51151 2010-12-15 16:53:44 -0800 NSScrollView-based scrollbars and ScrollBarTheme-based scrollbars don't look the same 2011-01-21 12:29:42 -0800 1 1 1 Unclassified WebKit Layout and Rendering 528+ (Nightly build) Mac OS X 10.5 RESOLVED FIXED P2 Normal --- 51406 51507 51728 1 dglazkov webkit-unassigned abarth andersca avi eric hyatt jam jamesr mihaip sam webkit.review.bot oldest_to_newest 322938 0 dglazkov 2010-12-15 16:53:44 -0800 It's an odd thing, but it appears that painting scrollbars using HIThemeDrawTrack and just relying on NSScrollView to make them doesn't produce the same result. Here's an example: Exhibit A: http://svn.webkit.org/repository/webkit/trunk/LayoutTests/platform/mac/fast/block/float/008-expected.png (uses NSScrollView to paint scrollbars) Exhibit B: http://build.webkit.org/results/Chromium%20Mac%20Release%20(Tests)/r74045%20(3464)/fast/block/float/008-actual.png (uses ScrollBarTheme) Exhibit C: http://build.webkit.org/results/Chromium%20Mac%20Release%20(Tests)/r74045%20(3464)/fast/block/float/008-diff.png (the diff between the two) I spent a day trying to hack ScrollbarThemeChromiumMac.mm into submission, just trying to see what the problem is. Turns out there are lots of funny little issues: 1) The proportions of the thumb are different for each path. Not sure why, but it requires tweaking SliderTrackInfo::viewsize to get them to match. The amount seems to be some proportional value, because it varies depending on the value of max/value in ThemeTrackDrawInfo. 2) The wavey Aqua texture on NSScrollView is off by exactly -11px compared to hand-drawn scrollbars. Again, not sure why. So I wrote an extra step to shift the thumb down, paint into a buffer, then paint back onto the track shifted upwards by the same amount. 3) Even after that, the caps on the thumb don't match! For some reason, they are also drawn differently. To summarize, I think it might be worth pursuing other options, such as: 1) convert Mac port to hand-drawn scrollbars and rebaseline the tests; 2) cry into a pillow. Unless someone has better ideas? :) 322941 1 dglazkov 2010-12-15 16:55:59 -0800 BTW, I started where Avi left off a while back. 322970 2 avi 2010-12-15 17:24:47 -0800 For those who are just joining us, I last dealt with this about a year ago when re-enabling Mac pixel tests for Chromium. http://www.mail-archive.com/chromium-dev@googlegroups.com/msg08115.html If we're OK with having different baselines for WebKit and Chromium, my conclusions still stand. It seems like you're trying to reconcile them, and that's going to be painful. Right now, for the sake of a Cocoa hierarchy, the top level scrollbars (and only those) are drawn with an NSScrollView. You're asking if we can ditch those in WebKit, and that's an answer I can't provide. Are you proposing using ScrollbarThemeMac for only tests and not Safari? (I suppose that's what you mean by "hand-drawn scrollbars".) 322971 3 avi 2010-12-15 17:26:51 -0800 BTW, the conclusion that I came to was that it simply wasn't worth the hassle to try to make an HIThemeDrawTrack scrollbar draw exactly like an NSScroller. They are two entirely different code paths and playing games is only going to leave someone in tears. And Mac OS X doesn't cry. 322972 4 jamesr 2010-12-15 17:29:08 -0800 In the long term, nobody will use NSScrollView for scrollbars on mac - Chromium already does not and WebKit2 will not either. Where are the WebKit2 pixel baselines expected to go? 322977 5 dglazkov 2010-12-15 17:34:15 -0800 (In reply to comment #4) > In the long term, nobody will use NSScrollView for scrollbars on mac - Chromium already does not and WebKit2 will not either. Where are the WebKit2 pixel baselines expected to go? Adding fearless WebKit2 peeps. 323204 6 76774 dglazkov 2010-12-16 09:01:16 -0800 Created attachment 76774 The closest I could come to matching I got pretty close :) 323498 7 mihaip 2010-12-16 16:18:22 -0800 H(In reply to comment #3) > BTW, the conclusion that I came to was that it simply wasn't worth the hassle to try to make an HIThemeDrawTrack scrollbar draw exactly like an NSScroller. They are two entirely different code paths and playing games is only going to leave someone in tears. Can we just use NSScroller directly (instantiate one, use setKnobProportion, etc, and then have it draw itself)? 