrong > Dirk. No, the patch is ok. It just depends on anotherone. I'll commot-queue+ the patch again, once Niko fixed the marker bug. 177810 44 45658 krit 2010-01-06 15:33:07 -0800 Comment on attachment 45658 Paint RenderPath only when necessary Markers and repaintRects are fixed. Lets try it again. 177980 45 45658 commit-queue 2010-01-06 21:30:37 -0800 Comment on attachment 45658 Paint RenderPath only when necessary Clearing flags on attachment: 45658 Committed r52900: <http://trac.webkit.org/changeset/52900> 177981 46 commit-queue 2010-01-06 21:30:50 -0800 All reviewed patches have been landed. Closing bug. 177983 47 45658 eric 2010-01-06 21:34:53 -0800 Comment on attachment 45658 Paint RenderPath only when necessary This type of fix would be more interesting to apply to the hit-testing code path. But this fix should be fine. Thanks for the patch Patrick and the review Dirk! 178235 48 dglazkov 2010-01-07 09:53:32 -0800 This change broke a bunch of SVG tests, all pixel results. For illustration, visit this at ToT: http://trac.webkit.org/export/52929/trunk/LayoutTests/fast/backgrounds/size/contain-and-cover.html I think we should roll this change out. 178242 49 dglazkov 2010-01-07 10:09:13 -0800 Reverted as http://trac.webkit.org/changeset/52930. 178251 50 krit 2010-01-07 10:23:32 -0800 (In reply to comment #48) > This change broke a bunch of SVG tests, all pixel results. For illustration, > visit this at ToT: > http://trac.webkit.org/export/52929/trunk/LayoutTests/fast/backgrounds/size/contain-and-cover.html > > I think we should roll this change out. A problem of transparent and half transparent parts of some objects? 179179 51 46243 oliver 2010-01-10 15:18:39 -0800 Created attachment 46243 Bad DOM performance in large SVG files 179180 52 46244 oliver 2010-01-10 15:19:10 -0800 Created attachment 46244 Bad DOM performance in large SVG files 179199 53 46243 zimmermann 2010-01-10 17:02:31 -0800 Comment on attachment 46243 Bad DOM performance in large SVG files LGTM, I assume there are no pixel test regressions, r+. 179200 54 46244 zimmermann 2010-01-10 17:03:27 -0800 Comment on attachment 46244 Bad DOM performance in large SVG files Interessting that CGPathIsEmpty is that slow... I assume you see a good benefit, so r=me. 179213 55 oliver 2010-01-10 19:05:16 -0800 First patch landed in r53057. I'm actually slightly uneasy about landing the isempty caching so i'm going to see if there's something that can be done to get a bigger win. That said i think that this patch should resolve all of the perf issues in the main testcase attached to this patch. 179260 56 dglazkov 2010-01-10 21:03:50 -0800 Sorry guys, still no luck on pixel tests. Running "run-webkit-tests --pixel-tests svg" produces: Enabling pixel tests with a tolerance of 0.1% WARNING: Temporarily changing the main display color profile: The colors on your screen will change for the duration of the testing. This allows the pixel tests to have consistent color values across all machines. Testing 945 test cases. svg/W3C-SVG-1.1 ....................................................................... svg/W3C-SVG-1.1/coords-units-01-b.svg -> pixel test failed ......... svg/W3C-SVG-1.1/filters-composite-02-b.svg -> pixel test failed .......................................... svg/W3C-SVG-1.1/masking-intro-01-f.svg -> pixel test failed . svg/W3C-SVG-1.1/masking-mask-01-b.svg -> pixel test failed ...................................... svg/W3C-SVG-1.1/pservers-grad-03-b.svg -> pixel test failed ... svg/W3C-SVG-1.1/pservers-grad-06-b.svg -> pixel test failed ....... svg/W3C-SVG-1.1/pservers-grad-13-b.svg -> pixel hash failed (but pixel test still passes) .... svg/W3C-SVG-1.1/pservers-grad-17-b.svg -> pixel hash failed (but pixel test still passes) ... svg/W3C-SVG-1.1/pservers-pattern-01-b.svg -> pixel test failed ....................................................................... svg/W3C-SVG-1.1/text-deco-01-b.svg -> pixel hash failed (but pixel test still passes) .... svg/W3C-SVG-1.1/text-path-01-b.svg -> pixel hash failed (but pixel test still passes) ... svg/W3C-SVG-1.1/text-text-03-b.svg -> pixel hash failed (but pixel test still passes) .... svg/W3C-SVG-1.1/text-text-07-t.svg -> pixel hash failed (but pixel test still passes) ........ svg/animations . svg/batik/filters .. svg/batik/masking . svg/batik/masking/maskRegions.svg -> pixel test failed svg/batik/paints .. svg/batik/paints/patternPreserveAspectRatioA.svg -> pixel test failed . svg/batik/paints/patternRegionA.svg -> pixel test failed . svg/batik/paints/patternRegions.svg -> pixel test failed svg/batik/text ..... svg/batik/text/textDecoration2.svg -> pixel hash failed (but pixel test still passes) .. svg/batik/text/textEffect3.svg -> pixel hash failed (but pixel test still passes) .. svg/batik/text/textFeatures.svg -> pixel hash failed (but pixel test still passes) . svg/batik/text/textGlyphOrientationHorizontal.svg -> pixel hash failed (but pixel test still passes) ....... svg/batik/text/textOnPathSpaces.svg -> pixel hash failed (but pixel test still passes) ..... svg/batik/text/textProperties.svg -> pixel hash failed (but pixel test still passes) . svg/batik/text/textStyles.svg -> pixel hash failed (but pixel test still passes) . svg/batik/text/verticalText.svg -> pixel hash failed (but pixel test still passes) . svg/batik/text/verticalTextOnPath.svg -> pixel hash failed (but pixel test still passes) . svg/carto.net .......... svg/css . svg/css/arrow-with-shadow.svg -> pixel hash failed (but pixel test still passes) . svg/css/circle-in-mask-with-shadow.svg -> pixel test failed .. svg/css/composite-shadow-example.html -> pixel hash failed (but pixel test still passes) . svg/css/composite-shadow-with-opacity.html -> pixel hash failed (but pixel test still passes) .... svg/css/group-with-shadow.svg -> pixel hash failed (but pixel test still passes) . svg/css/mask-with-shadow.svg -> pixel test failed . svg/css/path-with-shadow.svg -> pixel hash failed (but pixel test still passes) . svg/css/stars-with-shadow.html -> pixel hash failed (but pixel test still passes) svg/custom ....................................... svg/custom/deep-dynamic-updates.svg -> pixel test failed ................................................... svg/custom/grayscale-gradient-mask.svg -> pixel test failed ............ svg/custom/image-with-transform-clip-filter.svg -> pixel hash failed (but pixel test still passes) ................. svg/custom/js-late-mask-and-object-creation.svg -> pixel test failed . svg/custom/js-late-mask-creation.svg -> pixel test failed . svg/custom/js-late-pattern-and-object-creation.svg -> pixel test failed . svg/custom/js-late-pattern-creation.svg -> pixel test failed ............ svg/custom/js-update-pattern-child.svg -> pixel test failed . svg/custom/js-update-pattern.svg -> pixel test failed ....................... svg/custom/marker-changes.svg -> pixel hash failed (but pixel test still passes) . svg/custom/marker-child-changes.svg -> pixel hash failed (but pixel test still passes) ...... svg/custom/marker-strokeWidth-changes.svg -> pixel hash failed (but pixel test still passes) .. svg/custom/mask-changes.svg -> pixel test failed . svg/custom/mask-child-changes.svg -> pixel test failed . svg/custom/mask-excessive-malloc.svg -> pixel test failed . svg/custom/mask-inside-defs.svg -> pixel test failed . svg/custom/mask-on-multiple-objects.svg -> pixel test failed . svg/custom/mask-with-all-units.svg -> pixel test failed .............. svg/custom/pattern-cycle-detection.svg -> pixel test failed . svg/custom/pattern-deep-referencing.svg -> pixel test failed . svg/custom/pattern-in-defs.svg -> pixel test failed . svg/custom/pattern-rotate.svg -> pixel test failed . svg/custom/pattern-with-transformation.svg -> pixel test failed . svg/custom/pattern-y-offset.svg -> pixel test failed ............... svg/custom/radial-gradient-with-outstanding-focalPoint.svg -> pixel test failed .... svg/custom/resource-invalidate-on-target-update.svg -> pixel test failed ................ svg/custom/stroked-pattern.svg -> pixel test failed ........................................................................... svg/custom/use-on-symbol-inside-pattern.svg -> pixel test failed ......................... svg/custom/visibility-override-mask.svg -> pixel test failed .... svg/dom ......................... svg/dom/SVGRectElement . svg/dom/SVGScriptElement .... svg/dom/SVGStyleElement . svg/dynamic-updates ................................................................................................................................ svg/filters .......... svg/filters/filter-clip.svg -> pixel test failed ........ svg/hittest .. svg/hixie/cascade .. svg/hixie/data-types .. svg/hixie/dynamic ...... svg/hixie/error .................. svg/hixie/links ... svg/hixie/mixed ......... svg/hixie/painting . svg/hixie/perf ..... svg/hixie/processing-model ... svg/hixie/rendering-model ..... svg/hixie/shapes/path . svg/hixie/text ..... svg/hixie/transform . svg/hixie/use ... svg/hixie/viewbox .... svg/hixie/viewbox/preserveAspectRatio .. svg/text .... svg/text/selection-background-color.xhtml -> pixel hash failed (but pixel test still passes) .... svg/text/text-align-04-b.svg -> pixel hash failed (but pixel test still passes) .... svg/text/text-deco-01-b.svg -> pixel hash failed (but pixel test still passes) ....... svg/text/text-path-01-b.svg -> pixel hash failed (but pixel test still passes) .... svg/text/text-text-01-b.svg -> pixel hash failed (but pixel test still passes) . svg/text/text-text-03-b.svg -> pixel hash failed (but pixel test still passes) . svg/text/text-text-04-t.svg -> pixel hash failed (but pixel test still passes) ... svg/text/text-text-07-t.svg -> pixel hash failed (but pixel test still passes) . svg/text/text-text-08-b.svg -> pixel hash failed (but pixel test still passes) ....... svg/transforms .. svg/transforms/text-with-pattern-inside-transformed-html.xhtml -> pixel test failed . svg/transforms/text-with-pattern-with-svg-transform.svg -> pixel test failed svg/webarchive ... 72.68s total testing time 904 test cases (95%) succeeded 41 test cases (4%) had incorrect layout 1 test case (<1%) had stderr output 179263 57 dglazkov 2010-01-10 21:06:03 -0800 You can see the failures clearly here: http://build.chromium.org/buildbot/waterfall.fyi/waterfall?branch=&builder=Webkit+(webkit.org)&builder=Webkit+Linux+(webkit.org)&builder=Webkit+Mac+(webkit.org)&builder=XP+Perf+(webkit.org)&builder=Linux+Perf+(webkit.org)&builder=Webkit+(purify+webkit.org) 179281 58 dglazkov 2010-01-10 21:22:09 -0800 Just to make sure, with the patch rolled out locally, same command produces: 944 test cases (99%) succeeded 1 test case (<1%) had incorrect layout 1 test case (<1%) had stderr output 179311 59 46243 zimmermann 2010-01-11 02:42:42 -0800 Comment on attachment 46243 Bad DOM performance in large SVG files Clearing review flag, this patch needs to be revisited. Anyone who works on this has to be aware that regressions can only be spotted when running _PIXEL TESTS_. 179312 60 46244 zimmermann 2010-01-11 02:43:00 -0800 Comment on attachment 46244 Bad DOM performance in large SVG files Taking back review, as discussed with Oliver. 181186 61 46667 oliver 2010-01-15 03:49:24 -0800 Created attachment 46667 Patch 181189 62 46668 oliver 2010-01-15 03:51:56 -0800 Created attachment 46668 Patch 181198 63 46668 zimmermann 2010-01-15 04:19:41 -0800 Comment on attachment 46668 Patch Okay, let's hope it fixes it this time... Can you file a follow-up bug regarding PaintInfo fixes is renderSubtreeToImage? 181201 64 oliver 2010-01-15 04:25:40 -0800 Committed r53331: <http://trac.webkit.org/changeset/53331> 181250 65 dglazkov 2010-01-15 09:21:33 -0800 Darnit! One test still broke: http://trac.webkit.org/browser/trunk/LayoutTests/fast/borders/svg-as-border-image-2.html Can you take a look and see what's happening there? 