22150 2008-11-09 13:23:01 -0800 Color.cpp gives wrong colors for alpha < 1 2010-03-10 19:26:43 -0800 1 1 1 Unclassified WebKit New Bugs 528+ (Nightly build) PC All RESOLVED FIXED P2 Normal --- 1 krit cshu ariya.hidayat cshu darin mitz oliver oldest_to_newest 98097 0 krit 2008-11-09 13:23:01 -0800 Color.cpp gives wrong colors, for an alpha lower than 1. This is a problem for canvas in Qt and Cairo. Cg seems _not_ to be affected. It's difficult to messure it or make a real LayoutTest for it, since Qt and Cairo lacks a fully working getImageData support. But it is difficult to implement getImageData as long as we can't test it because of wrong colors. I'll upload a simple test. If you make a screenshot and compare the color values (rgb) with the color-values of ff or opera, you can see that they differs from ff and opera. ff and opera give the right values. I checked if it is a problem with canvas. But canvas has the right colors. The colors in canvas are parsed by CSSParser::parseColor and are given to the GraphicsContext by CanvasStyle.cpp. But the GraphicsContext gets the wrong colors. I can't say if it is a problem of Color.cpp or CSSParser.cpp. See also: https://bugs.webkit.org/show_bug.cgi?id=21575#c5 And again: Cg seems not to be affected. 98098 1 25005 krit 2008-11-09 13:25:28 -0800 Created attachment 25005 Short test case in canvas 98099 2 25006 krit 2008-11-09 13:30:27 -0800 Created attachment 25006 Screenshot left ff, right the result of qt/gtkwebkit 98144 3 25014 krit 2008-11-10 05:08:54 -0800 Created attachment 25014 Short test case in CSS A div box with the same background color like the canvas example but this time via CSS. This shows that it is not canvas' fault. The difference is minimal (that means it's not visible to the naked eye), but has a big effect on getImageData for canvas. (Chrome has the same wrong behavior like Qt and Cairo) 104409 4 oliver 2009-01-05 01:39:32 -0800 I'll need to look at Color's implementation, but i would guess it's multiply/dividing at the wrong time 119441 5 ariya.hidayat 2009-04-30 02:43:05 -0700 Seems to be precision problem. For example, if alpha is 1.0/255.0, when the last few digits are stripped (i.e. the value is truncated), multiplying back to 255 does not give 1.0, but rather 0.99998. Static casting it back to integer gives 0, instead of 1. 119442 6 29905 ariya.hidayat 2009-04-30 02:46:06 -0700 Created attachment 29905 potential fix for the problem The problem with this patch is the rounding. Now (0.5/255.0) as the alpha value will be treated as 1 (of max 255). Need to find out what CSS specs say about this, or if this is just our implementation limitation (of using RGBA32). 119443 7 29906 krit 2009-04-30 03:17:15 -0700 Created attachment 29906 Result of getImageData with ariya's patch This is the new result after ariya's patch. Many more tests passes now. There are still some tests, that fail (beside outside canvas bounds). But the alpha channel is correct on all tests now. 119628 8 krit 2009-05-01 13:33:20 -0700 Without going into to much details, the following tests seem to fail on MacOS X 10.5 with this patch: fast/block/basic/fieldset-stretch-to-legend.html fast/block/float/intruding-painted-twice.html fast/block/float/relative-painted-twice.html fast/borders/fieldsetBorderRadius.html fast/canvas/canvas-alphaImageData-behavior.html fast/css/percentage-non-integer.html fast/css/shadow-multiple.html fast/js/date-DST-time-cusps.html fast/js/date-big-setdate.html fast/overflow/float-in-relpositioned.html fast/overflow/overflow-focus-ring.html fast/repaint/fixed.html fast/table/border-collapsing/equal-precedence-resolution.html fast/text/shadow-no-blur.html svg/custom/pointer-events-path.svg I did not run the tests without the patch, but the buildbots were green. 119824 9 ariya.hidayat 2009-05-04 04:51:41 -0700 Some of the failure make sense, because now the color alpha is rounded. Some of them are certainly invalid (nothing to do with color), e.g. date-big-setdate.html. 120826 10 ariya.hidayat 2009-05-12 04:51:46 -0700 I change my position on this. As long as Qt is internally using premultiplied ARGB32 for doing the alpha-blending, there will be discrepancies with respect to the precise ARGB32 color that will be retrieved from the pixmap. As agreed by Simon Hausmann, for the Qt port we will just create a specialized test result to take this into account. 120836 11 staikos 2009-05-12 06:59:06 -0700 Seems to me that this is wrong. I think the test should just be skipped, not putting a different result into place. 