323503 8 avi 2010-12-16 16:24:06 -0800 (In reply to comment #7) > Can we just use NSScroller directly (instantiate one, use setKnobProportion, etc, and then have it draw itself)? I suppose we could. But if we're dropping NSScrollViews altogether and moving towards only ScrollbarThemeMac ones, I don't see the advantage of dropping the current implementation of ScrollbarThemeMac for a new one based on Cocoa widgets. 323530 9 mihaip 2010-12-16 17:21:35 -0800 (In reply to comment #8) > I suppose we could. But if we're dropping NSScrollViews altogether and moving towards only ScrollbarThemeMac ones, I don't see the advantage of dropping the current implementation of ScrollbarThemeMac for a new one based on Cocoa widgets. Will HIThemeDrawTrack be around indefinitely? It's part of the "legacy" Appearance Manager API. 323549 10 avi 2010-12-16 17:56:40 -0800 (In reply to comment #9) > Will HIThemeDrawTrack be around indefinitely? It's part of the "legacy" Appearance Manager API. HIThemeDrawTrack isn't Appearance Manager, it's HITheme. When Carbon was killed for 64-bits, people at WWDC asked explicitly about HITheme and using it to draw appearance primitives, and Apple said that it would stick around. For now, I'm not worried about its future. 324060 11 76930 mihaip 2010-12-17 17:48:48 -0800 Created attachment 76930 Patch 324061 12 mihaip 2010-12-17 17:52:02 -0800 The attached patch seems to produce identical results to the native NSScrollView scrollbars. It has a bunch of FIXMEs, but if this seems like a reasonable approach, I can look into getting it landed. The biggest hack it in is the use of FakeActiveWindow and overriding the private _hasActiveControls NSWindow method to control (in)active appearance. Is there any way to improve on that? 324623 13 mihaip 2010-12-20 16:09:51 -0800 Avi and I chatted about this, and he feels like the right thing to do is to ge the Safari/Mac port (or at least its DRT) off of NSScrollView for the root scrollable area, so that it matches the other ports. However, from talking to James, I realized that that means a mass rebaseline (in platform/mac and platform/mac-leopard) of all layout tests that have scrollbars in them. The whole point of Dimitri's and my exploration in this area is to get Chromium/Mac to match Safari/Mac so that we can use the existing baselines. If we're going to be doing a mass rebaseline anyway, we might as well just maintain platform/chromium-mac baselines with our different scrollbar rendering (since we're already down that path -- we started with ~400 scrollbar-related failures in platform/chromium/test_expectations.txt and are now down to 229). Avi, how about I limit the scope of this patch to only use NSScroller when running under the Chromium DRT _and_ we're rendering the top-level scrollbar? That way any brittleness will be isolated. I'm also open to any suggestions that you might have to lessen that (e.g. can we get at the real NSWindow that we're in, instead of needing FakeActiveWindow?) 324711 14 avi 2010-12-20 19:19:55 -0800 (In reply to comment #13) > Avi, how about I limit the scope of this patch to only use NSScroller when running under the Chromium DRT _and_ we're rendering the top-level scrollbar? That way any brittleness will be isolated. Yes, that would likely work, though I'm not thrilled at the code flow (if we're in DRT and we're at the top level then run this entirely different code path). BTW, how do you plan to pipe in that state so that the scrollbar code would know? > I'm also open to any suggestions that you might have to lessen that (e.g. can we get at the real NSWindow that we're in, instead of needing FakeActiveWindow?) Um, the window that the renderer's output is drawn into is in a different process. No, you may not get at it. 324726 15 mihaip 2010-12-20 19:37:33 -0800 (In reply to comment #14) > (In reply to comment #13) > > Avi, how about I limit the scope of this patch to only use NSScroller when running under the Chromium DRT _and_ we're rendering the top-level scrollbar? That way any