181255 66 dglazkov 2010-01-15 09:48:27 -0800 Rolled out Oliver's patch in http://trac.webkit.org/changeset/53334, per IRC discussion w/Nikolas. But you're really, really close now! Don't give up! 181326 67 46701 oliver 2010-01-15 12:48:23 -0800 Created attachment 46701 Patch 181337 68 oliver 2010-01-15 12:54:47 -0800 Committed r53343: <http://trac.webkit.org/changeset/53343> 181422 69 46715 oliver 2010-01-15 16:04:23 -0800 Created attachment 46715 Patch 181423 70 oliver 2010-01-15 16:05:44 -0800 Just opening for a improvement to culling 181425 71 46715 sam 2010-01-15 16:14:56 -0800 Comment on attachment 46715 Patch Please add more complete explanation to the ChangeLog. r=me 181426 72 oliver 2010-01-15 16:16:35 -0800 Committed r53349: <http://trac.webkit.org/changeset/53349> 41379 2009-10-18 09:10:27 -0700 2009-12-30 05:47:41 -0800 Paint RenderPath only when necessary bug-30055-20091018155000.patch text/plain 1375 paroga SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA0OTc0OCkKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTQgQEAKKzIwMDktMTAtMTggIFBhdHJpY2sgR2Fuc3RlcmVyICA8cGFyb2dhQHBh cm9nYS5jb20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAg ICAgQ2hlY2sgdGhlIGRpcnR5IGFyZWEgd2hlbiBwYWludGluZyBhIHBhdGguCisKKyAgICAgICAg Tm8gbmV3IHRlc3RzLiAoT09QUyEpCisKKyAgICAgICAgKiByZW5kZXJpbmcvUmVuZGVyUGF0aC5j cHA6CisgICAgICAgIChXZWJDb3JlOjpSZW5kZXJQYXRoOjpwYWludCk6CisKIDIwMDktMTAtMTcg IE5pa29sYXMgWmltbWVybWFubiAgPG56aW1tZXJtYW5uQHJpbS5jb20+CiAKICAgICAgICAgUmV2 aWV3ZWQgYnkgR2VvcmdlIFN0YWlrb3MuCkluZGV4OiBXZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJQ YXRoLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09Ci0tLSBXZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJQYXRoLmNwcAko cmV2aXNpb24gNDk3MzkpCisrKyBXZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJQYXRoLmNwcAkod29y a2luZyBjb3B5KQpAQCAtMTgwLDEzICsxODAsMTYgQEAgdm9pZCBSZW5kZXJQYXRoOjpwYWludChQ YWludEluZm8mIHBhaW50SQogewogICAgIGlmIChwYWludEluZm8uY29udGV4dC0+cGFpbnRpbmdE aXNhYmxlZCgpIHx8IHN0eWxlKCktPnZpc2liaWxpdHkoKSA9PSBISURERU4gfHwgbV9wYXRoLmlz RW1wdHkoKSkKICAgICAgICAgcmV0dXJuOwotICAgICAgICAgICAgCisKKyAgICBGbG9hdFJlY3Qg Ym91bmRpbmdCb3ggPSByZXBhaW50UmVjdEluTG9jYWxDb29yZGluYXRlcygpOworICAgIGlmICgh Ym91bmRpbmdCb3guaW50ZXJzZWN0cyhwYWludEluZm8ucmVjdCkpCisgICAgICAgIHJldHVybjsK KwogICAgIHBhaW50SW5mby5jb250ZXh0LT5zYXZlKCk7CiAgICAgcGFpbnRJbmZvLmNvbnRleHQt PmNvbmNhdENUTShsb2NhbFRvUGFyZW50VHJhbnNmb3JtKCkpOwogCiAgICAgU1ZHUmVzb3VyY2VG aWx0ZXIqIGZpbHRlciA9IDA7CiAKLSAgICBGbG9hdFJlY3QgYm91bmRpbmdCb3ggPSByZXBhaW50 UmVjdEluTG9jYWxDb29yZGluYXRlcygpOwogICAgIGlmIChwYWludEluZm8ucGhhc2UgPT0gUGFp bnRQaGFzZUZvcmVncm91bmQpIHsKICAgICAgICAgUGFpbnRJbmZvIHNhdmVkSW5mbyhwYWludElu Zm8pOwogCg== 45658 2009-12-30 05:47:41 -0800 2010-01-10 15:18:33 -0800 Paint RenderPath only when necessary bug-30055-20091230.patch text/plain 1547 paroga SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA1MjYwNikKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTUgQEAKKzIwMDktMTAtMTggUGF0cmljayBHYW5zdGVyZXIgPHBhcm9nYUBwYXJv Z2EuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAg IEltcHJvdmUgRE9NIHBlcmZvcm1hbmNlIGluIGxhcmdlIFNWRyBmaWxlcy4KKyAgICAgICAgQ2hl Y2sgdGhlIGRpcnR5IGFyZWEgYmVmb3JlIHBhaW50aW5nIGEgcGF0aCwgd2hpY2ggcHJldmVudHMg dW5uZWNlc3NhcnkKKyAgICAgICAgY2xpcHBpbmcgYW5kIGNoYW5nZXMgdG8gdGhlIEdyYXBoaWNz Q29udGV4dC4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lk PTMwMDU1CisKKyAgICAgICAgKiByZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHA6CisgICAgICAgIChX ZWJDb3JlOjpSZW5kZXJQYXRoOjpwYWludCk6CisKIDIwMDktMTItMjggIEd1c3Rhdm8gTm9yb25o YSBTaWx2YSAgPGd1c3Rhdm8ubm9yb25oYUBjb2xsYWJvcmEuY28udWs+CiAKICAgICAgICAgUnVi YmVyLXN0YW1wZWQgYnkgWGFuIExvcGV6LgpJbmRleDogV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVy UGF0aC5jcHAKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PQotLS0gV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAJ KHJldmlzaW9uIDUyNTk1KQorKysgV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAJKHdv cmtpbmcgY29weSkKQEAgLTE4MCwxMyArMTgwLDE2IEBAIHZvaWQgUmVuZGVyUGF0aDo6cGFpbnQo UGFpbnRJbmZvJiBwYWludEkKIHsKICAgICBpZiAocGFpbnRJbmZvLmNvbnRleHQtPnBhaW50aW5n RGlzYWJsZWQoKSB8fCBzdHlsZSgpLT52aXNpYmlsaXR5KCkgPT0gSElEREVOIHx8IG1fcGF0aC5p c0VtcHR5KCkpCiAgICAgICAgIHJldHVybjsKLSAgICAgICAgICAgIAorCisgICAgRmxvYXRSZWN0 IGJvdW5kaW5nQm94ID0gcmVwYWludFJlY3RJbkxvY2FsQ29vcmRpbmF0ZXMoKTsKKyAgICBpZiAo IWJvdW5kaW5nQm94LmludGVyc2VjdHMocGFpbnRJbmZvLnJlY3QpKQorICAgICAgICByZXR1cm47 CisKICAgICBwYWludEluZm8uY29udGV4dC0+c2F2ZSgpOwogICAgIHBhaW50SW5mby5jb250ZXh0 LT5jb25jYXRDVE0obG9jYWxUb1BhcmVudFRyYW5zZm9ybSgpKTsKIAogICAgIFNWR1Jlc291cmNl RmlsdGVyKiBmaWx0ZXIgPSAwOwogCi0gICAgRmxvYXRSZWN0IGJvdW5kaW5nQm94ID0gcmVwYWlu dFJlY3RJbkxvY2FsQ29vcmRpbmF0ZXMoKTsKICAgICBpZiAocGFpbnRJbmZvLnBoYXNlID09IFBh aW50UGhhc2VGb3JlZ3JvdW5kKSB7CiAgICAgICAgIFBhaW50SW5mbyBzYXZlZEluZm8ocGFpbnRJ bmZvKTsKIAo= 46243 2010-01-10 15:18:39 -0800 2010-01-15 03:49:07 -0800 Bad DOM performance in large SVG files bug-30055-20100110151838.patch text/plain 3885 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvQ2hhbmdlTG9nIGIvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXgg N2E2MmNkMS4uNTUyZmQxYSAxMDA2NDQKLS0tIGEvV2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvV2Vi Q29yZS9DaGFuZ2VMb2cKQEAgLTEsMyArMSwxNiBAQAorMjAxMC0wMS0xMCAgT2xpdmVyIEh1bnQg IDxvbGl2ZXJAYXBwbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEp LgorCisgICAgICAgIEJhZCBET00gcGVyZm9ybWFuY2UgaW4gbGFyZ2UgU1ZHIGZpbGVzCisgICAg ICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMDA1NQorCisgICAg ICAgIEN1bGwgUmVuZGVyUGF0aHMgYmVmb3JlIHBhc3Npbmcgb24gdG8gdGhlIHVuZGVybHlpbmcg Z3JhcGhpY3Mgc3lzdGVtLgorCisgICAgICAgICogcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwOgor ICAgICAgICAoV2ViQ29yZTo6UmVuZGVyUGF0aDo6cGFpbnQpOgorICAgICAgICAoV2ViQ29yZTo6 UmVuZGVyUGF0aDo6bm9kZUF0RmxvYXRQb2ludCk6CisKIDIwMTAtMDEtMDggIEx1aXogQWdvc3Rp bmkgIDxsdWl6LmFnb3N0aW5pQG9wZW5ib3NzYS5vcmc+CiAKICAgICAgICAgUmV2aWV3ZWQgYnkg S2VubmV0aCBSb2hkZSBDaHJpc3RpYW5zZW4uCmRpZmYgLS1naXQgYS9XZWJDb3JlL3JlbmRlcmlu Zy9SZW5kZXJQYXRoLmNwcCBiL1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwCmluZGV4 IDMwY2JlZGYuLmEyYWMzMzEgMTAwNjQ0Ci0tLSBhL1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlclBh dGguY3BwCisrKyBiL1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwCkBAIC0yMTQsMzIg KzIxNCwzNiBAQCB2b2lkIFJlbmRlclBhdGg6OnBhaW50KFBhaW50SW5mbyYgcGFpbnRJbmZvLCBp bnQsIGludCkKIHsKICAgICBpZiAocGFpbnRJbmZvLmNvbnRleHQtPnBhaW50aW5nRGlzYWJsZWQo KSB8fCBzdHlsZSgpLT52aXNpYmlsaXR5KCkgPT0gSElEREVOIHx8IG1fcGF0aC5pc0VtcHR5KCkp