120873 12 krit 2009-05-12 09:14:32 -0700 Qt and Cairo pass many tests, that are not related to a change of the alpha channel. This is the only test for getImageData and it's important to know if someone breaks something for getImageData. That means skipping is the wrong solution. 120891 13 ariya.hidayat 2009-05-12 10:15:37 -0700 On IRC, George suggested to split the test into two parts: one which does not use alpha channel and one which uses it. This is good and something we should consider as well. 195737 14 49939 cshu 2010-03-03 13:40:27 -0800 Created attachment 49939 fix patch I came up with the same solution as Ariya before I found the existing bug with Ariya's fix. To avoid changing the behavior of rounding up alpha value, I used a small tolerance number. I tested on mac and the patch didn't break anything. 195755 15 49939 darin 2010-03-03 14:12:51 -0800 Comment on attachment 49939 fix patch > - colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255); > + colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255 + 0.0000000001); That doesn't look right to me. I believe the correct fix for this is: colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * nextafter(256.0, 0.0)); See code using nextafter in Color.cpp. We'd have to do some testing to be sure this does not cause problems. 195772 16 cshu 2010-03-03 14:55:22 -0800 (In reply to comment #15) > (From update of attachment 49939 [details]) > > - colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255); > > + colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255 + 0.0000000001); > > That doesn't look right to me. I believe the correct fix for this is: > > colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * > nextafter(256.0, 0.0)); > > See code using nextafter in Color.cpp. We'd have to do some testing to be sure > this does not cause problems. Thanks for showing me the right way to do this fix! I will submit the new patch after running all layout tests. 196083 17 cshu 2010-03-04 07:51:58 -0800 (In reply to comment #16) > (In reply to comment #15) > > (From update of attachment 49939 [details] [details]) > > > - colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255); > > > + colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * 255 + 0.0000000001); > > > > That doesn't look right to me. I believe the correct fix for this is: > > > > colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * > > nextafter(256.0, 0.0)); > > > > See code using nextafter in Color.cpp. We'd have to do some testing to be sure > > this does not cause problems. Hi, Darin, your algorithm seems problematic. Note that we do NOT want to round up alpha in general but only when the precision plays a role. For example, let alpha = 0.995, the result is expected to be floor(0.995 * 255) = floor(253.72) = 253. While in your algorithm, the result is floor(0.995 * nextafter(256,0)) = floor(~254.72499) = 254. Another example, let alpha = 0.0078431372549, which is 2.0/255.0, but lost some precision due to float and string representation, the result is expected to be floor(2.0/255.0*255.0) = 2, but actual result(current algorithm) is floor(0.0078431372549 * 255.0) = floor(1.9999999999995002) = 1. 196193 18 49939 eric 2010-03-04 13:17:22 -0800 Comment on attachment 49939 fix patch This change at minimum needs an explanitory comment in the code. Otherwise future contributers will be as confused as I am now. 196480 19 50100 cshu 2010-03-05 09:24:14 -0800 Created attachment 50100 fix patch 2 I used an "intelligent" number in this patch. I also put in some comments as suggested by Eric. 198174 20 darin 2010-03-10 09:28:03 -0800 (In reply to comment #17) > Hi, Darin, your algorithm seems problematic. Note that we do NOT want to round > up alpha in general but only when the precision plays a role. > For example, let alpha = 0.995, the result is expected to be > floor(0.995 * 255) = floor(253.72) = 253. While in your algorithm, the result > is > floor(0.995 * nextafter(256,0)) = floor(~254.72499) = 254. I don't understand why 0.995 should be 253. Maybe to match other web browsers? 253 / 255 is 0.992156862745098 245 / 255 is 0.996078431372549 The value 0.995 is closer to 0.996 than to 0.992 and so 0.995 should give the result 254, not 253. This is rounding to nearest, not rounding up. > Another example, let alpha = 0.0078431372549, which is 2.0/255.0, but lost some > precision due to float and string representation, the result is expected to be > floor(2.0/255.0*255.0) = 2, but actual result(current algorithm) is > floor(0.0078431372549 * 255.0) = floor(1.9999999999995002) = 1. Yes, we agree on the correct result for 0.0078431372549. It must be 2. 