brittleness will be isolated. > > Yes, that would likely work, though I'm not thrilled at the code flow (if we're in DRT and we're at the top level then run this entirely different code path). BTW, how do you plan to pipe in that state so that the scrollbar code would know? I'd use a similar mechanism to the Windows port, which does something similar. It uses DRT-specific widget rendering code to get consistency between Win7, Vista and XP (see comments in http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/chromium/WebThemeEngineDRT.h#L31). Specifically: ScrollbarThemeChromiumWin::paintTrackPiece (http://trac.webkit.org/browser/trunk/WebCore/platform/chromium/ScrollbarThemeChromiumWin.cpp#L107) calls: ChromiumBridge::paintScrollbarTrack (http://trac.webkit.org/browser/trunk/WebKit/chromium/src/ChromiumBridge.cpp#L742) which calls: WebThemeEngine:: paintScrollbarTrack (defined at http://trac.webkit.org/browser/trunk/WebKit/chromium/public/win/WebThemeEngine.h#L67) WebThemeEngine:: paintScrollbarTrack has multiple implementations: http://www.google.com/codesearch?hl=en&vert=chromium&lr=&q=engine.*::paintscrollbartrack&sbtn=Search The one in WebThemeEngineDRT::paintScrollbarTrack is pulled in via a .gyp condition: http://trac.webkit.org/browser/trunk/Tools/DumpRenderTree/DumpRenderTree.gypi#L61 Linux has its own WebThemeEngine definition/implementation (added pretty recently: http://trac.webkit.org/changeset/69311), Mac will join the club. 324734 16 avi 2010-12-20 19:46:39 -0800 OK. I have no doubts that you're capable of doing it. I only have opinions left. 1. Yes, we can have two different scrollbar drawing implementations. It may solve a real problem. That doesn't mean it doesn't give me a bad feeling and scream "backward-looking". My gut says that fixing the top-level scrollbar != other scrollbar problem for Safari and rebaselining everything is a real solution and that this is just a hack. 2. You're doing WebThemeEngine for the Mac? Wasn't jam working on that? There is no way to draw the separate pieces of a scrollbar on the Mac (arrows, track, etc) with HITheme. Almost certainly not with the public API of NSScroller either. Avi 324739 17 mihaip 2010-12-20 19:57:57 -0800 (In reply to comment #16) > 1. Yes, we can have two different scrollbar drawing implementations. It may solve a real problem. That doesn't mean it doesn't give me a bad feeling and scream "backward-looking". My gut says that fixing the top-level scrollbar != other scrollbar problem for Safari and rebaselining everything is a real solution and that this is just a hack. Yes, this is a hack, but it seems preferable to either not looking at the results of hundreds of tests or maintaining parallel baselines for them. > 2. You're doing WebThemeEngine for the Mac? Wasn't jam working on that? Not sure, adding him to the bug. > There is no way to draw the separate pieces of a scrollbar on the Mac (arrows, track, etc) with HITheme. Almost certainly not with the public API of NSScroller either. The only part that's actually different is the thumb, so that's the only part we'd need to move into WebThemeEngineMac. We already draw that separately with HITheme: http://trac.webkit.org/browser/trunk/WebCore/platform/chromium/ScrollbarThemeChromiumMac.mm#L450 And NSScroller has drawKnob: http://developer.apple.com/library/mac/#documentation/Cocoa/Reference/ApplicationKit/Classes/NSScroller_Class/Reference/Reference.html#//apple_ref/occ/instm/NSScroller/drawKnob 324748 18 avi 2010-12-20 20:23:17 -0800 So kinda like we have today, where the Lin and Win WebThemeEngines are totally different? Can we put a TODO in this so that someday when we drop using the NSScrollView for the outer scroller we'll remember to drop these weird dual code paths? 328094 19 mihaip 2010-12-30 20:08:31 -0800 I've landed all the necessary code changes to fix this, all that remains is to flip the USE_WEB_THEME_ENGINE_TO_PAINT_THUMB #define in ScrollbarThemeChromiumMac. I'll try to do that in the next couple of days, while I'm still gardening, so that the inevitable mass rebaseline is less disruptive. 