CiAgICAgICAgIHJldHVybjsKLSAgICAgICAgICAgIAotICAgIHBhaW50SW5mby5jb250ZXh0LT5z YXZlKCk7Ci0gICAgcGFpbnRJbmZvLmNvbnRleHQtPmNvbmNhdENUTShsb2NhbFRvUGFyZW50VHJh bnNmb3JtKCkpOworICAgIFBhaW50SW5mbyBjaGlsZFBhaW50SW5mbyhwYWludEluZm8pOworICAg IGNoaWxkUGFpbnRJbmZvLmNvbnRleHQtPnNhdmUoKTsKKyAgICBhcHBseVRyYW5zZm9ybVRvUGFp bnRJbmZvKGNoaWxkUGFpbnRJbmZvLCBtX2xvY2FsVHJhbnNmb3JtKTsKKyAgICBGbG9hdFJlY3Qg Ym91bmRpbmdCb3ggPSByZXBhaW50UmVjdEluTG9jYWxDb29yZGluYXRlcygpOworICAgIGlmICgh Ym91bmRpbmdCb3guaW50ZXJzZWN0cyhjaGlsZFBhaW50SW5mby5yZWN0KSkgeworICAgICAgICBj aGlsZFBhaW50SW5mby5jb250ZXh0LT5yZXN0b3JlKCk7CisgICAgICAgIHJldHVybjsKKyAgICB9 CiAKICAgICBTVkdSZXNvdXJjZUZpbHRlciogZmlsdGVyID0gMDsKIAotICAgIEZsb2F0UmVjdCBi b3VuZGluZ0JveCA9IHJlcGFpbnRSZWN0SW5Mb2NhbENvb3JkaW5hdGVzKCk7Ci0gICAgaWYgKHBh aW50SW5mby5waGFzZSA9PSBQYWludFBoYXNlRm9yZWdyb3VuZCkgewotICAgICAgICBQYWludElu Zm8gc2F2ZWRJbmZvKHBhaW50SW5mbyk7CisgICAgaWYgKGNoaWxkUGFpbnRJbmZvLnBoYXNlID09 IFBhaW50UGhhc2VGb3JlZ3JvdW5kKSB7CisgICAgICAgIFBhaW50SW5mbyBzYXZlZEluZm8oY2hp bGRQYWludEluZm8pOwogCi0gICAgICAgIGlmIChwcmVwYXJlVG9SZW5kZXJTVkdDb250ZW50KHRo aXMsIHBhaW50SW5mbywgYm91bmRpbmdCb3gsIGZpbHRlcikpIHsKKyAgICAgICAgaWYgKHByZXBh cmVUb1JlbmRlclNWR0NvbnRlbnQodGhpcywgY2hpbGRQYWludEluZm8sIGJvdW5kaW5nQm94LCBm aWx0ZXIpKSB7CiAgICAgICAgICAgICBpZiAoc3R5bGUoKS0+c3ZnU3R5bGUoKS0+c2hhcGVSZW5k ZXJpbmcoKSA9PSBTUl9DUklTUEVER0VTKQotICAgICAgICAgICAgICAgIHBhaW50SW5mby5jb250 ZXh0LT5zZXRTaG91bGRBbnRpYWxpYXMoZmFsc2UpOwotICAgICAgICAgICAgZmlsbEFuZFN0cm9r ZVBhdGgobV9wYXRoLCBwYWludEluZm8uY29udGV4dCwgc3R5bGUoKSwgdGhpcyk7CisgICAgICAg ICAgICAgICAgY2hpbGRQYWludEluZm8uY29udGV4dC0+c2V0U2hvdWxkQW50aWFsaWFzKGZhbHNl KTsKKyAgICAgICAgICAgIGZpbGxBbmRTdHJva2VQYXRoKG1fcGF0aCwgY2hpbGRQYWludEluZm8u Y29udGV4dCwgc3R5bGUoKSwgdGhpcyk7CiAKICAgICAgICAgICAgIGlmIChzdGF0aWNfY2FzdDxT VkdTdHlsZWRFbGVtZW50Kj4obm9kZSgpKS0+c3VwcG9ydHNNYXJrZXJzKCkpCi0gICAgICAgICAg ICAgICAgbV9tYXJrZXJMYXlvdXRJbmZvLmRyYXdNYXJrZXJzKHBhaW50SW5mbyk7CisgICAgICAg ICAgICAgICAgbV9tYXJrZXJMYXlvdXRJbmZvLmRyYXdNYXJrZXJzKGNoaWxkUGFpbnRJbmZvKTsK ICAgICAgICAgfQotICAgICAgICBmaW5pc2hSZW5kZXJTVkdDb250ZW50KHRoaXMsIHBhaW50SW5m bywgZmlsdGVyLCBzYXZlZEluZm8uY29udGV4dCk7CisgICAgICAgIGZpbmlzaFJlbmRlclNWR0Nv bnRlbnQodGhpcywgY2hpbGRQYWludEluZm8sIGZpbHRlciwgc2F2ZWRJbmZvLmNvbnRleHQpOwog ICAgIH0KIAotICAgIGlmICgocGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VPdXRsaW5lIHx8 IHBhaW50SW5mby5waGFzZSA9PSBQYWludFBoYXNlU2VsZk91dGxpbmUpICYmIHN0eWxlKCktPm91 dGxpbmVXaWR0aCgpKQotICAgICAgICBwYWludE91dGxpbmUocGFpbnRJbmZvLmNvbnRleHQsIHN0 YXRpY19jYXN0PGludD4oYm91bmRpbmdCb3gueCgpKSwgc3RhdGljX2Nhc3Q8aW50Pihib3VuZGlu Z0JveC55KCkpLAorICAgIGlmICgoY2hpbGRQYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZU91 dGxpbmUgfHwgY2hpbGRQYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZVNlbGZPdXRsaW5lKSAm JiBzdHlsZSgpLT5vdXRsaW5lV2lkdGgoKSkKKyAgICAgICAgcGFpbnRPdXRsaW5lKGNoaWxkUGFp bnRJbmZvLmNvbnRleHQsIHN0YXRpY19jYXN0PGludD4oYm91bmRpbmdCb3gueCgpKSwgc3RhdGlj X2Nhc3Q8aW50Pihib3VuZGluZ0JveC55KCkpLAogICAgICAgICAgICAgc3RhdGljX2Nhc3Q8aW50 Pihib3VuZGluZ0JveC53aWR0aCgpKSwgc3RhdGljX2Nhc3Q8aW50Pihib3VuZGluZ0JveC5oZWln aHQoKSksIHN0eWxlKCkpOwogICAgIAotICAgIHBhaW50SW5mby5jb250ZXh0LT5yZXN0b3JlKCk7 CisgICAgY2hpbGRQYWludEluZm8uY29udGV4dC0+cmVzdG9yZSgpOwogfQogCiAvLyBUaGlzIG1l dGhvZCBpcyBjYWxsZWQgZnJvbSBpbnNpZGUgcGFpbnRPdXRsaW5lKCkgc2luY2Ugd2UgY2FsbCBw YWludE91dGxpbmUoKQpAQCAtMjU3LDcgKzI2MSw3IEBAIGJvb2wgUmVuZGVyUGF0aDo6bm9kZUF0 RmxvYXRQb2ludChjb25zdCBIaXRUZXN0UmVxdWVzdCYsIEhpdFRlc3RSZXN1bHQmIHJlc3VsdCwK ICAgICBpZiAoaGl0VGVzdEFjdGlvbiAhPSBIaXRUZXN0Rm9yZWdyb3VuZCkKICAgICAgICAgcmV0 dXJuIGZhbHNlOwogCi0gICAgRmxvYXRQb2ludCBsb2NhbFBvaW50ID0gbG9jYWxUb1BhcmVudFRy YW5zZm9ybSgpLmludmVyc2UoKS5tYXBQb2ludChwb2ludEluUGFyZW50KTsKKyAgICBGbG9hdFBv aW50IGxvY2FsUG9pbnQgPSBtX2xvY2FsVHJhbnNmb3JtLmludmVyc2UoKS5tYXBQb2ludChwb2lu dEluUGFyZW50KTsKIAogICAgIFBvaW50ZXJFdmVudHNIaXRSdWxlcyBoaXRSdWxlcyhQb2ludGVy RXZlbnRzSGl0UnVsZXM6OlNWR19QQVRIX0hJVFRFU1RJTkcsIHN0eWxlKCktPnBvaW50ZXJFdmVu dHMoKSk7CiAK 46244 2010-01-10 15:19:10 -0800 2010-01-15 03:49:17 -0800 Bad DOM performance in large SVG files bug-30055-20100110151909.patch text/plain 5658 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvQ2hhbmdlTG9nIGIvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXgg NTUyZmQxYS4uYzVhYzNlMiAxMDA2NDQKLS0tIGEvV2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvV2Vi Q29yZS9DaGFuZ2VMb2cKQEAgLTUsNiArNSwzNCBAQAogICAgICAgICBCYWQgRE9NIHBlcmZvcm1h bmNlIGluIGxhcmdlIFNWRyBmaWxlcwogICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9z aG93X2J1Zy5jZ2k/aWQ9MzAwNTUKIAorICAgICAgICBDR1BhdGhJc0VtcHR5IGlzIHZlcnkgZXhw ZW5zaXZlLCBzbyB3ZSBuZWVkIHRvIGNhY2hlIHRoZSByZXN1bHQuCisKKyAgICAgICAgKiBwbGF0 Zm9ybS9ncmFwaGljcy9QYXRoLmg6CisgICAgICAgICogcGxhdGZvcm0vZ3JhcGhpY3MvY2cvUGF0 aENHLmNwcDoKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6OlBhdGgpOgorICAgICAgICAoV2ViQ29y ZTo6UGF0aDo6b3BlcmF0b3I9KToKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6OnRyYW5zbGF0ZSk6 CisgICAgICAgIChXZWJDb3JlOjpQYXRoOjpzdHJva2VCb3VuZGluZ1JlY3QpOgorICAgICAgICAo V2ViQ29yZTo6UGF0aDo6bW92ZVRvKToKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6OmFkZExpbmVU byk6CisgICAgICAgIChXZWJDb3JlOjpQYXRoOjphZGRRdWFkQ3VydmVUbyk6CisgICAgICAgIChX ZWJDb3JlOjpQYXRoOjphZGRCZXppZXJDdXJ2ZVRvKToKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6 OmFkZEFyY1RvKToKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6OmNsb3NlU3VicGF0aCk6CisgICAg ICAgIChXZWJDb3JlOjpQYXRoOjphZGRBcmMpOgorICAgICAgICAoV2ViQ29yZTo6UGF0aDo6YWRk UmVjdCk6CisgICAgICAgIChXZWJDb3JlOjpQYXRoOjphZGRFbGxpcHNlKToKKyAgICAgICAgKFdl YkNvcmU6OlBhdGg6OmNsZWFyKToKKyAgICAgICAgKFdlYkNvcmU6OlBhdGg6OmlzRW1wdHkpOgor ICAgICAgICAoV2ViQ29yZTo6UGF0aDo6dHJhbnNmb3JtKToKKworMjAxMC0wMS0xMCAgT2xpdmVy IEh1bnQgIDxvbGl2ZXJAYXBwbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAo T09QUyEpLgorCisgICAgICAgIEJhZCBET00gcGVyZm9ybWFuY2UgaW4gbGFyZ2UgU1ZHIGZpbGVz CisgICAgICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMDA1NQor CiAgICAgICAgIEN1bGwgUmVuZGVyUGF0aHMgYmVmb3JlIHBhc3Npbmcgb24gdG8gdGhlIHVuZGVy bHlpbmcgZ3JhcGhpY3Mgc3lzdGVtLgogCiAgICAgICAgICogcmVuZGVyaW5nL1JlbmRlclBhdGgu Y3BwOgpkaWZmIC0tZ2l0IGEvV2ViQ29yZS9wbGF0Zm9ybS9ncmFwaGljcy9QYXRoLmggYi9XZWJD