198187 21 cshu 2010-03-10 10:05:58 -0800 > > For example, let alpha = 0.995, the result is expected to be > > floor(0.995 * 255) = floor(253.72) = 253. While in your algorithm, the result > > is > > floor(0.995 * nextafter(256,0)) = floor(~254.72499) = 254. > > I don't understand why 0.995 should be 253. Maybe to match other web browsers? I also think 254 makes more sense (round to nearest). I just didn't want to change the current behavior and break an existing test case: fast/canvas/canvas-alphaImageData-behavior.html. 253 was expected over there. I also verified it on firefox and it gives 254. Looks like we should update the expected results for all the failed tests. 198193 22 oliver 2010-03-10 10:24:14 -0800 (In reply to comment #21) > > > For example, let alpha = 0.995, the result is expected to be > > > floor(0.995 * 255) = floor(253.72) = 253. While in your algorithm, the result > > > is > > > floor(0.995 * nextafter(256,0)) = floor(~254.72499) = 254. > > > > I don't understand why 0.995 should be 253. Maybe to match other web browsers? > > I also think 254 makes more sense (round to nearest). I just didn't want to > change the current behavior and break an existing test case: > fast/canvas/canvas-alphaImageData-behavior.html. 253 was expected over there. > I also verified it on firefox and it gives 254. Looks like we should update the > expected results for all the failed tests. That seems like the correct approach to me, I can't recall what rounding mode i was targeting back when i wrote this but i don't recall deliberately choosing one model over another. 198232 23 50100 darin 2010-03-10 11:06:54 -0800 Comment on attachment 50100 fix patch 2 Lets do a fix more like the one I suggested as per our current discussion in the bug. 198261 24 cshu 2010-03-10 12:07:15 -0800 (In reply to comment #23) > (From update of attachment 50100 [details]) > Lets do a fix more like the one I suggested as per our current discussion in > the bug. working on that... 198272 25 50426 cshu 2010-03-10 12:49:31 -0800 Created attachment 50426 fix patch 3 198280 26 50426 cshu 2010-03-10 13:00:07 -0800 Comment on attachment 50426 fix patch 3 the styles checks, etc. do not seem to start if I don't put r?. but please do not approve the patch before all lights show green. 198381 27 50426 darin 2010-03-10 15:24:52 -0800 Comment on attachment 50426 fix patch 3 Will this change pixel results for any tests? If so, we need those to be updated too. Assuming it does not, r=me 198386 28 50426 darin 2010-03-10 15:29:41 -0800 Comment on attachment 50426 fix patch 3 > + // Round the float alpha to its nearest integer value. That comment is not what the code does. After all, the float is between 0 and 1, so the nearest integer is either 0 or 1. What the code does is convert the floating point number to an integer in the range [0, 256), with an equal distribution across all 256 values. > + colorArray[3] = static_cast<int>(max(0.0, min(1.0, v->fValue)) * (nextafter(256.0, 0.0))); There's an extra set of parentheses around the nextafter function that are not needed. 198388 29 darin 2010-03-10 15:30:01 -0800 But the r=me still stands. Those are just a couple of minor criticisms/ways this could be improved. 198459 30 cshu 2010-03-10 19:26:43 -0800 I updated the code with Darin's latest comments on patch 3 and manually committed the patch. Committed revision 55827. Set the bug to closed. 25005 2008-11-09 13:25:28 -0800 2008-11-09 13:25:28 -0800 Short test case in canvas test.html text/html 411 krit PGh0bWw+DQogIDxoZWFkPg0KICAgIDxzY3JpcHQgdHlwZT0iYXBwbGljYXRpb24veC1qYXZhc2Ny aXB0Ij4NCiAgICBmdW5jdGlvbiBkcmF3KCkgew0KICAgICAgdmFyIGN0eCA9IGRvY3VtZW50Lmdl dEVsZW1lbnRCeUlkKCdjYW52YXMnKS5nZXRDb250ZXh0KCcyZCcpOw0KICAgICAgICAgIGN0eC5m aWxsU3R5bGUgPSAncmdiYSgwLCAyNTUsIDAsIDAuMjUpJzsNCiAgICAgICAgICBjdHguZmlsbFJl Y3QoMCwwLDE1MCwxNTApOw0KICAgIH0NCiAgICA8L3NjcmlwdD4NCiAgPC9oZWFkPg0KICA8Ym9k eSBvbmxvYWQ9ImRyYXcoKTsiPg0KICAgIDxkaXY+DQogICAgICA8Y2FudmFzIGlkPSJjYW52YXMi IHdpZHRoPSIxNTAiIGhlaWdodD0iMTUwIj48L2NhbnZhcz4NCiAgICA8L2Rpdj4NCiAgPC9ib2R5 Pg0KPC9odG1sPg0K 25006 2008-11-09 13:30:27 -0800 2008-11-09 13:30:27 -0800 Screenshot test.png image/png 753 krit iVBORw0KGgoAAAANSUhEUgAAAVwAAACiCAYAAAAN+XXxAAAAAXNSR0IArs4c6QAAAAZiS0dEAP8A /wD/oL2nkwAAAAlwSFlzAAALEwAACxMBAJqcGAAAAAd0SU1FB9gLCRUbLlds7bcAAAAZdEVYdENv bW1lbnQAQ3JlYXRlZCB3aXRoIEdJTVBXgQ4XAAACTElEQVR42u3YsW3CUBiF0fseUFCwgEdgAyp7 CCqmJVXWobGMbCOnSp9EEAE6p38ubvHpl0vTNMvpdMrxeMx+v89utws8i2VZcrvdMs9zpmnKOI65 Xq8ZhiF936dt21wuF0PxEmqS1FpTa7UGgOACvElwV6uV4AL814VbSrEGgAsX4PWtvy/cnwb3nLPV uLsu3cO+/ZkPA3N3h7R/u3C32202m40FAR7IfwQAwQUQXAAEF0BwAQTXBACCCyC4AAgugOACCC4A ggsguAAILoDgAgguAIILILgACC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIILILgA CC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIILILgACC6A4AIILgCCCyC4AAgugOAC CC4AggsguAAILoDgAgguAIILILgACC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIIL ILgACC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIIL8BbBLaVYA8CFCyC4AAgugOAC CC4AggsguAAILoDgAgguAIILILgACC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIIL ILgACC6A4AIILgCCCyC4AAgugOACCC4AggsguAAILoDgAgguAIILILgACC6A4AIILgCCCyC4AAgu gOACCC4AggsguAAILoDgAgguAPe2/u2DLp3VeCmHtEbAhQsguAAILoDgAiC4AM/gC4JRJbSa4Ynt AAAAAElFTkSuQmCC 25014 2008-11-10 05:08:54 -0800 2008-11-10 05:08:54 -0800 Short test case in CSS test2.html text/html 126 krit PGh0bWw+DQogIDxib2R5Pg0KICAgIDxkaXYgc3R5bGU9ImJhY2tncm91bmQ6IHJnYmEoMCwgMjU1 LCAwLCAwLjI1KTsgd2lkdGg6IDE1MHB4OyBoZWlnaHQ6IDE1MHB4Ij48L2Rpdj4NCiAgPC9ib2R5 Pg0KPC9odG1sPg0K 29905 2009-04-30 02:46:06 -0700 2010-03-03 13:40:27 -0800 potential fix for the problem webkit-rgba-parser.patch text/plain 554 ariya.hidayat ZGlmZiAtLWdpdCBhL1dlYkNvcmUvY3NzL0NTU1BhcnNlci5jcHAgYi9XZWJDb3JlL2Nzcy9DU1NQ YXJzZXIuY3BwCmluZGV4IGQyYThkZmYuLmE4Y2JkYTQgMTAwNjQ0Ci0tLSBhL1dlYkNvcmUvY3Nz L0NTU1BhcnNlci5jcHAKKysrIGIvV2ViQ29yZS9jc3MvQ1NTUGFyc2VyLmNwcApAQCAtMzMxMiw3 ICszMzEyLDcgQEAgYm9vbCBDU1NQYXJzZXI6OnBhcnNlQ29sb3JQYXJhbWV0ZXJzKENTU1BhcnNl clZhbHVlKiB2YWx1ZSwgaW50KiBjb2xvckFycmF5LCBib28KICAgICAgICAgdiA9IGFyZ3MtPm5l eHQoKTsKICAgICAgICAgaWYgKCF2YWxpZFVuaXQodiwgRk51bWJlciwgdHJ1ZSkpCiAgICAgICAg ICAgICByZXR1cm4gZmFsc2U7Ci0gICAgICAgIGNvbG9yQXJyYXlbM10gPSBzdGF0aWNfY2FzdDxp bnQ+KG1heCgwLjAsIG1pbigxLjAsIHYtPmZWYWx1ZSkpICogMjU1KTsKKyAgICAgICAgY29sb3JB cnJheVszXSA9IG1heCgwLCBtaW4oMjU1LCBzdGF0aWNfY2FzdDxpbnQ+KHJvdW5kKHYtPmZWYWx1 ZSAqIDI1NSkpKSk7CiAgICAgfQogICAgIHJldHVybiB0cnVlOwogfQo= 29906 2009-04-30 03:17:15 -0700 2010-03-03 13:42:45 -0800 Result of getImageData with ariya's patch result text/plain 6448 krit VGhpcyB0ZXN0IGVuc3VyZXMgdGhhdCBnZXRJbWFnZURhdGEgd29ya3MgY29ycmVjdGx5LgpQQVNT OiBwaXhlbCBhdCAoMCwwKSB3YXMgWzAsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDQsMCkgd2Fz IFswLDExLDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAoOCwwKSB3YXMgWzAsMjEsMCwyNTVdClBBU1M6 IHBpeGVsIGF0ICgxMiwwKSB3YXMgWzAsMzIsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICgxNiwwKSB3 YXMgWzAsNDMsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICgyMCwwKSB3YXMgWzAsNTMsMCwyNTVdClBB U1M6IHBpeGVsIGF0ICgyNCwwKSB3YXMgWzAsNjQsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICgyOCww KSB3YXMgWzAsNzQsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICgzMiwwKSB3YXMgWzAsODUsMCwyNTVd ClBBU1M6IHBpeGVsIGF0ICgzNiwwKSB3YXMgWzAsOTYsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICg0 MCwwKSB3YXMgWzAsMTA2LDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAoNDQsMCkgd2FzIFswLDExNyww LDI1NV0KUEFTUzogcGl4ZWwgYXQgKDQ4LDApIHdhcyBbMCwxMjgsMCwyNTVdClBBU1M6IHBpeGVs IGF0ICg1MiwwKSB3YXMgWzAsMTM4LDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAoNTYsMCkgd2FzIFsw LDE0OSwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDYwLDApIHdhcyBbMCwxNTksMCwyNTVdClBBU1M6 IHBpeGVsIGF0ICg2NCwwKSB3YXMgWzAsMTcwLDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAoNjgsMCkg d2FzIFswLDE4MSwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDcyLDApIHdhcyBbMCwxOTEsMCwyNTVd ClBBU1M6IHBpeGVsIGF0ICg3NiwwKSB3YXMgWzAsMjAyLDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAo ODAsMCkgd2FzIFswLDIxMywwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDg0LDApIHdhcyBbMCwyMjMs MCwyNTVdClBBU1M6IHBpeGVsIGF0ICg4OCwwKSB3YXMgWzAsMjM0LDAsMjU1XQpQQVNTOiBwaXhl bCBhdCAoOTIsMCkgd2FzIFswLDI0NCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDk2LDApIHdhcyBb MCwyNTUsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICg1LDUpIHdhcyBbNjQsMTI4LDE5MSwyNTVdClBB U1M6IENvcnJlY3RseSByZXRyaWV2ZWQgZXZlcnkgcGl4ZWwgaW4gYSByb3cKRkFJTDogRGlkIG5v dCBnZXQgY29ycmVjdCBkYXRhIGZvciBjb250ZW50IG91dHNpZGUgY2FudmFzIGJvdW5kczogMjU1 LDI1NSwyNTUsMjU1LDI1NSwyNTUsMjU1LDI1NSwwLDAsMSwwLDM1LDAsMCwwClBBU1M6IHBpeGVs IGF0ICg1MCw1MCkgd2FzIFswLDAsMCwwXQpQQVNTOiBwaXhlbCBhdCAoNTAsNTQpIHdhcyBbMCww LDAsMV0KUEFTUzogcGl4ZWwgYXQgKDUwLDU4KSB3YXMgWzAsMCwwLDJdClBBU1M6IHBpeGVsIGF0 