328573 20 mihaip 2011-01-03 08:16:58 -0800 Committed r74892: <http://trac.webkit.org/changeset/74892> 328575 21 mihaip 2011-01-03 08:19:15 -0800 Keeping this open until rebaselines are done. 328601 22 webkit.review.bot 2011-01-03 09:40:02 -0800 http://trac.webkit.org/changeset/74892 might have broken Leopard Intel Debug (Tests) 334774 23 mihaip 2011-01-14 17:29:21 -0800 Marking this as fixed since I'm done with rebaselines. Bug 52414 is still open about a small remaining mac vs. chromium-mac scrollbar difference. 334778 24 eric 2011-01-14 17:38:01 -0800 Did we ever file a radar about this API mismatch? 336170 25 mihaip 2011-01-18 15:57:20 -0800 (In reply to comment #24) > Did we ever file a radar about this API mismatch? Filed rdar://8881862 (http://openradar.appspot.com/8881862) 336173 26 eric 2011-01-18 16:00:42 -0800 Ideally you should mention this (or any other WebKit/Chromium) bugs in your Radar. I've poked relevant people @ Apple. 336183 27 mihaip 2011-01-18 16:13:34 -0800 (In reply to comment #26) > Ideally you should mention this (or any other WebKit/Chromium) bugs in your Radar. I've poked relevant people @ Apple. Done. 337674 28 mihaip 2011-01-20 15:52:16 -0800 Curiouser and curiouser: http://trac.webkit.org/changeset/76292 337683 29 eric 2011-01-20 16:05:07 -0800 You should probably mention that commit and the associated Radar in your Radar if you haven't already. 338006 30 avi 2011-01-21 04:37:42 -0800 (In reply to comment #28) > Curiouser and curiouser: http://trac.webkit.org/changeset/76292 Sigh. Because all we needed was more opaque SPI. :( 338262 31 eric 2011-01-21 12:29:42 -0800 I suspect the opaque API may be related to the need for the chromium hack and mihai's radar. We just need to get the right Apple people talkign to one another so we can remove our hack and they can remove their opaque API. :) 76774 2010-12-16 09:01:16 -0800 2010-12-16 09:01:16 -0800 The closest I could come to matching 008-diff.png image/png 17377 dglazkov iVBORw0KGgoAAAANSUhEUgAAAyAAAAJYCAYAAACadoJwAAAgAElEQVR4nOzdv87t33bf9fcYc37X 3ggrxA5S6OJIacgFENGHiFwEDRIXgSiOXZCaC6CmoojEJVBGyBEtKAWSC6eIEvv4OM+afwbFXM9e m8b27/vEshLeL2lpHa2zx09P+9EYY46oqt9DkiRJkv6S/uk//aeMMfid3/kd/uzP/owxBhHB9+/f eTweXNfF4/H48b9ba6y1+LM/+zP6X/cfL0mSJOnfL713/s7f+Tv83b/7d8lMns8nv/71r/nNb35D ZvJ4PPj27duPEPIZQP70T//UACJJkiTpl2mtkZn/n7DxW7/1W/ybf/Nv+PWvf811XWQmvXd677TW iAhaawYQSZIkSb/M51hVRPz4wAkmn+NX375949u3bz/CyFqLb9++GUAkSZIk/XJV9eN7782f/Mmf 8Md//Mf03tl7//gAPwIKYACRJEmS9MtU1Y8A8vHxwR/90R/xJ3/yJ0QE3759Y4zB8/kEIDMNIJIk SZLu+wwVHx8f/It/8S/45//8n/Mv/+W/5F//63/Nr3/9a75//87f+lt/i7/39/4e/+Af/AP+9t/+ 27TWqCoDiCRJkqRfZu/NnJM/+IM/4L/6R//oV//Fn/Nvf+9Xv/r9f/gP/yF//+///b9cAInfj7/o n0iSJEn6D0j9qv78/7+KP/qjP+K//sf/+Fd/0X/r937/93/1P4zx+5nJ7/7u75L/zv5KSZIkSf+/ 8Qf/5//zl/63/+M/+Se/+mf/7J/xx3/8xwYQSZIkSb/M8/nk//i/rl9U84d/+If84R/+oQFEkiRJ 0i/zb//tv+V//Z//279w/Opnv/nNb/hX/+pfGUAkSZIk/TKf9z1+id/85jeMMQwgkiRJkn6Zzxsg v8SPuyD/rv8YSZIkSf9ha6394pqqIjMNIJIkSZL+6n1eQzeASJIkSforFxFkppfQJUmSJP1y8V/+ b7//3/93/zm//Z/9R+wKfv0B//v//eQ//e2/wW//9n/C3/ybyd/4LfiP/yZ8/Br4n/6bE0L+uv9w SZIkSf9+2muTwNhFVTEXjLmYc7IKFjAXPOf59xFhB0SSJEnSPZUJGTyA2YOrQW/B1ZMroF/w7QH7 +7vGDogkSZKkewJ6QBS0hIzi0eDKE0AuoBU8fkodBhBJkiRJt0RPIoPs54Wr63HRrkbrQXbIPF2Q ZgCRJEmS9GUBQZEBGUmjaBRJ0Rs8LmhxPp/cAZEkSZJ0TwQtgqLR8xU2WtIzaQ0C6A3GT3cL7YBI kiRJuqWiyJZknDGrFpAUGUUPuK6zI3I5giVJkiTpyzKJgNYgM+i90VvQe9CuEzaiQfw0gmUAkSRJ knRLAbugdp0g0oOrtfMiVp2OSO9nEf2TAUSSJEnSLVeD2lAEaxdzbmadULI5AWVOGM93jQFEkiRJ 0j2rCDZrLapgzc2cizlhjnMBvYqTRF58BUuSJEnSLXMCJMRrHCuSuYu5YXF+GwM+5rvGACJJkiTp lqrFrs2qgh3MOdm7zujVgLFgbM57vC+OYEmSJEm6JTasgqrgYyVrw9ow63yvdboga71rDCCSJEmS 7tnnBawxir0XY26eYzGfk1X82AX5cAldkiRJ0lc14LkKAuaCRbBpjEqeC8aE56tL8skAIkmSJOmW vTYJjF1UFXPBmIs5TwdkcYLJ0yV0SZIkSV9VmZDBA5g9uBr0Flw9ueIcIPz2gP39XWMHRJIkSdI9 AT0ginP9PIpHgytPALmAVvD4KXUYQCRJkiTdEj2JDLKfd3avx0W7Gq0H2SHzdEGaAUSSJEnSlwUE RQZkJI2iUSRFb/C4oMX5fHIHRJIkSdI9EbQIikbPV9hoSc+ktXN/sDcY7V1iB0SSJEnSLRVFtiTj jFm1gKTIKHrAdZ0dkcsRLEmSJElflkkEtAaZQe+N3oLeg3adsBEN4qcRLAOIJEmSpFsK2HWuoUdA 68HV2nkRq05HpPeziP7JACJJkiTplqtBbSiCtYs5N7NOKNmcgDInjOe7xgAiSZIk6Z5VBJu1FlWw 5mbOxZwwx7mAXsVJIi++giVJkiTpljkBEuI1jhXJ3MXcsDi/jQEf811jAJEkSZJ0S9Vi12ZVwQ7m nOxdZ/RqwFgwNuc93hdHsCRJkiTdEhtWQVXwsZK1YW2Ydb7XOl2Qtd41BhBJkiRJ9+zzAtYYxd6L MTfPsZjPySp+7IJ8uIQuSZIk6asa8FwFAXPBItg0RiXPBWPC89Ul+WQAkSRJknTLXpsExi6qirlg