b3JlL3BsYXRmb3JtL2dyYXBoaWNzL1BhdGguaAppbmRleCBiZjRjZDlkLi4xODRhOWMyIDEwMDY0 NAotLS0gYS9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL1BhdGguaAorKysgYi9XZWJDb3JlL3Bs YXRmb3JtL2dyYXBoaWNzL1BhdGguaApAQCAtMTQ1LDYgKzE0NSwxMSBAQCBuYW1lc3BhY2UgV2Vi Q29yZSB7CiAKICAgICBwcml2YXRlOgogICAgICAgICBQbGF0Zm9ybVBhdGgqIG1fcGF0aDsKKyNp ZiBQTEFURk9STShDRykKKyAgICAgICAgLy8gQ0dQYXRoSXNFbXB0eSBpcyBleHBlbnNpdmUgc28g d2UgbmVlZCB0byBjYWNoZSB0aGUgcmVzdWx0CisgICAgICAgIG11dGFibGUgYm9vbCBtX2lzRW1w dHlDYWNoZVZhbGlkIDogMTsKKyAgICAgICAgbXV0YWJsZSBib29sIG1fY2FjaGVkSXNFbXB0eSA6 IDE7CisjZW5kaWYKICAgICB9OwogCiB9CmRpZmYgLS1naXQgYS9XZWJDb3JlL3BsYXRmb3JtL2dy YXBoaWNzL2NnL1BhdGhDRy5jcHAgYi9XZWJDb3JlL3BsYXRmb3JtL2dyYXBoaWNzL2NnL1BhdGhD Ry5jcHAKaW5kZXggM2IwNTY0MS4uOWY5MjM5MCAxMDA2NDQKLS0tIGEvV2ViQ29yZS9wbGF0Zm9y bS9ncmFwaGljcy9jZy9QYXRoQ0cuY3BwCisrKyBiL1dlYkNvcmUvcGxhdGZvcm0vZ3JhcGhpY3Mv Y2cvUGF0aENHLmNwcApAQCAtNjcsNiArNjcsOCBAQCBzdGF0aWMgaW5saW5lIENHQ29udGV4dFJl ZiBzY3JhdGNoQ29udGV4dCgpCiAKIFBhdGg6OlBhdGgoKQogICAgIDogbV9wYXRoKENHUGF0aENy ZWF0ZU11dGFibGUoKSkKKyAgICAsIG1faXNFbXB0eUNhY2hlVmFsaWQodHJ1ZSkKKyAgICAsIG1f Y2FjaGVkSXNFbXB0eSh0cnVlKQogewogfQogCkBAIC03Nyw2ICs3OSw4IEBAIFBhdGg6On5QYXRo KCkKIAogUGF0aDo6UGF0aChjb25zdCBQYXRoJiBvdGhlcikKICAgICA6IG1fcGF0aChDR1BhdGhD cmVhdGVNdXRhYmxlQ29weShvdGhlci5tX3BhdGgpKQorICAgICwgbV9pc0VtcHR5Q2FjaGVWYWxp ZChmYWxzZSkKKyAgICAsIG1fY2FjaGVkSXNFbXB0eShmYWxzZSkKIHsKIH0KIApAQCAtODUsNiAr ODksNyBAQCBQYXRoJiBQYXRoOjpvcGVyYXRvcj0oY29uc3QgUGF0aCYgb3RoZXIpCiAgICAgQ0dN dXRhYmxlUGF0aFJlZiBwYXRoID0gQ0dQYXRoQ3JlYXRlTXV0YWJsZUNvcHkob3RoZXIubV9wYXRo KTsKICAgICBDR1BhdGhSZWxlYXNlKG1fcGF0aCk7CiAgICAgbV9wYXRoID0gcGF0aDsKKyAgICBt X2lzRW1wdHlDYWNoZVZhbGlkID0gZmFsc2U7CiAgICAgcmV0dXJuICp0aGlzOwogfQogCkBAIC0x NTksNiArMTY0LDcgQEAgdm9pZCBQYXRoOjp0cmFuc2xhdGUoY29uc3QgRmxvYXRTaXplJiBzaXpl KQogICAgIENHUGF0aEFkZFBhdGgobmV3UGF0aCwgJnRyYW5zbGF0aW9uLCBtX3BhdGgpOwogICAg IENHUGF0aFJlbGVhc2UobV9wYXRoKTsKICAgICBtX3BhdGggPSBuZXdQYXRoOworICAgIG1faXNF bXB0eUNhY2hlVmFsaWQgPSBmYWxzZTsKIH0KIAogRmxvYXRSZWN0IFBhdGg6OmJvdW5kaW5nUmVj dCgpIGNvbnN0CkBAIC0xODIsNjcgKzE4OCw4NCBAQCBGbG9hdFJlY3QgUGF0aDo6c3Ryb2tlQm91 bmRpbmdSZWN0KFN0cm9rZVN0eWxlQXBwbGllciogYXBwbGllcikKICAgICBDR0NvbnRleHRSZXBs YWNlUGF0aFdpdGhTdHJva2VkUGF0aChjb250ZXh0KTsKICAgICBDR1JlY3QgYm94ID0gQ0dDb250 ZXh0SXNQYXRoRW1wdHkoY29udGV4dCkgPyBDR1JlY3RaZXJvIDogQ0dDb250ZXh0R2V0UGF0aEJv dW5kaW5nQm94KGNvbnRleHQpOwogICAgIENHQ29udGV4dFJlc3RvcmVHU3RhdGUoY29udGV4dCk7 Ci0KKyAgICAKKyAgICBtX2lzRW1wdHlDYWNoZVZhbGlkID0gZmFsc2U7CiAgICAgcmV0dXJuIGJv eDsKIH0KIAogdm9pZCBQYXRoOjptb3ZlVG8oY29uc3QgRmxvYXRQb2ludCYgcG9pbnQpCiB7CiAg ICAgQ0dQYXRoTW92ZVRvUG9pbnQobV9wYXRoLCAwLCBwb2ludC54KCksIHBvaW50LnkoKSk7Cisg ICAgbV9pc0VtcHR5Q2FjaGVWYWxpZCA9IGZhbHNlOwogfQogCiB2b2lkIFBhdGg6OmFkZExpbmVU byhjb25zdCBGbG9hdFBvaW50JiBwKQogewogICAgIENHUGF0aEFkZExpbmVUb1BvaW50KG1fcGF0 aCwgMCwgcC54KCksIHAueSgpKTsKKyAgICBtX2lzRW1wdHlDYWNoZVZhbGlkID0gZmFsc2U7CiB9 CiAKIHZvaWQgUGF0aDo6YWRkUXVhZEN1cnZlVG8oY29uc3QgRmxvYXRQb2ludCYgY3AsIGNvbnN0 IEZsb2F0UG9pbnQmIHApCiB7CiAgICAgQ0dQYXRoQWRkUXVhZEN1cnZlVG9Qb2ludChtX3BhdGgs IDAsIGNwLngoKSwgY3AueSgpLCBwLngoKSwgcC55KCkpOworICAgIG1faXNFbXB0eUNhY2hlVmFs aWQgPSBmYWxzZTsKIH0KIAogdm9pZCBQYXRoOjphZGRCZXppZXJDdXJ2ZVRvKGNvbnN0IEZsb2F0 UG9pbnQmIGNwMSwgY29uc3QgRmxvYXRQb2ludCYgY3AyLCBjb25zdCBGbG9hdFBvaW50JiBwKQog ewogICAgIENHUGF0aEFkZEN1cnZlVG9Qb2ludChtX3BhdGgsIDAsIGNwMS54KCksIGNwMS55KCks IGNwMi54KCksIGNwMi55KCksIHAueCgpLCBwLnkoKSk7CisgICAgbV9pc0VtcHR5Q2FjaGVWYWxp ZCA9IGZhbHNlOwogfQogCiB2b2lkIFBhdGg6OmFkZEFyY1RvKGNvbnN0IEZsb2F0UG9pbnQmIHAx LCBjb25zdCBGbG9hdFBvaW50JiBwMiwgZmxvYXQgcmFkaXVzKQogewogICAgIENHUGF0aEFkZEFy Y1RvUG9pbnQobV9wYXRoLCAwLCBwMS54KCksIHAxLnkoKSwgcDIueCgpLCBwMi55KCksIHJhZGl1 cyk7CisgICAgbV9pc0VtcHR5Q2FjaGVWYWxpZCA9IGZhbHNlOwogfQogCiB2b2lkIFBhdGg6OmNs b3NlU3VicGF0aCgpCiB7Ci0gICAgaWYgKCFDR1BhdGhJc0VtcHR5KG1fcGF0aCkpIC8vIHRvIHNp bGVuY2UgYSB3YXJuaW5nIHdoZW4gdHJ5aW5nIHRvIGNsb3NlIGFuIGVtcHR5IHBhdGgKKyAgICBp ZiAoIWlzRW1wdHkoKSkgeyAvLyB0byBzaWxlbmNlIGEgd2FybmluZyB3aGVuIHRyeWluZyB0byBj bG9zZSBhbiBlbXB0eSBwYXRoCiAgICAgICAgIENHUGF0aENsb3NlU3VicGF0aChtX3BhdGgpOwor ICAgICAgICBtX2lzRW1wdHlDYWNoZVZhbGlkID0gZmFsc2U7CisgICAgfQogfQogCiB2b2lkIFBh dGg6OmFkZEFyYyhjb25zdCBGbG9hdFBvaW50JiBwLCBmbG9hdCByLCBmbG9hdCBzYSwgZmxvYXQg ZWEsIGJvb2wgY2xvY2t3aXNlKQogewogICAgIC8vIFdvcmthcm91bmQgZm9yIDxyZGFyOi8vcHJv YmxlbS81MTg5MjMzPiBDR1BhdGhBZGRBcmMgaGFuZ3Mgb3IgY3Jhc2hlcyB3aGVuIHBhc3NlZCBp bmYgYXMgc3RhcnQgb3IgZW5kIGFuZ2xlCi0gICAgaWYgKGlzZmluaXRlKHNhKSAmJiBpc2Zpbml0 ZShlYSkpCisgICAgaWYgKGlzZmluaXRlKHNhKSAmJiBpc2Zpbml0ZShlYSkpIHsKICAgICAgICAg Q0dQYXRoQWRkQXJjKG1fcGF0aCwgMCwgcC54KCksIHAueSgpLCByLCBzYSwgZWEsIGNsb2Nrd2lz ZSk7CisgICAgICAgIG1faXNFbXB0eUNhY2hlVmFsaWQgPSBmYWxzZTsKKyAgICB9CiB9CiAKIHZv aWQgUGF0aDo6YWRkUmVjdChjb25zdCBGbG9hdFJlY3QmIHIpCiB7CiAgICAgQ0dQYXRoQWRkUmVj dChtX3BhdGgsIDAsIHIpOworICAgIG1faXNFbXB0eUNhY2hlVmFsaWQgPSBmYWxzZTsKIH0KIAog dm9pZCBQYXRoOjphZGRFbGxpcHNlKGNvbnN0IEZsb2F0UmVjdCYgcikKIHsKICAgICBDR1BhdGhB ZGRFbGxpcHNlSW5SZWN0KG1fcGF0aCwgMCwgcik7CisgICAgbV9pc0VtcHR5Q2FjaGVWYWxpZCA9 IGZhbHNlOwogfQogCiB2b2lkIFBhdGg6OmNsZWFyKCkKIHsKICAgICBDR1BhdGhSZWxlYXNlKG1f cGF0aCk7CiAgICAgbV9wYXRoID0gQ0dQYXRoQ3JlYXRlTXV0YWJsZSgpOworICAgIG1faXNFbXB0 eUNhY2hlVmFsaWQgPSB0cnVlOworICAgIG1fY2FjaGVkSXNFbXB0eSA9IHRydWU7CiB9CiAKIGJv b2wgUGF0aDo6aXNFbXB0eSgpIGNvbnN0CiB7Ci0gICAgcmV0dXJuIENHUGF0aElzRW1wdHkobV9w YXRoKTsKKyAgICBpZiAobV9pc0VtcHR5Q2FjaGVWYWxpZCkKKyAgICAgICAgcmV0dXJuIG1fY2Fj aGVkSXNFbXB0eTsKKyAgICBtX2lzRW1wdHlDYWNoZVZhbGlkID0gdHJ1ZTsKKyAgICByZXR1cm4g bV9jYWNoZWRJc0VtcHR5ID0gQ0dQYXRoSXNFbXB0eShtX3BhdGgpOwogfQogCiBib29sIFBhdGg6 Omhhc0N1cnJlbnRQb2ludCgpIGNvbnN0CkBAIC0zNTMsNiArMzc2LDcgQEAgdm9pZCBQYXRoOjp0 cmFuc2Zvcm0oY29uc3QgVHJhbnNmb3JtYXRpb25NYXRyaXgmIHRyYW5zZm9ybSkKICAgICBDR1Bh dGhBZGRQYXRoKHBhdGgsICZ0cmFuc2Zvcm1DRywgbV9wYXRoKTsKICAgICBDR1BhdGhSZWxlYXNl KG1fcGF0aCk7CiAgICAgbV9wYXRoID0gcGF0aDsKKyAgICBtX2lzRW1wdHlDYWNoZVZhbGlkID0g ZmFsc2U7CiB9CiAKIH0K 46667 2010-01-15 03:49:24 -0800 2010-01-15 03:51:49 -0800 Patch bug-30055-20100115034923.patch text/plain 3055 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwIGIvV2ViQ29yZS9y ZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAKaW5kZXggOGNlODdjZC4uODA5ZGZlYyAxMDA2NDQKLS0t IGEvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAKKysrIGIvV2ViQ29yZS9yZW5kZXJp