ICg1MCw2Mikgd2FzIFswLDAsMCwxNl0KUEFTUzogcGl4ZWwgYXQgKDUwLDY2KSB3YXMgWzAsMCww LDMyXQpQQVNTOiBwaXhlbCBhdCAoNTAsNzApIHdhcyBbMCwwLDAsNjRdClBBU1M6IHBpeGVsIGF0 ICg1MCw3NCkgd2FzIFswLDAsMCw5Nl0KUEFTUzogcGl4ZWwgYXQgKDUwLDc4KSB3YXMgWzAsMCww LDEyOF0KUEFTUzogcGl4ZWwgYXQgKDUwLDgyKSB3YXMgWzAsMCwwLDE5Ml0KUEFTUzogcGl4ZWwg YXQgKDUwLDg2KSB3YXMgWzAsMCwwLDI1NF0KUEFTUzogcGl4ZWwgYXQgKDUwLDkwKSB3YXMgWzAs MCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDU0LDUwKSB3YXMgWzAsMCwwLDBdClBBU1M6IHBpeGVs IGF0ICg1NCw1NCkgd2FzIFswLDAsMCwxXQpQQVNTOiBwaXhlbCBhdCAoNTQsNTgpIHdhcyBbMCww LDAsMl0KUEFTUzogcGl4ZWwgYXQgKDU0LDYyKSB3YXMgWzAsMCwwLDE2XQpQQVNTOiBwaXhlbCBh dCAoNTQsNjYpIHdhcyBbMCwwLDAsMzJdClBBU1M6IHBpeGVsIGF0ICg1NCw3MCkgd2FzIFswLDAs MCw2NF0KUEFTUzogcGl4ZWwgYXQgKDU0LDc0KSB3YXMgWzAsMCwwLDk2XQpGQUlMOiBwaXhlbCBh dCAoNTQsNzgpIHdhcyBbMCwwLDAsMTI4XSwgZXhwZWN0ZWQgWzEsMCwwLDEyOF0KRkFJTDogcGl4 ZWwgYXQgKDU0LDgyKSB3YXMgWzAsMCwwLDE5Ml0sIGV4cGVjdGVkIFsxLDAsMCwxOTJdCkZBSUw6 IHBpeGVsIGF0ICg1NCw4Nikgd2FzIFswLDAsMCwyNTRdLCBleHBlY3RlZCBbMSwwLDAsMjU0XQpQ QVNTOiBwaXhlbCBhdCAoNTQsOTApIHdhcyBbMSwwLDAsMjU1XQpQQVNTOiBwaXhlbCBhdCAoNTgs NTApIHdhcyBbMCwwLDAsMF0KUEFTUzogcGl4ZWwgYXQgKDU4LDU0KSB3YXMgWzAsMCwwLDFdClBB U1M6IHBpeGVsIGF0ICg1OCw1OCkgd2FzIFswLDAsMCwyXQpQQVNTOiBwaXhlbCBhdCAoNTgsNjIp IHdhcyBbMCwwLDAsMTZdClBBU1M6IHBpeGVsIGF0ICg1OCw2Nikgd2FzIFswLDAsMCwzMl0KRkFJ TDogcGl4ZWwgYXQgKDU4LDcwKSB3YXMgWzAsMCwwLDY0XSwgZXhwZWN0ZWQgWzMsMCwwLDY0XQpG QUlMOiBwaXhlbCBhdCAoNTgsNzQpIHdhcyBbMCwwLDAsOTZdLCBleHBlY3RlZCBbMiwwLDAsOTZd ClBBU1M6IHBpeGVsIGF0ICg1OCw3OCkgd2FzIFsxLDAsMCwxMjhdCkZBSUw6IHBpeGVsIGF0ICg1 OCw4Mikgd2FzIFsxLDAsMCwxOTJdLCBleHBlY3RlZCBbMiwwLDAsMTkyXQpGQUlMOiBwaXhlbCBh dCAoNTgsODYpIHdhcyBbMSwwLDAsMjU0XSwgZXhwZWN0ZWQgWzIsMCwwLDI1NF0KUEFTUzogcGl4 ZWwgYXQgKDU4LDkwKSB3YXMgWzIsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDYyLDUwKSB3YXMg WzAsMCwwLDBdClBBU1M6IHBpeGVsIGF0ICg2Miw1NCkgd2FzIFswLDAsMCwxXQpQQVNTOiBwaXhl bCBhdCAoNjIsNTgpIHdhcyBbMCwwLDAsMl0KUEFTUzogcGl4ZWwgYXQgKDYyLDYyKSB3YXMgWzE1 LDAsMCwxNl0KUEFTUzogcGl4ZWwgYXQgKDYyLDY2KSB3YXMgWzE1LDAsMCwzMl0KUEFTUzogcGl4 ZWwgYXQgKDYyLDcwKSB3YXMgWzE1LDAsMCw2NF0KUEFTUzogcGl4ZWwgYXQgKDYyLDc0KSB3YXMg WzE1LDAsMCw5Nl0KUEFTUzogcGl4ZWwgYXQgKDYyLDc4KSB3YXMgWzE1LDAsMCwxMjhdClBBU1M6 IHBpeGVsIGF0ICg2Miw4Mikgd2FzIFsxNSwwLDAsMTkyXQpGQUlMOiBwaXhlbCBhdCAoNjIsODYp IHdhcyBbMTUsMCwwLDI1NF0sIGV4cGVjdGVkIFsxNiwwLDAsMjU0XQpQQVNTOiBwaXhlbCBhdCAo NjIsOTApIHdhcyBbMTYsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDY2LDUwKSB3YXMgWzAsMCww LDBdClBBU1M6IHBpeGVsIGF0ICg2Niw1NCkgd2FzIFswLDAsMCwxXQpQQVNTOiBwaXhlbCBhdCAo NjYsNTgpIHdhcyBbMCwwLDAsMl0KUEFTUzogcGl4ZWwgYXQgKDY2LDYyKSB3YXMgWzMxLDAsMCwx Nl0KUEFTUzogcGl4ZWwgYXQgKDY2LDY2KSB3YXMgWzMxLDAsMCwzMl0KUEFTUzogcGl4ZWwgYXQg KDY2LDcwKSB3YXMgWzMxLDAsMCw2NF0KUEFTUzogcGl4ZWwgYXQgKDY2LDc0KSB3YXMgWzMxLDAs MCw5Nl0KUEFTUzogcGl4ZWwgYXQgKDY2LDc4KSB3YXMgWzMxLDAsMCwxMjhdClBBU1M6IHBpeGVs IGF0ICg2Niw4Mikgd2FzIFszMSwwLDAsMTkyXQpGQUlMOiBwaXhlbCBhdCAoNjYsODYpIHdhcyBb MzEsMCwwLDI1NF0sIGV4cGVjdGVkIFszMiwwLDAsMjU0XQpQQVNTOiBwaXhlbCBhdCAoNjYsOTAp IHdhcyBbMzIsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDcwLDUwKSB3YXMgWzAsMCwwLDBdClBB U1M6IHBpeGVsIGF0ICg3MCw1NCkgd2FzIFswLDAsMCwxXQpGQUlMOiBwaXhlbCBhdCAoNzAsNTgp IHdhcyBbMCwwLDAsMl0sIGV4cGVjdGVkIFsxMjcsMCwwLDJdClBBU1M6IHBpeGVsIGF0ICg3MCw2 Mikgd2FzIFs2MywwLDAsMTZdClBBU1M6IHBpeGVsIGF0ICg3MCw2Nikgd2FzIFs2MywwLDAsMzJd ClBBU1M6IHBpeGVsIGF0ICg3MCw3MCkgd2FzIFs2MywwLDAsNjRdClBBU1M6IHBpeGVsIGF0ICg3 MCw3NCkgd2FzIFs2MywwLDAsOTZdClBBU1M6IHBpeGVsIGF0ICg3MCw3OCkgd2FzIFs2MywwLDAs MTI4XQpQQVNTOiBwaXhlbCBhdCAoNzAsODIpIHdhcyBbNjMsMCwwLDE5Ml0KRkFJTDogcGl4ZWwg YXQgKDcwLDg2KSB3YXMgWzYzLDAsMCwyNTRdLCBleHBlY3RlZCBbNjQsMCwwLDI1NF0KUEFTUzog cGl4ZWwgYXQgKDcwLDkwKSB3YXMgWzY0LDAsMCwyNTVdClBBU1M6IHBpeGVsIGF0ICg3NCw1MCkg d2FzIFswLDAsMCwwXQpQQVNTOiBwaXhlbCBhdCAoNzQsNTQpIHdhcyBbMCwwLDAsMV0KRkFJTDog cGl4ZWwgYXQgKDc0LDU4KSB3YXMgWzAsMCwwLDJdLCBleHBlY3RlZCBbMTI3LDAsMCwyXQpQQVNT OiBwaXhlbCBhdCAoNzQsNjIpIHdhcyBbOTUsMCwwLDE2XQpQQVNTOiBwaXhlbCBhdCAoNzQsNjYp IHdhcyBbOTUsMCwwLDMyXQpQQVNTOiBwaXhlbCBhdCAoNzQsNzApIHdhcyBbOTUsMCwwLDY0XQpQ QVNTOiBwaXhlbCBhdCAoNzQsNzQpIHdhcyBbOTUsMCwwLDk2XQpQQVNTOiBwaXhlbCBhdCAoNzQs NzgpIHdhcyBbOTUsMCwwLDEyOF0KUEFTUzogcGl4ZWwgYXQgKDc0LDgyKSB3YXMgWzk1LDAsMCwx OTJdCkZBSUw6IHBpeGVsIGF0ICg3NCw4Nikgd2FzIFs5NSwwLDAsMjU0XSwgZXhwZWN0ZWQgWzk2 LDAsMCwyNTRdClBBU1M6IHBpeGVsIGF0ICg3NCw5MCkgd2FzIFs5NiwwLDAsMjU1XQpQQVNTOiBw aXhlbCBhdCAoNzgsNTApIHdhcyBbMCwwLDAsMF0KRkFJTDogcGl4ZWwgYXQgKDc4LDU0KSB3YXMg WzAsMCwwLDFdLCBleHBlY3RlZCBbMjU1LDAsMCwxXQpQQVNTOiBwaXhlbCBhdCAoNzgsNTgpIHdh