zMWcpwOyOMHk6RK6JEmSpK+qTMjgAcweXA16C66eXHEOEH57wP7+rrEDIkmSJOmegB4Qxbl+HsWj wZUngFxAK3j8lDoMIJIkSZJuiZ5EBtnPO7vX46JdjdaD7JB5uiDNACJJkiTpywKCIgMykkbRKJKi N3hc0OJ8PrkDIkmSJOmeCFoERaPnK2y0pGfS2rk/2BuM9i6xAyJJkiTplooiW5JxxqxaQFJkFD3g us6OyOUIliRJkqQvyyQCWoPMoPdGb0HvQbtO2IgG8dMIlgFEkiRJ0i0F7DrX0COg9eBq7byIVacj 0vtZRP9kAJEkSZJ0y9WgNhTB2sWcm1knlGxOQJkTxvNdYwCRJEmSdM8qgs1aiypYczPnYk6Y41xA r+IkkRdfwZIkSZJ0y5wACfEax4pk7mJuWJzfxoCP+a4xgEiSJEm6pWqxa7OqYAdzTvauM3o1YCwY m/Me74sjWJIkSZJuiQ2roCr4WMnasDbMOt9rnS7IWu8aA4gkSZKke/Z5AWuMYu/FmJvnWMznZBU/ dkE+XEKXJEmS9FUNeK6CgLlgEWwao5LngjHh+eqSfDKASJIkSbplr00CYxdVxVww5mLO0wFZnGDy dAldkiRJ0ldVJmTwAGYPrga9BVdPrjgHCL89YH9/19gBkSRJknRPQA+I4lw/j+LR4MoTQC6gFTx+ Sh0GEEmSJEm3RE8ig+znnd3rcdGuRutBdsg8XZBmAJEkSZL0ZQFBkQEZSaNoFEnRGzwuaHE+n9wB kSRJknRPBC2CotHzFTZa0jNp7dwf7A1Ge5fYAZEkSZJ0S0WRLck4Y1YtICkyih5wXWdH5HIES5Ik SdKXZRIBrUFm0Hujt6D3oF0nbESD+GkEywAiSZIk6ZYCdp1r6BHQenC1dl7EqtMR6f0son8ygEiS JEm65WpQG4pg7WLOzawTSjYnoMwJ4/muMYBIkiRJumcVwWatRRWsuZlzMSfMcS6gV3GSyIuvYEmS JEm6ZU6AhHiNY0UydzE3LM5vY8DHfNcYQCRJkiTdUrXYtVlVsIM5J3vXGb0aMBaMzXmP98URLEmS JEm3xIZVUBV8rGRtWBtmne+1ThdkrXeNAUSSJEnSPfu8gDVGsfdizM1zLOZzsoofuyAfLqFLkiRJ +qoGPFdBwFywCDaNUclzwZjwfHVJPhlAJEmSJN2y1yaBsYuqYi4YczHn6YAsTjB5uoQuSZIk6asq EzJ4ALMHV4PegqsnV5wDhN8esL+/a+yASJIkSbonoAdEca6fR/FocOUJIBfQCh4/pQ4DiCRJkqRb oieRQfbzzu71uGhXo/UgO2SeLkgzgEiSJEn6soCgyICMpFE0iqToDR4XtDifT+6ASJIkSbonghZB 0ej5Chst6Zm0du4P9gajvUvsgEiSJEm6paLIlmScMasWkBQZRQ+4rrMjcjmCJUmSJOnLMomA1iAz 6L3RW9B70K4TNqJB/DSCZQCRJEmSdEsBu8419AhoPbhaOy9i1emI9H4W0T8ZQCRJkiTdcjWoDUWw djHnZtYJJZsTUOaE8XzXGEAkSZIk3bOKYLPWogrW3My5mBPmOBfQqzhJ5MVXsCRJkiTdMidAQrzG sSKZu5gbFue3MeBjvmsMIJIkSZJuqVrs2qwq2MGck73rjF4NGAvG5rzH++IIliRJkqRbYsMqqAo+ VrI2rA2zzvdapwuy1rvGACJJkiTpnn1ewBqj2Hsx5uY5FvM5WcWPXZAPl9AlSZIkfVUDnqsgYC5Y BJvGqOS5YEx4vroknwwgkiRJkm7Za5PA2EVVMReMuZjzdEAWJ5g8XUKXJEmS9FWVCRk8gNmDq0Fv wdWTK84Bwm8P2N/fNXZAJEmSJN0T0AOiONfPo3g0uPIEkAtoBY+fUocBRJIkSdIt0ZPIIPt5Z/d6 XLSr0XqQHTJPF6QZQCRJkiR9WUBQZEBG0igaRVL0Bo8LWpzPJ3dAJEmSJN0TQYugaPR8hY2W9Exa O/cHe4PR3iV2QCRJkiTdUlFkSzLOmFULSIqMogdc19kRuRzBkiRJkvRlmURAa5AZ9N7oLeg9aNcJ G9EgfhrBMoBIkiRJuqWAXecaegS0HlytnRex6nREej+L6J8MIJIkSZJuuRrUhiJYu5hzM+uEks0J KHPCeL5r/sIl9N/7vd/7K/uD9cv9qn711/0nSJL0H4b/Jf7ifyPpz7eKYDMXVDXW3My5mBPmOBfQ vxcnibz4CpYkSZKkW+YESIjXOFYkcxdzw+L8NgZ8zHeNAUSSJEnSLVWLXZtVBTuYc7J3ndGrAWPB 2Jz3eF/cAZEkSZJ0S2xYBVXBx0rWhrVh1vle63RB1nrXGEAkSZIk3bPPC1hjFHsvxtw8x2I+J6v4 sQvy8dMSugFEkiRJ0i0NeK6CgLlgEWwao5LngjHh+eqSfDKASJIkSbplr00CYxdVxVww5mLO0wFZ nGDydAldkiRJ0ldVJmTwAGYPrga9BVdPrjgHCL89YH9/19gBkSRJknRPQA+I4lw/j+LR4MoTQC6g FTx+Sh0GEEmSJEm3RE8ig+znnd3rcdGuRutBdsg8XZBmAJEkSZL0ZQFBkQEZSaNoFEnRGzwuaHE+ n9wBkSRJknRPBC2CotHzFTZa0jNp7dwf7A1Ge5fYAZEkSZJ0S0WRLck4Y1YtICkyih5wXWdH5HIE S5IkSdKXZRIBrUFm0Hujt6D3oF0nbESD+GkEywAiSZIk6ZYCdp1r6BHQenC1dl7EqtMR6f0son8y gEiSJEm65WpQG4pg7WLOzawTSjYnoMwJ4/muMYBIkiRJumcVwWatRRWsuZlzMSfMcS6gV3GSyIuv YEmSJEm6ZU6AhHiNY0UydzE3LM5vY8DHfNcYQCRJkiTdUrXYtVlVsIM5J3vXGb0aMBaMzXmP98UR LEmSJEm3xIZVUBV8rGRtWBtmne+1ThdkrXeNAUSSJEnSPfu8gDVGsfdizM1zLOZzsoofuyAfLqFL kiRJ+qoGPFdBwFywCDaNUclzwZjwfHVJPhlAJEmSJN2y1yaBsYuqYi4YczHn6YAsTjB5uoQuSZIk 6asqEzJ4ALMHV4PegqsnV5wDhN8esL+/a+yASJIkSbonoAdEca6fR/FocOUJIBfQCh4/pQ4DiCRJ kqRboieRQfbzzu71uGhXo/UgO2SeLkgzgEiSJEn6soCgyICMpFE0iqToDR4XtDifT+6ASJIkSbon ghZB0ej5Chst6Zm0du4P9gajvUvsgEiSJEm6paLIlmScMasWkBQZRQ+4rrMjcjmCJUmSJOnLMomA 1iAz6L3RW9B70K4TNqJB/DSCZQCRJEmSdEsBu8419AhoPbhaOy9i1emI9H4W0T8ZQCRJkiTdcjWo DUWwdjHnZtYJJZsTUOaE8XzXGEAkSZIk3bOKYLPWogrW3My5mBPmOBfQqzhJ5MVXsCRJkiTdMidA QrzGsSKZu5gbFue3MeBjvmsMIJIkSZJuqVrs2qwq2MGck73rjF4NGAvG5rzH++IIliRJkqRbYsMq qAo+VrI2rA2zzvdapwuy1rvGACJJkiTpnn1ewBqj2Hsx5uY5FvM5WcWPXZAPl9AlSZIkfVUDnqsg YC5YBJvGqOS5YEx4vroknwwgkiRJkm7Za5PA2EVVMReMuZjzdEAWJ5g8XUKXJEmS9FWVCRk8gNmD q0FvwdWTK84Bwm8P2N/fNXZAJEmSJN0T0AOiONfPo3g0uPIEkAtoBY+fUocBRJIkSdIt0ZPIIPt5 Z/d6XLSr0XqQHTJPF6QZQCRJkiR9WUBQZEBG0igaRVL0Bo8LWpzPJ3dAJEmSJN0TQYugaPR8hY2W 9ExaO/cHe4PR3iV2QCRJkiTdUlFkSzLOmFULSIqMogdc19kRuRzBkiRJkvRlmURAa5AZ9N7oLeg9 aNcJG9EgfhrBMoBIkiRJuqWAXecaegS0HlytnRex6nREej+L6J8MIJIkSZJuuRrUhiJYu5hzM+uE ks0JKHPCeL5rDCCSJEmS7llFsFlrUQVrbuZczAlznAvoVZwk8uIrWJIkSZJumRMgIV7jWJHMXcwN i/PbGPAx3zUGEEmSJEm3VC12bVYV7GDOyd51Rq8GjAVjc97jfXEES5IkSdItsWEVVAUfK1kb1oZZ 