bmcvUmVuZGVyUGF0aC5jcHAKQEAgLTIxMywzMiArMjEzLDQwIEBAIHZvaWQgUmVuZGVyUGF0aDo6 cGFpbnQoUGFpbnRJbmZvJiBwYWludEluZm8sIGludCwgaW50KQogewogICAgIGlmIChwYWludElu Zm8uY29udGV4dC0+cGFpbnRpbmdEaXNhYmxlZCgpIHx8IHN0eWxlKCktPnZpc2liaWxpdHkoKSA9 PSBISURERU4gfHwgbV9wYXRoLmlzRW1wdHkoKSkKICAgICAgICAgcmV0dXJuOwotICAgICAgICAg ICAgCi0gICAgcGFpbnRJbmZvLmNvbnRleHQtPnNhdmUoKTsKLSAgICBwYWludEluZm8uY29udGV4 dC0+Y29uY2F0Q1RNKGxvY2FsVG9QYXJlbnRUcmFuc2Zvcm0oKSk7CisKKyAgICBQYWludEluZm8g Y2hpbGRQYWludEluZm8ocGFpbnRJbmZvKTsKKyAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5z YXZlKCk7CisgICAgYXBwbHlUcmFuc2Zvcm1Ub1BhaW50SW5mbyhjaGlsZFBhaW50SW5mbywgbV9s b2NhbFRyYW5zZm9ybSk7CisgICAgRmxvYXRSZWN0IGJvdW5kaW5nQm94ID0gcmVwYWludFJlY3RJ bkxvY2FsQ29vcmRpbmF0ZXMoKTsKKyAgICAvLyBGSVhNRTogVGhlIGVtcHR5IHJlY3QgY2hlY2sg aXMgdG8gZGVhbCB3aXRoIGluY29ycmVjdCBpbml0aWFsIGNsaXAgaW4gcmVuZGVyU3VidHJlZVRv SW1hZ2UKKyAgICAvLyB1bmZvcnR1bmF0ZWx5IGZpeGluZyB0aGF0IHByb2JsZW0gaXMgZmFpcmx5 IGNvbXBsZXggdW5sZXNzIHdlIHdlcmUgd2lsbGluZyB0byBqdXN0IGZ1dHogdGhlCisgICAgLy8g cmVjdCB0byBzb21ldGhpbmcgImNsb3NlIGVub3VnaCIKKyAgICBpZiAoIWJvdW5kaW5nQm94Lmlu dGVyc2VjdHMoY2hpbGRQYWludEluZm8ucmVjdCkgJiYgIWNoaWxkUGFpbnRJbmZvLnJlY3QuaXNF bXB0eSgpKSB7CisgICAgICAgIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQtPnJlc3RvcmUoKTsKKyAg ICAgICAgcmV0dXJuOworICAgIH0KIAogICAgIFNWR1Jlc291cmNlRmlsdGVyKiBmaWx0ZXIgPSAw OwogCi0gICAgRmxvYXRSZWN0IGJvdW5kaW5nQm94ID0gcmVwYWludFJlY3RJbkxvY2FsQ29vcmRp bmF0ZXMoKTsKLSAgICBpZiAocGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VGb3JlZ3JvdW5k KSB7Ci0gICAgICAgIFBhaW50SW5mbyBzYXZlZEluZm8ocGFpbnRJbmZvKTsKKyAgICBpZiAoY2hp bGRQYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZUZvcmVncm91bmQpIHsKKyAgICAgICAgUGFp bnRJbmZvIHNhdmVkSW5mbyhjaGlsZFBhaW50SW5mbyk7CiAKLSAgICAgICAgaWYgKHByZXBhcmVU b1JlbmRlclNWR0NvbnRlbnQodGhpcywgcGFpbnRJbmZvLCBib3VuZGluZ0JveCwgZmlsdGVyKSkg eworICAgICAgICBpZiAocHJlcGFyZVRvUmVuZGVyU1ZHQ29udGVudCh0aGlzLCBjaGlsZFBhaW50 SW5mbywgYm91bmRpbmdCb3gsIGZpbHRlcikpIHsKICAgICAgICAgICAgIGlmIChzdHlsZSgpLT5z dmdTdHlsZSgpLT5zaGFwZVJlbmRlcmluZygpID09IFNSX0NSSVNQRURHRVMpCi0gICAgICAgICAg ICAgICAgcGFpbnRJbmZvLmNvbnRleHQtPnNldFNob3VsZEFudGlhbGlhcyhmYWxzZSk7Ci0gICAg ICAgICAgICBmaWxsQW5kU3Ryb2tlUGF0aChtX3BhdGgsIHBhaW50SW5mby5jb250ZXh0LCBzdHls ZSgpLCB0aGlzKTsKKyAgICAgICAgICAgICAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5zZXRT aG91bGRBbnRpYWxpYXMoZmFsc2UpOworICAgICAgICAgICAgZmlsbEFuZFN0cm9rZVBhdGgobV9w YXRoLCBjaGlsZFBhaW50SW5mby5jb250ZXh0LCBzdHlsZSgpLCB0aGlzKTsKIAogICAgICAgICAg ICAgaWYgKHN0YXRpY19jYXN0PFNWR1N0eWxlZEVsZW1lbnQqPihub2RlKCkpLT5zdXBwb3J0c01h cmtlcnMoKSkKLSAgICAgICAgICAgICAgICBtX21hcmtlckxheW91dEluZm8uZHJhd01hcmtlcnMo cGFpbnRJbmZvKTsKKyAgICAgICAgICAgICAgICBtX21hcmtlckxheW91dEluZm8uZHJhd01hcmtl cnMoY2hpbGRQYWludEluZm8pOwogICAgICAgICB9Ci0gICAgICAgIGZpbmlzaFJlbmRlclNWR0Nv bnRlbnQodGhpcywgcGFpbnRJbmZvLCBmaWx0ZXIsIHNhdmVkSW5mby5jb250ZXh0KTsKKyAgICAg ICAgZmluaXNoUmVuZGVyU1ZHQ29udGVudCh0aGlzLCBjaGlsZFBhaW50SW5mbywgZmlsdGVyLCBz YXZlZEluZm8uY29udGV4dCk7CiAgICAgfQogCi0gICAgaWYgKChwYWludEluZm8ucGhhc2UgPT0g UGFpbnRQaGFzZU91dGxpbmUgfHwgcGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VTZWxmT3V0 bGluZSkgJiYgc3R5bGUoKS0+b3V0bGluZVdpZHRoKCkpCi0gICAgICAgIHBhaW50T3V0bGluZShw YWludEluZm8uY29udGV4dCwgc3RhdGljX2Nhc3Q8aW50Pihib3VuZGluZ0JveC54KCkpLCBzdGF0 aWNfY2FzdDxpbnQ+KGJvdW5kaW5nQm94LnkoKSksCisgICAgaWYgKChjaGlsZFBhaW50SW5mby5w aGFzZSA9PSBQYWludFBoYXNlT3V0bGluZSB8fCBjaGlsZFBhaW50SW5mby5waGFzZSA9PSBQYWlu dFBoYXNlU2VsZk91dGxpbmUpICYmIHN0eWxlKCktPm91dGxpbmVXaWR0aCgpKQorICAgICAgICBw YWludE91dGxpbmUoY2hpbGRQYWludEluZm8uY29udGV4dCwgc3RhdGljX2Nhc3Q8aW50Pihib3Vu ZGluZ0JveC54KCkpLCBzdGF0aWNfY2FzdDxpbnQ+KGJvdW5kaW5nQm94LnkoKSksCiAgICAgICAg ICAgICBzdGF0aWNfY2FzdDxpbnQ+KGJvdW5kaW5nQm94LndpZHRoKCkpLCBzdGF0aWNfY2FzdDxp bnQ+KGJvdW5kaW5nQm94LmhlaWdodCgpKSwgc3R5bGUoKSk7CiAgICAgCi0gICAgcGFpbnRJbmZv LmNvbnRleHQtPnJlc3RvcmUoKTsKKyAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5yZXN0b3Jl KCk7CiB9CiAKIC8vIFRoaXMgbWV0aG9kIGlzIGNhbGxlZCBmcm9tIGluc2lkZSBwYWludE91dGxp bmUoKSBzaW5jZSB3ZSBjYWxsIHBhaW50T3V0bGluZSgpCg== 46668 2010-01-15 03:51:56 -0800 2010-01-15 12:48:16 -0800 Patch bug-30055-20100115035155.patch text/plain 3670 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvQ2hhbmdlTG9nIGIvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXgg Yzg1ODBlMC4uMWM1Y2M4MCAxMDA2NDQKLS0tIGEvV2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvV2Vi Q29yZS9DaGFuZ2VMb2cKQEAgLTEsMyArMSwxNiBAQAorMjAxMC0wMS0xNSAgT2xpdmVyIEh1bnQg IDxvbGl2ZXJAYXBwbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEp LgorCisgICAgICAgIEJhZCBET00gcGVyZm9ybWFuY2UgaW4gbGFyZ2UgU1ZHIGZpbGVzCisgICAg ICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMDA1NQorCisgICAg ICAgIEFkZCBhbiBlYXJseSByZXR1cm4gd2hlbiB3ZSBnbyB0byBwYWludCBhIFJlbmRlclBhdGgg dGhhdAorICAgICAgICBpc24ndCBpbiB0aGUgY3VycmVudCBjbGlwLgorCisgICAgICAgICogcmVu ZGVyaW5nL1JlbmRlclBhdGguY3BwOgorICAgICAgICAoV2ViQ29yZTo6UmVuZGVyUGF0aDo6cGFp bnQpOgorCiAyMDEwLTAxLTE0ICBLd2FuZyBZdWwgU2VvICA8c2t5dWxAY29tcGFueTEwMC5uZXQ+ CiAKICAgICAgICAgUmV2aWV3ZWQgYnkgQWxleGV5IFByb3NrdXJ5YWtvdi4KZGlmZiAtLWdpdCBh L1dlYkNvcmUvcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwIGIvV2ViQ29yZS9yZW5kZXJpbmcvUmVu ZGVyUGF0aC5jcHAKaW5kZXggOGNlODdjZC4uODA5ZGZlYyAxMDA2NDQKLS0tIGEvV2ViQ29yZS9y ZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAKKysrIGIvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0 aC5jcHAKQEAgLTIxMywzMiArMjEzLDQwIEBAIHZvaWQgUmVuZGVyUGF0aDo6cGFpbnQoUGFpbnRJ bmZvJiBwYWludEluZm8sIGludCwgaW50KQogewogICAgIGlmIChwYWludEluZm8uY29udGV4dC0+ cGFpbnRpbmdEaXNhYmxlZCgpIHx8IHN0eWxlKCktPnZpc2liaWxpdHkoKSA9PSBISURERU4gfHwg bV9wYXRoLmlzRW1wdHkoKSkKICAgICAgICAgcmV0dXJuOwotICAgICAgICAgICAgCi0gICAgcGFp bnRJbmZvLmNvbnRleHQtPnNhdmUoKTsKLSAgICBwYWludEluZm8uY29udGV4dC0+Y29uY2F0Q1RN KGxvY2FsVG9QYXJlbnRUcmFuc2Zvcm0oKSk7CisKKyAgICBQYWludEluZm8gY2hpbGRQYWludElu Zm8ocGFpbnRJbmZvKTsKKyAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5zYXZlKCk7CisgICAg YXBwbHlUcmFuc2Zvcm1Ub1BhaW50SW5mbyhjaGlsZFBhaW50SW5mbywgbV9sb2NhbFRyYW5zZm9y bSk7CisgICAgRmxvYXRSZWN0IGJvdW5kaW5nQm94ID0gcmVwYWludFJlY3RJbkxvY2FsQ29vcmRp