cyBbMTI3LDAsMCwyXQpQQVNTOiBwaXhlbCBhdCAoNzgsNjIpIHdhcyBbMTI3LDAsMCwxNl0KUEFT UzogcGl4ZWwgYXQgKDc4LDY2KSB3YXMgWzEyNywwLDAsMzJdClBBU1M6IHBpeGVsIGF0ICg3OCw3 MCkgd2FzIFsxMjcsMCwwLDY0XQpQQVNTOiBwaXhlbCBhdCAoNzgsNzQpIHdhcyBbMTI3LDAsMCw5 Nl0KUEFTUzogcGl4ZWwgYXQgKDc4LDc4KSB3YXMgWzEyNywwLDAsMTI4XQpQQVNTOiBwaXhlbCBh dCAoNzgsODIpIHdhcyBbMTI3LDAsMCwxOTJdClBBU1M6IHBpeGVsIGF0ICg3OCw4Nikgd2FzIFsx MjcsMCwwLDI1NF0KUEFTUzogcGl4ZWwgYXQgKDc4LDkwKSB3YXMgWzEyOCwwLDAsMjU1XQpQQVNT OiBwaXhlbCBhdCAoODIsNTApIHdhcyBbMCwwLDAsMF0KRkFJTDogcGl4ZWwgYXQgKDgyLDU0KSB3 YXMgWzAsMCwwLDFdLCBleHBlY3RlZCBbMjU1LDAsMCwxXQpGQUlMOiBwaXhlbCBhdCAoODIsNTgp IHdhcyBbMTI3LDAsMCwyXSwgZXhwZWN0ZWQgWzI1NSwwLDAsMl0KUEFTUzogcGl4ZWwgYXQgKDgy LDYyKSB3YXMgWzE5MSwwLDAsMTZdClBBU1M6IHBpeGVsIGF0ICg4Miw2Nikgd2FzIFsxOTEsMCww LDMyXQpQQVNTOiBwaXhlbCBhdCAoODIsNzApIHdhcyBbMTkxLDAsMCw2NF0KUEFTUzogcGl4ZWwg YXQgKDgyLDc0KSB3YXMgWzE5MSwwLDAsOTZdClBBU1M6IHBpeGVsIGF0ICg4Miw3OCkgd2FzIFsx OTEsMCwwLDEyOF0KUEFTUzogcGl4ZWwgYXQgKDgyLDgyKSB3YXMgWzE5MiwwLDAsMTkyXQpQQVNT OiBwaXhlbCBhdCAoODIsODYpIHdhcyBbMTkxLDAsMCwyNTRdClBBU1M6IHBpeGVsIGF0ICg4Miw5 MCkgd2FzIFsxOTIsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDg2LDUwKSB3YXMgWzAsMCwwLDBd CkZBSUw6IHBpeGVsIGF0ICg4Niw1NCkgd2FzIFswLDAsMCwxXSwgZXhwZWN0ZWQgWzI1NSwwLDAs MV0KRkFJTDogcGl4ZWwgYXQgKDg2LDU4KSB3YXMgWzEyNywwLDAsMl0sIGV4cGVjdGVkIFsyNTUs MCwwLDJdCkZBSUw6IHBpeGVsIGF0ICg4Niw2Mikgd2FzIFsyMzksMCwwLDE2XSwgZXhwZWN0ZWQg WzI1NSwwLDAsMTZdCkZBSUw6IHBpeGVsIGF0ICg4Niw2Nikgd2FzIFsyNDcsMCwwLDMyXSwgZXhw ZWN0ZWQgWzI1NSwwLDAsMzJdCkZBSUw6IHBpeGVsIGF0ICg4Niw3MCkgd2FzIFsyNTEsMCwwLDY0 XSwgZXhwZWN0ZWQgWzI1NSwwLDAsNjRdCkZBSUw6IHBpeGVsIGF0ICg4Niw3NCkgd2FzIFsyNTIs MCwwLDk2XSwgZXhwZWN0ZWQgWzI1NSwwLDAsOTZdClBBU1M6IHBpeGVsIGF0ICg4Niw3OCkgd2Fz IFsyNTMsMCwwLDEyOF0KUEFTUzogcGl4ZWwgYXQgKDg2LDgyKSB3YXMgWzI1MywwLDAsMTkyXQpQ QVNTOiBwaXhlbCBhdCAoODYsODYpIHdhcyBbMjUzLDAsMCwyNTRdClBBU1M6IHBpeGVsIGF0ICg4 Niw5MCkgd2FzIFsyNTQsMCwwLDI1NV0KUEFTUzogcGl4ZWwgYXQgKDkwLDUwKSB3YXMgWzAsMCww LDBdClBBU1M6IHBpeGVsIGF0ICg5MCw1NCkgd2FzIFsyNTUsMCwwLDFdClBBU1M6IHBpeGVsIGF0 ICg5MCw1OCkgd2FzIFsyNTUsMCwwLDJdClBBU1M6IHBpeGVsIGF0ICg5MCw2Mikgd2FzIFsyNTUs MCwwLDE2XQpQQVNTOiBwaXhlbCBhdCAoOTAsNjYpIHdhcyBbMjU1LDAsMCwzMl0KUEFTUzogcGl4 ZWwgYXQgKDkwLDcwKSB3YXMgWzI1NSwwLDAsNjRdClBBU1M6IHBpeGVsIGF0ICg5MCw3NCkgd2Fz IFsyNTUsMCwwLDk2XQpQQVNTOiBwaXhlbCBhdCAoOTAsNzgpIHdhcyBbMjU1LDAsMCwxMjhdClBB U1M6IHBpeGVsIGF0ICg5MCw4Mikgd2FzIFsyNTUsMCwwLDE5Ml0KUEFTUzogcGl4ZWwgYXQgKDkw LDg2KSB3YXMgWzI1NSwwLDAsMjU0XQpQQVNTOiBwaXhlbCBhdCAoOTAsOTApIHdhcyBbMjU1LDAs MCwyNTVdCg== 49939 2010-03-03 13:40:27 -0800 2010-03-05 09:24:14 -0800 fix patch bug22150_1.patch text/plain 2131 cshu SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA1NTQ3MikKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTMgQEAKKzIwMTAtMDMtMDMgIFNodSBDaGFuZyAgPENoYW5nLlNodUBub2tpYS5j b20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgRml4 IHRoZSBwcmVjaXNpb24gcHJvYmxlbSBzaG93aW5nIHVwIG9uIGNlcnRhaW4gcGxhdGZvcm1zLCBz dWNoIGFzIFF0IExpbnV4LgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1 Zy5jZ2k/aWQ9MjIxNTAKKworICAgICAgICAqIGNzcy9DU1NQYXJzZXIuY3BwOgorICAgICAgICAo V2ViQ29yZTo6Q1NTUGFyc2VyOjpwYXJzZUNvbG9yUGFyYW1ldGVycyk6CisKIDIwMTAtMDMtMDMg IEFsZXhhbmRlciBQYXZsb3YgIDxhcGF2bG92QGNocm9taXVtLm9yZz4KIAogICAgICAgICBSZXZp ZXdlZCBieSBQYXZlbCBGZWxkbWFuLgpJbmRleDogV2ViQ29yZS9jc3MvQ1NTUGFyc2VyLmNwcAo9 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09Ci0tLSBXZWJDb3JlL2Nzcy9DU1NQYXJzZXIuY3BwCShyZXZpc2lvbiA1NTQyMykK KysrIFdlYkNvcmUvY3NzL0NTU1BhcnNlci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTM1NzYsNyAr MzU3Niw3IEBAIGJvb2wgQ1NTUGFyc2VyOjpwYXJzZUNvbG9yUGFyYW1ldGVycyhDU1MKICAgICAg ICAgdiA9IGFyZ3MtPm5leHQoKTsKICAgICAgICAgaWYgKCF2YWxpZFVuaXQodiwgRk51bWJlciwg dHJ1ZSkpCiAgICAgICAgICAgICByZXR1cm4gZmFsc2U7Ci0gICAgICAgIGNvbG9yQXJyYXlbM10g PSBzdGF0aWNfY2FzdDxpbnQ+KG1heCgwLjAsIG1pbigxLjAsIHYtPmZWYWx1ZSkpICogMjU1KTsK KyAgICAgICAgY29sb3JBcnJheVszXSA9IHN0YXRpY19jYXN0PGludD4obWF4KDAuMCwgbWluKDEu MCwgdi0+ZlZhbHVlKSkgKiAyNTUgKyAwLjAwMDAwMDAwMDEpOwogICAgIH0KICAgICByZXR1cm4g dHJ1ZTsKIH0KSW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBMYXlv dXRUZXN0cy9DaGFuZ2VMb2cJKHJldmlzaW9uIDU1NDcyKQorKysgTGF5b3V0VGVzdHMvQ2hhbmdl TG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTIgQEAKKzIwMTAtMDMtMDMgIFNodSBDaGFu ZyAgPENoYW5nLlNodUBub2tpYS5jb20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChP T1BTISkuCisKKyAgICAgICAgW1F0XSBSZW1vdmUgdGVzdCBjYXNlIGZyb20gU2tpcHBlZC4KKyAg ICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTIyMTUwCisKKyAg ICAgICAgKiBwbGF0Zm9ybS9xdC9Ta2lwcGVkOgorCiAyMDEwLTAzLTAzICBBbGV4ZXkgUHJvc2t1 cnlha292ICA8YXBAYXBwbGUuY29tPgogCiAgICAgICAgIE5vdCByZXZpZXdlZC4KSW5kZXg6IExh eW91dFRlc3RzL3BsYXRmb3JtL3F0L1NraXBwZWQKPT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMv cGxhdGZvcm0vcXQvU2tpcHBlZAkocmV2aXNpb24gNTU0MjMpCisrKyBMYXlvdXRUZXN0cy9wbGF0 Zm9ybS9xdC9Ta2lwcGVkCSh3b3JraW5nIGNvcHkpCkBAIC0xNDEsNyArMTQxLDYgQEAgZmFzdC93 b3JrZXJzL3NoYXJlZC13b3JrZXItZnJhbWUtbGlmZWN5YwogIyAtLS0tLS0tIHdvcmtzIG9uIDY0 IGJpdCBidXQgbm90IG9uIDMyIGJpdAogc3ZnL2N1c3RvbS9wYXRoLWdldFRvdGFsTGVuZ3RoLnN2 Zwogc3ZnL2N1c3RvbS9nbHlwaC1zZXR0aW5nLWQtYXR0cmlidXRlLnN2ZwotZmFzdC9jYW52YXMv Y2FudmFzLWdldEltYWdlRGF0YS5odG1sCiAKICMgLS0tLS0tLSBjcmFzaGVzCiBmYXN0L3RleHQv ZmluZC1oaWRkZW4tdGV4dC5odG1sCg== 50100 2010-03-05 09:24:14 -0800 2010-03-10 12:49:31 -0800 fix patch 2 bug22150_2.patch text/plain 2326 cshu SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA1NTU4NikKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTMgQEAKKzIwMTAtMDMtMDUgIFNodSBDaGFuZyAgPENoYW5nLlNodUBub2tpYS5j b20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgRml4 IHRoZSBwcmVjaXNpb24gcHJvYmxlbSBzaG93aW5nIHVwIG9uIGNlcnRhaW4gcGxhdGZvcm1zLCBz dWNoIGFzIFF0IExpbnV4LgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1 Zy5jZ2k/aWQ9MjIxNTAKKworICAgICAgICAqIGNzcy9DU1NQYXJzZXIuY3BwOgorICAgICAgICAo V2ViQ29yZTo6Q1NTUGFyc2VyOjpwYXJzZUNvbG9yUGFyYW1ldGVycyk6CisKIDIwMTAtMDMtMDQg IE5hdGUgQ2hhcGluICA8amFwaGV0QGNocm9taXVtLm9yZz4KIAogICAgICAgICBSZXZpZXdlZCBi eSBEaW1pdHJpIEdsYXprb3YuCkluZGV4OiBXZWJDb3JlL2Nzcy9DU1NQYXJzZXIuY3BwCj09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT0KLS0tIFdlYkNvcmUvY3NzL0NTU1BhcnNlci5jcHAJKHJldmlzaW9uIDU1NTg2KQorKysg V2ViQ29yZS9jc3MvQ1NTUGFyc2VyLmNwcAkod29ya2luZyBjb3B5KQpAQCAtMzU5MCw3ICszNTkw LDkgQEAgYm9vbCBDU1NQYXJzZXI6OnBhcnNlQ29sb3JQYXJhbWV0ZXJzKENTUwogICAgICAgICB2 ID0gYXJncy0+bmV4dCgpOwogICAgICAgICBpZiAoIXZhbGlkVW5pdCh2LCBGTnVtYmVyLCB0cnVl KSkKICAgICAgICAgICAgIHJldHVybiBmYWxzZTsKLSAgICAgICAgY29sb3JBcnJheVszXSA9IHN0 YXRpY19jYXN0PGludD4obWF4KDAuMCwgbWluKDEuMCwgdi0+ZlZhbHVlKSkgKiAyNTUpOworICAg ICAgICAvLyBDb21wZW5zYXRlIHRoZSBwcmVjaXNpb24gZXJyb3IgaW50cm9kdWNlZCBieSBmbG9h dGluZyBwb2ludGVyIG51bWJlci4KKyAgICAgICAgLy8gRm9yIGV4YW1wbGUsIHRoaXMgbWFrZXMg c3VyZSAoaW50KSgoMi4wLzI1NS4wKSoyNTUpIHJlc3VsdHMgMiBpbnN0ZWFkIG9mIDEuCisgICAg ICAgIGNvbG9yQXJyYXlbM10gPSBzdGF0aWNfY2FzdDxpbnQ+KG1heCgwLjAsIG1pbigxLjAsIHYt PmZWYWx1ZSkgKyAxLjAgLSBuZXh0YWZ0ZXIoMS4wLCAwLjApKSAqIDI1NSk7CiAgICAgfQogICAg IHJldHVybiB0cnVlOwogfQpJbmRleDogTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCj09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0K LS0tIExheW91dFRlc3RzL0NoYW5nZUxvZwkocmV2aXNpb24gNTU1ODYpCisrKyBMYXlvdXRUZXN0 cy9DaGFuZ2VMb2cJKHdvcmtpbmcgY29weSkKQEAgLTEsMyArMSwxMiBAQAorMjAxMC0wMy0wNSAg U2h1IENoYW5nICA8Q2hhbmcuU2h1QG5va2lhLmNvbT4KKworICAgICAgICBSZXZpZXdlZCBieSBO T0JPRFkgKE9PUFMhKS4KKworICAgICAgICBbUXRdIFJlbW92ZSB0ZXN0IGNhc2UgZnJvbSBTa2lw cGVkLgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93X2J1Zy5jZ2k/aWQ9MjIx NTAKKworICAgICAgICAqIHBsYXRmb3JtL3F0L1NraXBwZWQ6CisKIDIwMTAtMDMtMDMgIEFudG9u aW8gR29tZXMgIDx0b25pa2l0b29Ad2Via2l0Lm9yZz4KIAogICAgICAgICBSZXZpZXdlZCBieSBI b2xnZXIgRnJleXRoZXIKSW5kZXg6IExheW91dFRlc3RzL3BsYXRmb3JtL3F0L1NraXBwZWQKPT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PQotLS0gTGF5b3V0VGVzdHMvcGxhdGZvcm0vcXQvU2tpcHBlZAkocmV2aXNpb24gNTU1 ODYpCisrKyBMYXlvdXRUZXN0cy9wbGF0Zm9ybS9xdC9Ta2lwcGVkCSh3b3JraW5nIGNvcHkpCkBA IC0xNDIsNyArMTQyLDYgQEAgZmFzdC93b3JrZXJzL3NoYXJlZC13b3JrZXItZnJhbWUtbGlmZWN5 YwogIyAtLS0tLS0tIHdvcmtzIG9uIDY0IGJpdCBidXQgbm90IG9uIDMyIGJpdAogc3ZnL2N1c3Rv bS9wYXRoLWdldFRvdGFsTGVuZ3RoLnN2Zwogc3ZnL2N1c3RvbS9nbHlwaC1zZXR0aW5nLWQtYXR0 cmlidXRlLnN2ZwotZmFzdC9jYW52YXMvY2FudmFzLWdldEltYWdlRGF0YS5odG1sCiAKICMgLS0t LS0tLSBjcmFzaGVzCiBmYXN0L3RleHQvZmluZC1oaWRkZW4tdGV4dC5odG1sCg== 50426 2010-03-10 12:49:31 -0800 2010-03-10 15:29:41 -0800 fix patch 3 bug22150_3.patch text/plain 3982 cshu SW5kZXg6IFdlYkNvcmUvQ2hhbmdlTG9nCj09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT0KLS0tIFdlYkNvcmUvQ2hhbmdlTG9n CShyZXZpc2lvbiA1NTc5NykKKysrIFdlYkNvcmUvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBA IC0xLDMgKzEsMTQgQEAKKzIwMTAtMDMtMTAgIENoYW5nIFNodSAgPGNoYW5nLnNodUBub2tpYS5j b20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgV2hp bGUgY2FsY3VsYXRpbmcgYWxwaGEgY2hhbm5lbCwgcm91bmQgdGhlIGZsb2F0aW5nIHBvaW50IHZh bHVlIHRvIGl0cyBuZWFyZXN0IHBvc3NpYmxlCisgICAgICAgIGludGVnZXIgdmFsdWUgaW5zdGVh ZCBvZiByb3VuZGluZyBkb3duLgorICAgICAgICBodHRwczovL2J1Z3Mud2Via2l0Lm9yZy9zaG93 X2J1Zy5jZ2k/aWQ9MjIxNTAKKworICAgICAgICAqIGNzcy9DU1NQYXJzZXIuY3BwOgorICAgICAg