53ut0wVZ611jAJEkSZJ0zz4vYI1R7L0Yc/Mci/mcrOLHLsiHS+iSJEmSvqoBz1UQMBcsgk1jVPJc MCY8X12STwYQSZIkSbfstUlg7KKqmAvGXMx5OiCLE0yeLqFLkiRJ+qrKhAwewOzB1aC34OrJFecA 4bcH7O/vGjsgkiRJku4J6AFRnOvnUTwaXHkCyAW0gsdPqcMAIkmSJOmW6ElkkP28s3s9LtrVaD3I DpmnC9IMIJIkSZK+LCAoMiAjaRSNIil6g8cFLc7nkzsgkiRJku6JoEVQNHq+wkZLeiatnfuDvcFo 7xI7IJIkSZJuqSiyJRlnzKoFJEVG0QOu6+yIXI5gSZIkSfqyTCKgNcgMem/0FvQetOuEjWgQP41g GUAkSZIk3VLArnMNPQJaD67WzotYdToivZ9F9E8GEEmSJEm3XA1qQxGsXcy5mXVCyeYElDlhPN81 BhBJkiRJ96wi2Ky1qII1N3Mu5oQ5zgX0Kk4SefEVLEmSJEm3zAmQEK9xrEjmLuaGxfltDPiY7xoD iCRJkqRbqha7NqsKdjDnZO86o1cDxoKxOe/xvjiCJUmSJOmW2LAKqoKPlawNa8Os873W6YKs9a4x gEiSJEm6Z58XsMYo9l6MuXmOxXxOVvFjF+TDJXRJkiRJX9WA5yoImAsWwaYxKnkuGBOery7JJwOI JEmSpFv22iQwdlFVzAVjLuY8HZDFCSZPl9AlSZIkfVVlQgYPYPbgatBbcPXkinOA8NsD9vd3jR0Q SZIkSfcE9IAozvXzKB4NrjwB5AJaweOn1GEAkSRJknRL9CQyyH7e2b0eF+1qtB5kh8zTBWkGEEmS JElfFhAUGZCRNIpGkRS9weOCFufzyR0QSZIkSfdE0CIoGj1fYaMlPZPWzv3B3mC0d4kdEEmSJEm3 VBTZkowzZtUCkiKj6AHXdXZELkewJEmSJH1ZJhHQGmQGvTd6C3oP2nXCRjSIn0awDCCSJEmSbilg 17mGHgGtB1dr50WsOh2R3s8i+icDiCRJkqRbrga1oQjWLubczDqhZHMCypwwnu8aA4gkSZKke1YR bNZaVMGamzkXc8Ic5wJ6FSeJvPgKliRJkqRb5gRIiNc4ViRzF3PD4vw2BnzMd40BRJIkSdItVYtd m1UFO5hzsned0asBY8HYnPd4XxzBkiRJknRLbFgFVcHHStaGtWHW+V7rdEHWetcYQCRJkiTds88L WGMUey/G3DzHYj4nq/ixC/LhErokSZKkr2rAcxUEzAWLYNMYlTwXjAnPV5fkkwFEkiRJ0i17bRIY u6gq5oIxF3OeDsjiBJOnS+iSJEmSvqoyIYMHMHtwNegtuHpyxTlA+O0B+/u7xg6IJEmSpHsCekAU 5/p5FI8GV54AcgGt4PFT6jCASJIkSbolehIZZD/v7F6Pi3Y1Wg+yQ+bpgjQDiCRJkqQvCwiKDMhI GkWjSIre4HFBi/P55A6IJEmSpHsiaBEUjZ6vsNGSnklr5/5gbzDau8QOiCRJkqRbKopsScYZs2oB SZFR9IDrOjsilyNYkiRJkr4skwhoDTKD3hu9Bb0H7TphIxrETyNYBhBJkiRJtxSw61xDj4DWg6u1 8yJWnY5I72cR/ZMBRJIkSdItV4PaUARrF3NuZp1QsjkBZU4Yz3eNAUSSJEnSPasINmstqmDNzZyL OWGOcwG9ipNEXnwFS5IkSdItcwIkxGscK5K5i7lhcX4bAz7mu8YAIkmSJOmWqsWuzaqCHcw52bvO 6NWAsWBsznu8L45gSZIkSbolNqyCquBjJWvD2jDrfK91uiBrvWv+wg7Ir+pXf4V/siRJkqR/b+3z AtYYsPdizM1zLOZzsqr/2AX5cAldkiRJ0lc14LkKAuaCRbBpjEqeC8aE56tL8skAIkmSJOmWvTYJ jF1UFXPBmIs5J6vOIvpc5zWsTy6hS5IkSbqlMiGDBzB7cDXoLbh6csU5QPjtAfv7u8YOiCRJkqR7 AnpAFOf6eRSPBleeAHIBreDxU+owgEiSJEm6JXoSGWQ/7+xej4t2NVoPskPm6YI0A4gkSZKkLwsI igzISBpFo0iK3uBxQYvz+eQOiCRJkqR7ImgRFI2er7DRkp5Ja+f+YG8w2rvEDogkSZKkWyqKbEnG GbNqAUmRUfSA6zo7IpcjWJIkSZK+LJMIaA0yg94bvQW9B+06YSMaxE8jWAYQSZIkSbcUsOtcQ4+A 1oOrtfMiVp2OSO9nEf2TAUSSJEnSLVeD2lAEaxdzbmadULI5AWVOGM93jQFEkiRJ0j2rCDZrLapg zc2cizlhjnMBvYqTRF58BUuSJEnSLXMCJMRrHCuSuYu5YXF+GwM+5rvGACJJkiTplqrFrs2qgh3M Odm7zujVgLFgbM57vC+OYEmSJEm6JTasgqrgYyVrw9ow63yvdboga71rDCCSJEmS7tnnBawxir0X Y26eYzGfk1X82AX5cAldkiRJ0lc14LkKAuaCRbBpjEqeC8aE56tL8skAIkmSJOmWvTYJjF1UFXPB mIs5TwdkcYLJ0yV0SZIkSV9VmZDBA5g9uBr0Flw9ueIcIPz2gP39XWMHRJIkSdI9AT0ginP9PIpH gytPALmAVvD4KXUYQCRJkiTdEj2JDLKfd3avx0W7Gq0H2SHzdEGaAUSSJEnSlwUERQZkJI2iUSRF b/C4oMX5fHIHRJIkSdI9EbQIikbPV9hoSc+ktXN/sDcY7V1iB0SSJEnSLRVFtiTjjFm1gKTIKHrA dZ0dkcsRLEmSJElflkkEtAaZQe+N3oLeg3adsBEN4qcRLAOIJEmSpFsK2HWuoUdA68HV2nkRq05H pPeziP7JACJJkiTplqtBbSiCtYs5N7NOKNmcgDInjOe7xgAiSZIk6Z5VBJu1FlWw5mbOxZwwx7mA XsVJIi++giVJkiTpljkBEuI1jhXJ3MXcsDi/jQEf811jAJEkSZJ0S9Vi12ZVwQ7mnOxdZ/RqwFgw Nuc93hdHsCRJkiTdEhtWQVXwsZK1YW2Ydb7XOl2Qtd41BhBJkiRJ9+zzAtYYxd6LMTfPsZjPySp+ 7IJ8uIQuSZIk6asa8FwFAXPBItg0RiXPBWPC89Ul+WQAkSRJknTLXpsExi6qirlgzMWcpwOyOMHk 6RK6JEmSpK+qTMjgAcweXA16C66eXHEOEH57wP7+rrEDIkmSJOmegB4Qxbl+HsWjwZUngFxAK3j8 lDoMIJIkSZJuiZ5EBtnPO7vX46JdjdaD7JB5uiDNACJJkiTpywKCIgMykkbRKJKiN3hc0OJ8PrkD IkmSJOmeCFoERaPnK2y0pGfS2rk/2BuM9i6xAyJJkiTplooiW5JxxqxaQFJkFD3gus6OyOUIliRJ kqQvyyQCWoPMoPdGb0HvQbtO2IgG8dMIlgFEkiRJ0i0F7DrX0COg9eBq7byIVacj0vtZRP9kAJEk SZJ0y9WgNhTB2sWcm1knlGxOQJkTxvNdYwCRJEmSdM8qgs1aiypYczPnYk6Y41xAr+IkkRdfwZIk SZJ0y5wACfEax4pk7mJuWJzfxoCP+a4xgEiSJEm6pWqxa7OqYAdzTvauM3o1YCwYm/Me74sjWJIk SZJuiQ2roCr4WMnasDbMOt9rnS7IWu8aA4gkSZKke/Z5AWuMYu/FmJvnWMznZBU/dkE+XEKXJEmS 9FUNeK6CgLlgEWwao5LngjHh+eqSfDKASJIkSbplr00CYxdVxVww5mLO0wFZnGDydAldkiRJ0ldV JmTwAGYPrga9BVdPrjgHCL89YH9/19gBkSRJknRPQA+I4lw/j+LR4MoTQC6gFTx+Sh0GEEmSJEm3 RE8ig+znnd3rcdGuRutBdsg8XZBmAJEkSZL0ZQFBkQEZSaNoFEnRGzwuaHE+n9wBkSRJknRPBC2C otHzFTZa0jNp7dwf7A1Ge5fYAZEkSZJ0S0WRLck4Y1YtICkyih5wXWdH5HIES5IkSdKXZRIBrUFm 0Hujt6D3oF0nbESD+GkEywAiSZIk6ZYCdp1r6BHQenC1dl7EqtMR6f0son8ygEiSJEm65WpQG4pg 7WLOzawTSjYnoMwJ4/muMYBIkiRJumcVwWatRRWsuZlzMSfMcS6gV3GSyIuvYEmSJEm6ZU6AhHiN