bmF0ZXMoKTsKKyAgICAvLyBGSVhNRTogVGhlIGVtcHR5IHJlY3QgY2hlY2sgaXMgdG8gZGVhbCB3 aXRoIGluY29ycmVjdCBpbml0aWFsIGNsaXAgaW4gcmVuZGVyU3VidHJlZVRvSW1hZ2UKKyAgICAv LyB1bmZvcnR1bmF0ZWx5IGZpeGluZyB0aGF0IHByb2JsZW0gaXMgZmFpcmx5IGNvbXBsZXggdW5s ZXNzIHdlIHdlcmUgd2lsbGluZyB0byBqdXN0IGZ1dHogdGhlCisgICAgLy8gcmVjdCB0byBzb21l dGhpbmcgImNsb3NlIGVub3VnaCIKKyAgICBpZiAoIWJvdW5kaW5nQm94LmludGVyc2VjdHMoY2hp bGRQYWludEluZm8ucmVjdCkgJiYgIWNoaWxkUGFpbnRJbmZvLnJlY3QuaXNFbXB0eSgpKSB7Cisg ICAgICAgIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQtPnJlc3RvcmUoKTsKKyAgICAgICAgcmV0dXJu OworICAgIH0KIAogICAgIFNWR1Jlc291cmNlRmlsdGVyKiBmaWx0ZXIgPSAwOwogCi0gICAgRmxv YXRSZWN0IGJvdW5kaW5nQm94ID0gcmVwYWludFJlY3RJbkxvY2FsQ29vcmRpbmF0ZXMoKTsKLSAg ICBpZiAocGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VGb3JlZ3JvdW5kKSB7Ci0gICAgICAg IFBhaW50SW5mbyBzYXZlZEluZm8ocGFpbnRJbmZvKTsKKyAgICBpZiAoY2hpbGRQYWludEluZm8u cGhhc2UgPT0gUGFpbnRQaGFzZUZvcmVncm91bmQpIHsKKyAgICAgICAgUGFpbnRJbmZvIHNhdmVk SW5mbyhjaGlsZFBhaW50SW5mbyk7CiAKLSAgICAgICAgaWYgKHByZXBhcmVUb1JlbmRlclNWR0Nv bnRlbnQodGhpcywgcGFpbnRJbmZvLCBib3VuZGluZ0JveCwgZmlsdGVyKSkgeworICAgICAgICBp ZiAocHJlcGFyZVRvUmVuZGVyU1ZHQ29udGVudCh0aGlzLCBjaGlsZFBhaW50SW5mbywgYm91bmRp bmdCb3gsIGZpbHRlcikpIHsKICAgICAgICAgICAgIGlmIChzdHlsZSgpLT5zdmdTdHlsZSgpLT5z aGFwZVJlbmRlcmluZygpID09IFNSX0NSSVNQRURHRVMpCi0gICAgICAgICAgICAgICAgcGFpbnRJ bmZvLmNvbnRleHQtPnNldFNob3VsZEFudGlhbGlhcyhmYWxzZSk7Ci0gICAgICAgICAgICBmaWxs QW5kU3Ryb2tlUGF0aChtX3BhdGgsIHBhaW50SW5mby5jb250ZXh0LCBzdHlsZSgpLCB0aGlzKTsK KyAgICAgICAgICAgICAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5zZXRTaG91bGRBbnRpYWxp YXMoZmFsc2UpOworICAgICAgICAgICAgZmlsbEFuZFN0cm9rZVBhdGgobV9wYXRoLCBjaGlsZFBh aW50SW5mby5jb250ZXh0LCBzdHlsZSgpLCB0aGlzKTsKIAogICAgICAgICAgICAgaWYgKHN0YXRp Y19jYXN0PFNWR1N0eWxlZEVsZW1lbnQqPihub2RlKCkpLT5zdXBwb3J0c01hcmtlcnMoKSkKLSAg ICAgICAgICAgICAgICBtX21hcmtlckxheW91dEluZm8uZHJhd01hcmtlcnMocGFpbnRJbmZvKTsK KyAgICAgICAgICAgICAgICBtX21hcmtlckxheW91dEluZm8uZHJhd01hcmtlcnMoY2hpbGRQYWlu dEluZm8pOwogICAgICAgICB9Ci0gICAgICAgIGZpbmlzaFJlbmRlclNWR0NvbnRlbnQodGhpcywg cGFpbnRJbmZvLCBmaWx0ZXIsIHNhdmVkSW5mby5jb250ZXh0KTsKKyAgICAgICAgZmluaXNoUmVu ZGVyU1ZHQ29udGVudCh0aGlzLCBjaGlsZFBhaW50SW5mbywgZmlsdGVyLCBzYXZlZEluZm8uY29u dGV4dCk7CiAgICAgfQogCi0gICAgaWYgKChwYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZU91 dGxpbmUgfHwgcGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VTZWxmT3V0bGluZSkgJiYgc3R5 bGUoKS0+b3V0bGluZVdpZHRoKCkpCi0gICAgICAgIHBhaW50T3V0bGluZShwYWludEluZm8uY29u dGV4dCwgc3RhdGljX2Nhc3Q8aW50Pihib3VuZGluZ0JveC54KCkpLCBzdGF0aWNfY2FzdDxpbnQ+ KGJvdW5kaW5nQm94LnkoKSksCisgICAgaWYgKChjaGlsZFBhaW50SW5mby5waGFzZSA9PSBQYWlu dFBoYXNlT3V0bGluZSB8fCBjaGlsZFBhaW50SW5mby5waGFzZSA9PSBQYWludFBoYXNlU2VsZk91 dGxpbmUpICYmIHN0eWxlKCktPm91dGxpbmVXaWR0aCgpKQorICAgICAgICBwYWludE91dGxpbmUo Y2hpbGRQYWludEluZm8uY29udGV4dCwgc3RhdGljX2Nhc3Q8aW50Pihib3VuZGluZ0JveC54KCkp LCBzdGF0aWNfY2FzdDxpbnQ+KGJvdW5kaW5nQm94LnkoKSksCiAgICAgICAgICAgICBzdGF0aWNf Y2FzdDxpbnQ+KGJvdW5kaW5nQm94LndpZHRoKCkpLCBzdGF0aWNfY2FzdDxpbnQ+KGJvdW5kaW5n Qm94LmhlaWdodCgpKSwgc3R5bGUoKSk7CiAgICAgCi0gICAgcGFpbnRJbmZvLmNvbnRleHQtPnJl c3RvcmUoKTsKKyAgICBjaGlsZFBhaW50SW5mby5jb250ZXh0LT5yZXN0b3JlKCk7CiB9CiAKIC8v IFRoaXMgbWV0aG9kIGlzIGNhbGxlZCBmcm9tIGluc2lkZSBwYWludE91dGxpbmUoKSBzaW5jZSB3 ZSBjYWxsIHBhaW50T3V0bGluZSgpCg== 46701 2010-01-15 12:48:23 -0800 2010-01-15 16:04:16 -0800 Patch bug-30055-20100115124821.patch text/plain 4308 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvQ2hhbmdlTG9nIGIvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXgg NWZlNjM5OC4uYTI0NjljNyAxMDA2NDQKLS0tIGEvV2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvV2Vi Q29yZS9DaGFuZ2VMb2cKQEAgLTEsMyArMSwxOCBAQAorMjAxMC0wMS0xNSAgT2xpdmVyIEh1bnQg IDxvbGl2ZXJAYXBwbGUuY29tPgorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEp LgorCisgICAgICAgIEJhZCBET00gcGVyZm9ybWFuY2UgaW4gbGFyZ2UgU1ZHIGZpbGVzCisgICAg ICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMDA1NQorCisgICAg ICAgIEFkZCBhbiBlYXJseSByZXR1cm4gd2hlbiB3ZSBnbyB0byBwYWludCBhIFJlbmRlclBhdGgg dGhhdAorICAgICAgICBpc24ndCBpbiB0aGUgY3VycmVudCBjbGlwLgorCisgICAgICAgICogcmVu ZGVyaW5nL1JlbmRlclBhdGguY3BwOgorICAgICAgICAoV2ViQ29yZTo6UmVuZGVyUGF0aDo6cGFp bnQpOgorICAgICAgICAqIHN2Zy9ncmFwaGljcy9TVkdJbWFnZS5jcHA6CisgICAgICAgIChXZWJD b3JlOjpTVkdJbWFnZTo6ZHJhdyk6CisKIDIwMTAtMDEtMTUgIEFsZXhhbmRlciBQYXZsb3YgIDxh cGF2bG92QGNocm9taXVtLm9yZz4KIAogICAgICAgICBSZXZpZXdlZCBieSBQYXZlbCBGZWxkbWFu LgpkaWZmIC0tZ2l0IGEvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAgYi9XZWJDb3Jl L3JlbmRlcmluZy9SZW5kZXJQYXRoLmNwcAppbmRleCA4NmQyODBiLi44MWNkZGZhIDEwMDY0NAot LS0gYS9XZWJDb3JlL3JlbmRlcmluZy9SZW5kZXJQYXRoLmNwcAorKysgYi9XZWJDb3JlL3JlbmRl cmluZy9SZW5kZXJQYXRoLmNwcApAQCAtMjE1LDMyICsyMTUsNDAgQEAgdm9pZCBSZW5kZXJQYXRo OjpwYWludChQYWludEluZm8mIHBhaW50SW5mbywgaW50LCBpbnQpCiB7CiAgICAgaWYgKHBhaW50 SW5mby5jb250ZXh0LT5wYWludGluZ0Rpc2FibGVkKCkgfHwgc3R5bGUoKS0+dmlzaWJpbGl0eSgp ID09IEhJRERFTiB8fCBtX3BhdGguaXNFbXB0eSgpKQogICAgICAgICByZXR1cm47Ci0gICAgICAg ICAgICAKLSAgICBwYWludEluZm8uY29udGV4dC0+c2F2ZSgpOwotICAgIHBhaW50SW5mby5jb250 ZXh0LT5jb25jYXRDVE0obG9jYWxUb1BhcmVudFRyYW5zZm9ybSgpKTsKKworICAgIFBhaW50SW5m byBjaGlsZFBhaW50SW5mbyhwYWludEluZm8pOworICAgIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQt PnNhdmUoKTsKKyAgICBhcHBseVRyYW5zZm9ybVRvUGFpbnRJbmZvKGNoaWxkUGFpbnRJbmZvLCBt X2xvY2FsVHJhbnNmb3JtKTsKKyAgICBGbG9hdFJlY3QgYm91bmRpbmdCb3ggPSByZXBhaW50UmVj dEluTG9jYWxDb29yZGluYXRlcygpOworICAgIC8vIEZJWE1FOiBUaGUgZW1wdHkgcmVjdCBjaGVj ayBpcyB0byBkZWFsIHdpdGggaW5jb3JyZWN0IGluaXRpYWwgY2xpcCBpbiByZW5kZXJTdWJ0cmVl VG9JbWFnZQorICAgIC8vIHVuZm9ydHVuYXRlbHkgZml4aW5nIHRoYXQgcHJvYmxlbSBpcyBmYWly bHkgY29tcGxleCB1bmxlc3Mgd2Ugd2VyZSB3aWxsaW5nIHRvIGp1c3QgZnV0eiB0aGUKKyAgICAv LyByZWN0IHRvIHNvbWV0aGluZyAiY2xvc2UgZW5vdWdoIgorICAgIGlmICghYm91bmRpbmdCb3gu aW50ZXJzZWN0cyhjaGlsZFBhaW50SW5mby5yZWN0KSAmJiAhY2hpbGRQYWludEluZm8ucmVjdC5p c0VtcHR5KCkpIHsKKyAgICAgICAgY2hpbGRQYWludEluZm8uY29udGV4dC0+cmVzdG9yZSgpOwor ICAgICAgICByZXR1cm47CisgICAgfQogCiAgICAgU1ZHUmVzb3VyY2VGaWx0ZXIqIGZpbHRlciA9 IDA7CiAKLSAgICBGbG9hdFJlY3QgYm91bmRpbmdCb3ggPSByZXBhaW50UmVjdEluTG9jYWxDb29y ZGluYXRlcygpOwotICAgIGlmIChwYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZUZvcmVncm91 