ICAoV2ViQ29yZTo6Q1NTUGFyc2VyOjpwYXJzZUNvbG9yUGFyYW1ldGVycyk6CisKIDIwMTAtMDMt MTAgIEFudG9uaW8gR29tZXMgIDx0b25pa2l0b29Ad2Via2l0Lm9yZz4KIAogICAgICAgICBSZXZp ZXdlZCBieSBEYXZlIEh5YXR0LgpJbmRleDogV2ViQ29yZS9jc3MvQ1NTUGFyc2VyLmNwcAo9PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09Ci0tLSBXZWJDb3JlL2Nzcy9DU1NQYXJzZXIuY3BwCShyZXZpc2lvbiA1NTc3NikKKysr IFdlYkNvcmUvY3NzL0NTU1BhcnNlci5jcHAJKHdvcmtpbmcgY29weSkKQEAgLTM1OTEsNyArMzU5 MSw4IEBAIGJvb2wgQ1NTUGFyc2VyOjpwYXJzZUNvbG9yUGFyYW1ldGVycyhDU1MKICAgICAgICAg diA9IGFyZ3MtPm5leHQoKTsKICAgICAgICAgaWYgKCF2YWxpZFVuaXQodiwgRk51bWJlciwgdHJ1 ZSkpCiAgICAgICAgICAgICByZXR1cm4gZmFsc2U7Ci0gICAgICAgIGNvbG9yQXJyYXlbM10gPSBz dGF0aWNfY2FzdDxpbnQ+KG1heCgwLjAsIG1pbigxLjAsIHYtPmZWYWx1ZSkpICogMjU1KTsKKyAg ICAgICAgLy8gUm91bmQgdGhlIGZsb2F0IGFscGhhIHRvIGl0cyBuZWFyZXN0IGludGVnZXIgdmFs dWUuCisgICAgICAgIGNvbG9yQXJyYXlbM10gPSBzdGF0aWNfY2FzdDxpbnQ+KG1heCgwLjAsIG1p bigxLjAsIHYtPmZWYWx1ZSkpICogKG5leHRhZnRlcigyNTYuMCwgMC4wKSkpOwogICAgIH0KICAg ICByZXR1cm4gdHJ1ZTsKIH0KSW5kZXg6IExheW91dFRlc3RzL0NoYW5nZUxvZwo9PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 Ci0tLSBMYXlvdXRUZXN0cy9DaGFuZ2VMb2cJKHJldmlzaW9uIDU1Nzk3KQorKysgTGF5b3V0VGVz dHMvQ2hhbmdlTG9nCSh3b3JraW5nIGNvcHkpCkBAIC0xLDMgKzEsMTUgQEAKKzIwMTAtMDMtMTAg IENoYW5nIFNodSAgPGNoYW5nLnNodUBub2tpYS5jb20+CisKKyAgICAgICAgUmV2aWV3ZWQgYnkg Tk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgV2hpbGUgY2FsY3VsYXRpbmcgYWxwaGEgY2hhbm5l bCwgcm91bmQgdGhlIGZsb2F0aW5nIHBvaW50IHZhbHVlIHRvIGl0cyBuZWFyZXN0IHBvc3NpYmxl CisgICAgICAgIGludGVnZXIgdmFsdWUgaW5zdGVhZCBvZiByb3VuZGluZyBkb3duLiBVcGRhdGVk IHRlc3QgcmVzdWx0cyBiYXNlZCBvbiB0aGlzIG5ldyBiZWhhdmlvci4KKyAgICAgICAgaHR0cHM6 Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTIyMTUwCisKKyAgICAgICAgKiBmYXN0 L2NhbnZhcy9jYW52YXMtYWxwaGFJbWFnZURhdGEtYmVoYXZpb3ItZXhwZWN0ZWQudHh0OgorICAg ICAgICAqIGZhc3QvY2FudmFzL2NhbnZhcy1hbHBoYUltYWdlRGF0YS1iZWhhdmlvci5qczoKKyAg ICAgICAgKiBwbGF0Zm9ybS9xdC9Ta2lwcGVkOgorCiAyMDEwLTAzLTEwICBBbnRvbmlvIEdvbWVz ICA8dG9uaWtpdG9vQHdlYmtpdC5vcmc+CiAKICAgICAgICAgUmV2aWV3ZWQgYnkgRGF2ZSBIeWF0 dC4KSW5kZXg6IExheW91dFRlc3RzL2Zhc3QvY2FudmFzL2NhbnZhcy1hbHBoYUltYWdlRGF0YS1i ZWhhdmlvci1leHBlY3RlZC50eHQKPT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMvZmFzdC9jYW52 YXMvY2FudmFzLWFscGhhSW1hZ2VEYXRhLWJlaGF2aW9yLWV4cGVjdGVkLnR4dAkocmV2aXNpb24g NTU3NzYpCisrKyBMYXlvdXRUZXN0cy9mYXN0L2NhbnZhcy9jYW52YXMtYWxwaGFJbWFnZURhdGEt YmVoYXZpb3ItZXhwZWN0ZWQudHh0CSh3b3JraW5nIGNvcHkpCkBAIC02LDExICs2LDExIEBAIE9u IHN1Y2Nlc3MsIHlvdSB3aWxsIHNlZSBhIHNlcmllcyBvZiAiUEEKIFBBU1MgaW1nZGF0YVs0XSBp cyAwCiBQQVNTIGltZ2RhdGFbNV0gaXMgMjU1CiBQQVNTIGltZ2RhdGFbNl0gaXMgMAotUEFTUyBp bWdkYXRhWzddIGlzIDI1MworUEFTUyBpbWdkYXRhWzddIGlzIDI1NAogUEFTUyBpbWdkYXRhWzRd IGlzIDAKIFBBU1MgaW1nZGF0YVs1XSBpcyAyNTUKIFBBU1MgaW1nZGF0YVs2XSBpcyAwCi1QQVNT IGltZ2RhdGFbN10gaXMgMjUzCitQQVNTIGltZ2RhdGFbN10gaXMgMjU0CiBQQVNTIHN1Y2Nlc3Nm dWxseVBhcnNlZCBpcyB0cnVlCiAKIFRFU1QgQ09NUExFVEUKSW5kZXg6IExheW91dFRlc3RzL2Zh c3QvY2FudmFzL2NhbnZhcy1hbHBoYUltYWdlRGF0YS1iZWhhdmlvci5qcwo9PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0t LSBMYXlvdXRUZXN0cy9mYXN0L2NhbnZhcy9jYW52YXMtYWxwaGFJbWFnZURhdGEtYmVoYXZpb3Iu anMJKHJldmlzaW9uIDU1Nzc2KQorKysgTGF5b3V0VGVzdHMvZmFzdC9jYW52YXMvY2FudmFzLWFs cGhhSW1hZ2VEYXRhLWJlaGF2aW9yLmpzCSh3b3JraW5nIGNvcHkpCkBAIC05LDcgKzksNyBAQCB2 YXIgaW1nZGF0YSA9IGltYWdlRGF0YS5kYXRhOwogc2hvdWxkQmUoImltZ2RhdGFbNF0iLCAiMCIp Owogc2hvdWxkQmUoImltZ2RhdGFbNV0iLCAiMjU1Iik7CiBzaG91bGRCZSgiaW1nZGF0YVs2XSIs ICIwIik7Ci1zaG91bGRCZSgiaW1nZGF0YVs3XSIsICIyNTMiKTsKK3Nob3VsZEJlKCJpbWdkYXRh WzddIiwgIjI1NCIpOwogCiBjdHgucHV0SW1hZ2VEYXRhKGltYWdlRGF0YSwgMCwgMCk7CiAKQEAg LTE4LDYgKzE4LDYgQEAgaW1nZGF0YSA9IGN0eC5nZXRJbWFnZURhdGEoMCwgMCwgMjAwLCAyMAog c2hvdWxkQmUoImltZ2RhdGFbNF0iLCAiMCIpOwogc2hvdWxkQmUoImltZ2RhdGFbNV0iLCAiMjU1 Iik7CiBzaG91bGRCZSgiaW1nZGF0YVs2XSIsICIwIik7Ci1zaG91bGRCZSgiaW1nZGF0YVs3XSIs ICIyNTMiKTsKK3Nob3VsZEJlKCJpbWdkYXRhWzddIiwgIjI1NCIpOwogCiB2YXIgc3VjY2Vzc2Z1 bGx5UGFyc2VkID0gdHJ1ZTsKSW5kZXg6IExheW91dFRlc3RzL3BsYXRmb3JtL3F0L1NraXBwZWQK PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PQotLS0gTGF5b3V0VGVzdHMvcGxhdGZvcm0vcXQvU2tpcHBlZAkocmV2aXNpb24g NTU3OTgpCisrKyBMYXlvdXRUZXN0cy9wbGF0Zm9ybS9xdC9Ta2lwcGVkCSh3b3JraW5nIGNvcHkp CkBAIC0xNDMsNyArMTQzLDYgQEAgZmFzdC93b3JrZXJzL3NoYXJlZC13b3JrZXItZnJhbWUtbGlm ZWN5YwogIyAtLS0tLS0tIHdvcmtzIG9uIDY0IGJpdCBidXQgbm90IG9uIDMyIGJpdAogc3ZnL2N1 c3RvbS9wYXRoLWdldFRvdGFsTGVuZ3RoLnN2Zwogc3ZnL2N1c3RvbS9nbHlwaC1zZXR0aW5nLWQt YXR0cmlidXRlLnN2ZwotZmFzdC9jYW52YXMvY2FudmFzLWdldEltYWdlRGF0YS5odG1sCiAKICMg LS0tLS0tLSBjcmFzaGVzCiBmYXN0L3RleHQvZmluZC1oaWRkZW4tdGV4dC5odG1sCg==