bmQpIHsKLSAgICAgICAgUGFpbnRJbmZvIHNhdmVkSW5mbyhwYWludEluZm8pOworICAgIGlmIChj aGlsZFBhaW50SW5mby5waGFzZSA9PSBQYWludFBoYXNlRm9yZWdyb3VuZCkgeworICAgICAgICBQ YWludEluZm8gc2F2ZWRJbmZvKGNoaWxkUGFpbnRJbmZvKTsKIAotICAgICAgICBpZiAocHJlcGFy ZVRvUmVuZGVyU1ZHQ29udGVudCh0aGlzLCBwYWludEluZm8sIGJvdW5kaW5nQm94LCBmaWx0ZXIp KSB7CisgICAgICAgIGlmIChwcmVwYXJlVG9SZW5kZXJTVkdDb250ZW50KHRoaXMsIGNoaWxkUGFp bnRJbmZvLCBib3VuZGluZ0JveCwgZmlsdGVyKSkgewogICAgICAgICAgICAgaWYgKHN0eWxlKCkt PnN2Z1N0eWxlKCktPnNoYXBlUmVuZGVyaW5nKCkgPT0gU1JfQ1JJU1BFREdFUykKLSAgICAgICAg ICAgICAgICBwYWludEluZm8uY29udGV4dC0+c2V0U2hvdWxkQW50aWFsaWFzKGZhbHNlKTsKLSAg ICAgICAgICAgIGZpbGxBbmRTdHJva2VQYXRoKG1fcGF0aCwgcGFpbnRJbmZvLmNvbnRleHQsIHN0 eWxlKCksIHRoaXMpOworICAgICAgICAgICAgICAgIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQtPnNl dFNob3VsZEFudGlhbGlhcyhmYWxzZSk7CisgICAgICAgICAgICBmaWxsQW5kU3Ryb2tlUGF0aCht X3BhdGgsIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQsIHN0eWxlKCksIHRoaXMpOwogCiAgICAgICAg ICAgICBpZiAoc3RhdGljX2Nhc3Q8U1ZHU3R5bGVkRWxlbWVudCo+KG5vZGUoKSktPnN1cHBvcnRz TWFya2VycygpKQotICAgICAgICAgICAgICAgIG1fbWFya2VyTGF5b3V0SW5mby5kcmF3TWFya2Vy cyhwYWludEluZm8pOworICAgICAgICAgICAgICAgIG1fbWFya2VyTGF5b3V0SW5mby5kcmF3TWFy a2VycyhjaGlsZFBhaW50SW5mbyk7CiAgICAgICAgIH0KLSAgICAgICAgZmluaXNoUmVuZGVyU1ZH Q29udGVudCh0aGlzLCBwYWludEluZm8sIGZpbHRlciwgc2F2ZWRJbmZvLmNvbnRleHQpOworICAg ICAgICBmaW5pc2hSZW5kZXJTVkdDb250ZW50KHRoaXMsIGNoaWxkUGFpbnRJbmZvLCBmaWx0ZXIs IHNhdmVkSW5mby5jb250ZXh0KTsKICAgICB9CiAKLSAgICBpZiAoKHBhaW50SW5mby5waGFzZSA9 PSBQYWludFBoYXNlT3V0bGluZSB8fCBwYWludEluZm8ucGhhc2UgPT0gUGFpbnRQaGFzZVNlbGZP dXRsaW5lKSAmJiBzdHlsZSgpLT5vdXRsaW5lV2lkdGgoKSkKLSAgICAgICAgcGFpbnRPdXRsaW5l KHBhaW50SW5mby5jb250ZXh0LCBzdGF0aWNfY2FzdDxpbnQ+KGJvdW5kaW5nQm94LngoKSksIHN0 YXRpY19jYXN0PGludD4oYm91bmRpbmdCb3gueSgpKSwKKyAgICBpZiAoKGNoaWxkUGFpbnRJbmZv LnBoYXNlID09IFBhaW50UGhhc2VPdXRsaW5lIHx8IGNoaWxkUGFpbnRJbmZvLnBoYXNlID09IFBh aW50UGhhc2VTZWxmT3V0bGluZSkgJiYgc3R5bGUoKS0+b3V0bGluZVdpZHRoKCkpCisgICAgICAg IHBhaW50T3V0bGluZShjaGlsZFBhaW50SW5mby5jb250ZXh0LCBzdGF0aWNfY2FzdDxpbnQ+KGJv dW5kaW5nQm94LngoKSksIHN0YXRpY19jYXN0PGludD4oYm91bmRpbmdCb3gueSgpKSwKICAgICAg ICAgICAgIHN0YXRpY19jYXN0PGludD4oYm91bmRpbmdCb3gud2lkdGgoKSksIHN0YXRpY19jYXN0 PGludD4oYm91bmRpbmdCb3guaGVpZ2h0KCkpLCBzdHlsZSgpKTsKICAgICAKLSAgICBwYWludElu Zm8uY29udGV4dC0+cmVzdG9yZSgpOworICAgIGNoaWxkUGFpbnRJbmZvLmNvbnRleHQtPnJlc3Rv cmUoKTsKIH0KIAogLy8gVGhpcyBtZXRob2QgaXMgY2FsbGVkIGZyb20gaW5zaWRlIHBhaW50T3V0 bGluZSgpIHNpbmNlIHdlIGNhbGwgcGFpbnRPdXRsaW5lKCkKZGlmZiAtLWdpdCBhL1dlYkNvcmUv c3ZnL2dyYXBoaWNzL1NWR0ltYWdlLmNwcCBiL1dlYkNvcmUvc3ZnL2dyYXBoaWNzL1NWR0ltYWdl LmNwcAppbmRleCAwNTBkY2NhLi4zNDhkZjRmIDEwMDY0NAotLS0gYS9XZWJDb3JlL3N2Zy9ncmFw aGljcy9TVkdJbWFnZS5jcHAKKysrIGIvV2ViQ29yZS9zdmcvZ3JhcGhpY3MvU1ZHSW1hZ2UuY3Bw CkBAIC0xOTMsNyArMTkzLDcgQEAgdm9pZCBTVkdJbWFnZTo6ZHJhdyhHcmFwaGljc0NvbnRleHQq IGNvbnRleHQsIGNvbnN0IEZsb2F0UmVjdCYgZHN0UmVjdCwgY29uc3QgRmwKIAogICAgIGlmICh2 aWV3LT5uZWVkc0xheW91dCgpKQogICAgICAgICB2aWV3LT5sYXlvdXQoKTsKLSAgICB2aWV3LT5w YWludChjb250ZXh0LCBlbmNsb3NpbmdJbnRSZWN0KHNyY1JlY3QpKTsKKyAgICB2aWV3LT5wYWlu dChjb250ZXh0LCBJbnRSZWN0KDAsIDAsIHZpZXctPndpZHRoKCksIHZpZXctPmhlaWdodCgpKSk7 CiAKICAgICBpZiAoY29tcG9zaXRlT3AgIT0gQ29tcG9zaXRlU291cmNlT3ZlcikKICAgICAgICAg Y29udGV4dC0+ZW5kVHJhbnNwYXJlbmN5TGF5ZXIoKTsK 46715 2010-01-15 16:04:23 -0800 2010-01-15 16:14:56 -0800 Patch bug-30055-20100115160422.patch text/plain 2008 oliver ZGlmZiAtLWdpdCBhL1dlYkNvcmUvQ2hhbmdlTG9nIGIvV2ViQ29yZS9DaGFuZ2VMb2cKaW5kZXgg ZWZmNmYwZS4uMDM5YTgyNiAxMDA2NDQKLS0tIGEvV2ViQ29yZS9DaGFuZ2VMb2cKKysrIGIvV2Vi Q29yZS9DaGFuZ2VMb2cKQEAgLTEsNSArMSwxNyBAQAogMjAxMC0wMS0xNSAgT2xpdmVyIEh1bnQg IDxvbGl2ZXJAYXBwbGUuY29tPgogCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEp LgorCisgICAgICAgIEJhZCBET00gcGVyZm9ybWFuY2UgaW4gbGFyZ2UgU1ZHIGZpbGVzCisgICAg ICAgIGh0dHBzOi8vYnVncy53ZWJraXQub3JnL3Nob3dfYnVnLmNnaT9pZD0zMDA1NQorCisgICAg ICAgIEltcHJvdmUgdGhlIHBlcmZvcm1hbmNlIG9mIHRoZSBlYXJseSBjdWxsaW5nIGNoZWNrCisK KyAgICAgICAgKiByZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHA6CisgICAgICAgIChXZWJDb3JlOjpS ZW5kZXJQYXRoOjpwYWludCk6CisKKzIwMTAtMDEtMTUgIE9saXZlciBIdW50ICA8b2xpdmVyQGFw cGxlLmNvbT4KKwogICAgICAgICBSZXZpZXdlZCBieSBEaXJrIFNjaHVsemUuCiAKICAgICAgICAg QmFkIERPTSBwZXJmb3JtYW5jZSBpbiBsYXJnZSBTVkcgZmlsZXMKZGlmZiAtLWdpdCBhL1dlYkNv cmUvcmVuZGVyaW5nL1JlbmRlclBhdGguY3BwIGIvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0 aC5jcHAKaW5kZXggODFjZGRmYS4uMWQ0MmQ3ZSAxMDA2NDQKLS0tIGEvV2ViQ29yZS9yZW5kZXJp bmcvUmVuZGVyUGF0aC5jcHAKKysrIGIvV2ViQ29yZS9yZW5kZXJpbmcvUmVuZGVyUGF0aC5jcHAK QEAgLTIxNiwxOCArMjE2LDE3IEBAIHZvaWQgUmVuZGVyUGF0aDo6cGFpbnQoUGFpbnRJbmZvJiBw YWludEluZm8sIGludCwgaW50KQogICAgIGlmIChwYWludEluZm8uY29udGV4dC0+cGFpbnRpbmdE aXNhYmxlZCgpIHx8IHN0eWxlKCktPnZpc2liaWxpdHkoKSA9PSBISURERU4gfHwgbV9wYXRoLmlz RW1wdHkoKSkKICAgICAgICAgcmV0dXJuOwogCi0gICAgUGFpbnRJbmZvIGNoaWxkUGFpbnRJbmZv KHBhaW50SW5mbyk7Ci0gICAgY2hpbGRQYWludEluZm8uY29udGV4dC0+c2F2ZSgpOwotICAgIGFw cGx5VHJhbnNmb3JtVG9QYWludEluZm8oY2hpbGRQYWludEluZm8sIG1fbG9jYWxUcmFuc2Zvcm0p OwogICAgIEZsb2F0UmVjdCBib3VuZGluZ0JveCA9IHJlcGFpbnRSZWN0SW5Mb2NhbENvb3JkaW5h dGVzKCk7CisgICAgRmxvYXRSZWN0IG5vbkxvY2FsQm91bmRpbmdCb3ggPSBtX2xvY2FsVHJhbnNm b3JtLm1hcFJlY3QoYm91bmRpbmdCb3gpOwogICAgIC8vIEZJWE1FOiBUaGUgZW1wdHkgcmVjdCBj aGVjayBpcyB0byBkZWFsIHdpdGggaW5jb3JyZWN0IGluaXRpYWwgY2xpcCBpbiByZW5kZXJTdWJ0 cmVlVG9JbWFnZQogICAgIC8vIHVuZm9ydHVuYXRlbHkgZml4aW5nIHRoYXQgcHJvYmxlbSBpcyBm YWlybHkgY29tcGxleCB1bmxlc3Mgd2Ugd2VyZSB3aWxsaW5nIHRvIGp1c3QgZnV0eiB0aGUKICAg ICAvLyByZWN0IHRvIHNvbWV0aGluZyAiY2xvc2UgZW5vdWdoIgotICAgIGlmICghYm91bmRpbmdC b3guaW50ZXJzZWN0cyhjaGlsZFBhaW50SW5mby5yZWN0KSAmJiAhY2hpbGRQYWludEluZm8ucmVj dC5pc0VtcHR5KCkpIHsKLSAgICAgICAgY2hpbGRQYWludEluZm8uY29udGV4dC0+cmVzdG9yZSgp OworICAgIGlmICghbm9uTG9jYWxCb3VuZGluZ0JveC5pbnRlcnNlY3RzKHBhaW50SW5mby5yZWN0 KSAmJiAhcGFpbnRJbmZvLnJlY3QuaXNFbXB0eSgpKQogICAgICAgICByZXR1cm47Ci0gICAgfQog CisgICAgUGFpbnRJbmZvIGNoaWxkUGFpbnRJbmZvKHBhaW50SW5mbyk7CisgICAgY2hpbGRQYWlu dEluZm8uY29udGV4dC0+c2F2ZSgpOworICAgIGFwcGx5VHJhbnNmb3JtVG9QYWludEluZm8oY2hp bGRQYWludEluZm8sIG1fbG9jYWxUcmFuc2Zvcm0pOwogICAgIFNWR1Jlc291cmNlRmlsdGVyKiBm aWx0ZXIgPSAwOwogCiAgICAgaWYgKGNoaWxkUGFpbnRJbmZvLnBoYXNlID09IFBhaW50UGhhc2VG b3JlZ3JvdW5kKSB7Cg==