Y0UydzE3LM5vY8DHfNcYQCRJkiTdUrXYtVlVsIM5J3vXGb0aMBaMzXmP98URLEmSJEm3xIZVUBV8 rGRtWBtmne+1ThdkrXeNAUSSJEnSPfu8gDVGsfdizM1zLOZzsoofuyAfLqFLkiRJ+qoGPFdBwFyw CDaNUclzwZjwfHVJPhlAJEmSJN2y1yaBsYuqYi4YczHn6YAsTjB5uoQuSZIk6asqEzJ4ALMHV4Pe gqsnV5wDhN8esL+/a+yASJIkSbonoAdEca6fR/FocOUJIBfQCh4/pQ4DiCRJkqRboieRQfbzzu71 uGhXo/UgO2SeLkgzgEiSJEn6soCgyICMpFE0iqToDR4XtDifT+6ASJIkSbonghZB0ej5Chst6Zm0 du4P9gajvUvsgEiSJEm6paLIlmScMasWkBQZRQ+4rrMjcjmCJUmSJOnLMomA1iAz6L3RW9B70K4T NqJB/DSCZQCRJEmSdEsBu8419AhoPbhaO7yEpoQAACAASURBVC9i1emI9H4W0T8ZQCRJkiTdcjWo DUWwdjHnZtYJJZsTUOaE8XzXGEAkSZIk3bOKYLPWogrW3My5mBPmOBfQqzhJ5MVXsCRJkiTdMidA QrzGsSKZu5gbFue3MeBjvmsMIJIkSZJuqVrs2qwq2MGck73rjF4NGAvG5rzH++IIliRJkqRbYsMq qAo+VrI2rA2zzvdapwuy1rvGACJJkiTpnn1ewBqj2Hsx5uY5FvM5WcWPXZAPl9AlSZIkfVUDnqsg YC5YBJvGqOS5YEx4vroknwwgkiRJkm7Za5PA2EVVMReMuZjzdEAWJ5g8XUKXJEmS9FWVCRk8gNmD q0FvwdWTK84Bwm8P2N/fNXZAJEmSJN0T0AOiONfPo3g0uPIEkAtoBY+fUocBRJIkSdIt0ZPIIPt5 Z/d6XLSr0XqQHTJPF6QZQCRJkiR9WUBQZEBG0igaRVL0Bo8LWpzPJ3dAJEmSJN0TQYugaPR8hY2W 9ExaO/cHe4PR3iV2QCRJkiTdUlFkSzLOmFULSIqMogdc19kRuRzBkiRJkvRlmURAa5AZ9N7oLeg9 aNcJG9EgfhrBMoBIkiRJuqWAXecaegS0HlytnRex6nREej+L6J8MIJIkSZJuuRrUhiJYu5hzM+uE ks0JKHPCeL5rDCCSJEmS7llFsFlrUQVrbuZczAlznAvoVZwk8uIrWJIkSZJumRMgIV7jWJHMXcwN i/PbGPAx3zUGEEmSJEm3VC12bVYV7GDOyd51Rq8GjAVjc97jfXEES5IkSdItsWEVVAUfK1kb1oZZ 53ut0wVZ611jAJEkSZJ0zz4vYI1R7L0Yc/Mci/mcrOLHLsiHS+iSJEmSvqoBz1UQMBcsgk1jVPJc MCY8X12STwYQSZIkSbfstUlg7KKqmAvGXMx5OiCLE0yeLqFLkiRJ+qrKhAwewOzB1aC34OrJFecA 4bcH7O/vGjsgkiRJku4J6AFRnOvnUTwaXHkCyAW0gsdPqcMAIkmSJOmW6ElkkP28s3s9LtrVaD3I DpmnC9IMIJIkSZK+LCAoMiAjaRSNIil6g8cFLc7nkzsgkiRJku6JoEVQNHq+wkZLeiatnfuDvcFo 7xI7IJIkSZJuqSiyJRlnzKoFJEVG0QOu6+yIXI5gSZIkSfqyTCKgNcgMem/0FvQetOuEjWgQP41g GUAkSZIk3VLArnMNPQJaD67WzotYdToivZ9F9E8GEEmSJEm3XA1qQxGsXcy5mXVCyeYElDlhPN81 BhBJkiRJ96wi2Ky1qII1N3Mu5oQ5zgX0Kk4SefEVLEmSJEm3zAmQEK9xrEjmLuaGxfltDPiY7xoD iCRJkqRbqha7NqsKdjDnZO86o1cDxoKxOe/xvjiCJUmSJOmW2LAKqoKPlawNa8Os873W6YKs9a4x gEiSJEm6Z58XsMYo9l6MuXmOxXxOVvFjF+TDJXRJkiRJX9WA5yoImAsWwaYxKnkuGBOery7JJwOI JEmSpFv22iQwdlFVzAVjLuY8HZDFCSZPl9AlSZIkfVVlQgYPYPbgatBbcPXkinOA8NsD9vd3jR0Q SZIkSfcE9IAozvXzKB4NrjwB5AJaweOn1GEAkSRJknRL9CQyyH7e2b0eF+1qtB5kh8zTBWkGEEmS JElfFhAUGZCRNIpGkRS9weOCFufzyR0QSZIkSfdE0CIoGj1fYaMlPZPWzv3B3mC0d4kdEEmSJEm3 VBTZkowzZtUCkiKj6AHXdXZELkewJEmSJH1ZJhHQGmQGvTd6C3oP2nXCRjSIn0awDCCSJEmSbilg 17mGHgGtB1dr50WsOh2R3s8i+icDiCRJkqRbrga1oQjWLubczDqhZHMCypwwnu8aA4gkSZKke1YR bNZaVMGamzkXc8Ic5wJ6FSeJvPgKliRJkqRb5gRIiNc4ViRzF3PD4vw2BnzMd40BRJIkSdItVYtd m1UFO5hzsned0asBY8HYnPd4XxzBkiRJknRLbFgFVcHHStaGtWHW+V7rdEHWetcYQCRJkiTds88L WGMUey/G3DzHYj4nq/ixC/LhErokSZKkr2rAcxUEzAWLYNMYlTwXjAnPV5fkkwFEkiRJ0i17bRIY u6gq5oIxF3OeDsjiBJOnS+iSJEmSvqoyIYMHMHtwNegtuHpyxTlA+O0B+/u7xg6IJEmSpHsCekAU 5/p5FI8GV54AcgGt4PFT6jCASJIkSbolehIZZD/v7F6Pi3Y1Wg+yQ+bpgjQDiCRJkqQvCwiKDMhI GkWjSIre4HFBi/P55A6IJEmSpHsiaBEUjZ6vsNGSnklr5/5gbzDau8QOiCRJkqRbKopsScYZs2oB SZFR9IDrOjsilyNYkiRJkr4skwhoDTKD3hu9Bb0H7TphIxrETyNYBhBJkiRJtxSw61xDj4DWg6u1 8yJWnY5I72cR/ZMBRJIkSdItV4PaUARrF3NuZp1QsjkBZU4Yz3eNAUSSJEnSPasINmstqmDNzZyL OWGOcwG9ipNEXnwFS5IkSdItcwIkxGscK5K5i7lhcX4bAz7mu8YAIkmSJOmWqsWuzaqCHcw52bvO 6NWAsWBsznu8L45gSZIkSbolNqyCquBjJWvD2jDrfK91uiBrvWsMIJIkSZLu2ecFrDGKvRdjbp5j MZ+TVfzYBflwCV2SJEnSVzXguQoC5oJFsGmMSp4LxoTnq0vyyQAiSZIk6Za9NgmMXVQVc8GYizlP B2RxgsnTJXRJkiRJX1WZkMEDmD24GvQWXD254hwg/PaA/f1dYwdEkiRJ0j0BPSCKc/08ikeDK08A uYBW8PgpdRhAJEmSJN3y/7Z3967WZmmdx7/XWuve51Q91V1tC9W+BJZJjR2IiLmJmTAYaktPZqL/ wzD0X2FiKLTSk5jogGPSOBgYi1AYaKq0UFovffa91romWHuffZ7ueTv3ZkqE7wc2u9jPWVDpj+st WiFKUNras7udNupWqS0oDUpZVZBqAJEkSZJ0t4AgKQElCpWkkhSSVuG0QY31uXIGRJIkSdIxEdQI kkorl7BRC60Ual33B1uFvd6eWAGRJEmSdEhGUmqhxGqzqgGFpETSArZtzYhstmBJkiRJulspRECt UErQWqXVoLWgbitsRIV40YJlAJEkSZJ0SAIz1zX0CKgt2GpdG7FyVURaW4PoVwYQSZIkSYdsFXJC EoyZ9D7puULJZAWU3mE/394YQCRJkiQdM5JgMsYgE0af9D7oHfq+LqBnspLIhVuwJEmSJB3SO0CB uLRjRaHPpE8YrN/2HZ767Y0BRJIkSdIhmYOZk5EJM+i9M2eu1qsd9gH7ZO3jvbAFS5IkSdIhMWEk ZAZPozAmjAk91/cYqwoyxu2NAUSSJEnSMXNtwNr3ZM7B3ifnfdDPnZE8z4I8OYQuSZIk6V4VOI+E gD5gEEwqexbOA/YO50uV5MoAIkmSJOmQOSYF2GeSmfQBex/0vioggxVMzg6hS5IkSbpXlgIlOAG9 BVuFVoOtFbZYBwgfTjAfb2+sgEiSJEk6JqAFRLKun0dyqrCVFUA2oCacXqQOA4gkSZKkQ6IVogSl rT2722mjbpXagtKglFUFqQYQSZIkSXcLCJISUKJQSSpJIWkVThvUWJ8rZ0AkSZIkHRNBjSCptHIJ G7XQSqHWdX+wVdjr7YkVEEmSJEmHZCSlFkqsNqsaUEhKJC1g29aMyGYLliRJkqS7lUIE1AqlBK1V Wg1aC+q2wkZUiBctWAYQSZIkSYckMHNdQ4+A2oKt1rURK1dFpLU1iH5lAJEkSZJ0yFYhJyTBmEnv k54rlExWQOkd9vPtjQFEkiRJ0jEjCSZjDDJh9Envg96h7+sCeiYriVy4BUuSJEnSIb0DFIhLO1YU +kz6hMH6bd/hqd/eGEAkSZIkHZI5mDkZmTCD3jtz5mq92mEfsE/WPt4LW7AkSZIkHRITRkJm8DQK Y8KY0HN9j7GqIGPc3hhAJEmSJB0z1wasfU/mHOx9ct4H/dwZyfMsyJND6JIkSZLuVYHzSAjoAwbB pLJn4Txg73C+VEmuDCCSJEmSDpljUoB9JplJH7D3Qe+rAjJYweTsELokSZKke2UpUIIT0FuwVWg1 2Fphi3WA8OEE8/H2xgqIJEmSpGMCWkAk6/p5JKcKW1kBZANqwulF6jCASJIkSTokWiFKUNras7ud NupWqS0oDUpZVZBqAJEkSZJ0t4AgKQElCpWkkhSSVuG0QY31uXIGRJIkSdIxEdQIkkorl7BRC60U al33B1uFvd6eWAGRJEmSdEhGUmqhxGqzqgGFpETSArZtzYhstmBJkiRJulspRECtUErQWqXVoLWg bitsRIV40YJlAJEkSZJ0SAIz1zX0CKgt2GpdG7FyVURaW4PoVwYQSZIkSYdsFXJCEoyZ9D7puULJ ZAWU3mE/394YQCRJkiQdM5JgMsYgE0af9D7oHfq+LqBnspLIhVuwJEmSJB3SO0CBuLRjRaHPpE8Y rN/2HZ767Y0BRJIkSdIhmYOZk5EJM+i9M2eu1qsd9gH7ZO3jvbAFS5IkSdIhMWEkZAZPozAmjAk9 1/cYqwoyxu2NAUSSJEnSMXNtwNr3ZM7B3ifnfdDPnZE8z4I8OYQuSZIk6V4VOI+EgD5gEEwqexbO A/YO50uV5MoAIkmSJOmQOSYF2GeSmfQBex/0vioggxVMzg6hS5IkSbpXlgIlOAG9BVuFVoOtFbZY BwgfTjAfb2+sgEiSJEk6JqAFRLKun0dyqrCVFUA2oCacXqQOA4gkSZKkQ6IVogSlrT2722mjbpXa gtKglFUFqQYQSZIkSXcLCJISUKJQSSpJIWkVThvUWJ8rZ0AkSZIkHRNBjSCptHIJG7XQSqHWdX+w Vdjr7YkVEEmSJEmHZCSlFkqsNqsaUEhKJC1g29aMyGYLliRJkqS7lUIE1AqlBK1VWg1aC+q2wkZU iBctWAYQSZIkSYckMHNdQ4+A2oKt1rURK1dFpLU1iH5lAJEkSZJ0yFYhJyTBmEnvk54rlExWQOkd 9vPtjQFEkiRJ0jEjCSZjDDJh9Envg96h7+sCeiYriVy4BUuSJEnSIb0DFIhLO1YU+kz6hMH6bd/h qd/eGEAkSZIkHZI5mDkZmTCD3jtz5mq92mEfsE/WPt4LW7AkSZIkHRITRkJm8DQKY8KY0HN9j7Gq IGPc3hhAJEmSJB0z1wasfU/mHOx9ct4H/dwZyfMsyJND6JIkSZLuVYHzSAjoAwbBpLJn4Txg73C+ VEmuDCCSJEmSDpljUoB9JplJH7D3Qe+rAjJYweTsELokSZKke2UpUIIT0FuwVWg12Fphi3WA8OEE 8/H2xgqIJEmSpGMCWkAk6/p5JKcKW1kBZANqwulF6jCASJIkSTokWiFKUNras7udNupWqS0oDUpZ VZBqAJEkSZJ0t4AgKQElCpWkkhSSVuG0QY31uXIGRJIkSdIxEdQIkkorl7BRC60Ual33B1uFvd6e WAGRJEmSdEhGUmqhxGqzqgGFpETSArZtzYhstmBJkiRJulspRECtUErQWqXVoLWgbitsRIV40YJl AJEkSZJ0SAIz1zX0CKgt2GpdG7FyVURaW4PoVwYQSZIkSYdsFXJCEoyZ9D7puULJZAWU3mE/394Y QCRJkiQdM5JgMsYgE0af9D7oHfq+LqBnspLIhVuwJEmSJB3SO0CBuLRjRaHPpE8YrN/2HZ767Y0B RJIkSdIhmYOZk5EJM+i9M2eu1qsd9gH7ZO3jvbAFS5IkSdIhMWEkZAZPozAmjAk91/cYqwoyxu2N AUSSJEnSMXNtwNr3ZM7B3ifnfdDPnZE8z4I8OYQuSZIk6V4VOI+EgD5gEEwqexbOA/YO50uV5MoA IkmSJOmQOSYF2GeSmfQBex/0vioggxVMzg6hS5IkSbpXlgIlOAG9BVuFVoOtFbZYBwgfTjAfb2+s gEiSJEk6JqAFRLKun0dyqrCVFUA2oCacXqQOA4gkSZKkQ6IVogSlrT2722mjbpXagtKglFUFqQYQ SZIkSXcLCJISUKJQSSpJIWkVThvUWJ8rZ0AkSZIkHRNBjSCptHIJG7XQSqHWdX+wVdjr7YkVEEmS JEmHZCSlFkqsNqsaUEhKJC1g29aMyGYLliRJkqS7lUIE1AqlBK1VWg1aC+q2wkZUiBctWAYQSZIk SYckMHNdQ4+A2oKt1rURK1dFpLU1iH5lAJEkSZJ0yFYhJyTBmEnvk54rlExWQOkd9vPtjQFEkiRJ 0jEjCSZjDDJh9Envg96h7+sCeiYriVy4BUuSJEnSIb0DFIhLO1YU+kz6hMH6bd/hqd/eGEAkSZIk HZI5mDkZmTCD3jtz5mq92mEfsE/WPt4LW7AkSZIkHRITRkJm8DQKY8KY0HN9j7GqIGPc3hhAJEmS JB0z1wasfU/mHOx9ct4H/dwZyfMsyJND6JIkSZLuVYHzSAjoAwbBpLJn4Txg73C+VEmuDCCSJEmS DpljUoB9JplJH7D3Qe+rAjJYweTsELokSZKke2UpUIIT0FuwVWg12Fphi3WA8OEE8/H2xgqIJEmS pGMCWkAk6/p5JKcKW1kBZANqwulF6jCASJIkSTokWiFKUNras7udNupWqS0oDUpZVZBqAJEkSZJ0 t4AgKQElCpWkkhSSVuG0QY31uXIGRJIkSdIxEdQIkkorl7BRC60Ual33B1uFvd6eWAGRJEmSdEhG UmqhxGqzqgGFpETSArZtzYhstmBJkiRJulspRECtUErQWqXVoLWgbitsRIV40YJlAJEkSZJ0SAIz 1zX0CKgt2GpdG7FyVURaW4PoVwYQSZIkSYdsFXJCEoyZ9D7puULJZAWU3mE/394YQCRJkiQdM5Jg MsYgE0af9D7oHfq+LqBnspLIhVuwJEmSJB3SO0CBuLRjRaHPpE8YrN/2HZ767Y0BRJIkSdIhmYOZ k5EJM+i9M2eu1qsd9gH7ZO3jvbAFS5IkSdIhMWEkZAZPozAmjAk91/cYqwoyxu2NAUSSJEnSMXNt wNr3ZM7B3ifnfdDPnZE8z4I8OYQuSZIk6V4VOI+EgD5gEEwqexbOA/YO50uV5MoAIkmSJOmQOSYF 2GeSmfQBex/0vioggxVMzg6hS5IkSbpXlgIlOAG9BVuFVoOtFbZYBwgfTjAfb2+sgEiSJEk6JqAF RLKun0dyqrCVFUA2oCacXqQOA4gkSZKkQ6IVogSlrT2722mjbpXagtKglFUFqQYQSZIkSXcLCJIS UKJQSSpJIWkVThvUWJ8rZ0AkSZIkHRNBjSCptHIJG7XQSqHWdX+wVdjr7YkVEEmSJEmHZCSlFkqs NqsaUEhKJC1g29aMyGYLliRJkqS7lUIE1AqlBK1VWg1aC+q2wkZUiBctWAYQSZIkSYckMHNdQ4+A 2oKt1rURK1dFpLU1iH5lAJEkSZJ0yFYhJyTBmEnvk54rlExWQOkd9vPtjQFEkiRJ0jEjCSZjDDJh 9Envg96h7+sCeiYriVy4BUuSJEnSIb0DFIhLO1YU+kz6hMH6bd/hqd/eGEAkSZIkHZI5mDkZmTCD 3jtz5mq92mEfsE/WPt4LW7AkSZIkHRITRkJm8DQKY8KY0HN9j7GqIGPc3hhAJEmSJB0z1wasfU/m HOx9ct4H/dwZyfMsyJND6JIkSZLuVYHzSAjoAwbBpLJn4Txg73C+VEmuDCCSJEmSDpljUoB9JplJ H7D3Qe+rAjJYweTsELokSZKke2UpUIIT0FuwVWg12Fphi3WA8OEE8/H2xgqIJEmSpGMCWkAk6/p5 JKcKW1kBZANqwulF6jCASJIkSTokWiFKUNras7udNupWqS0oDUpZVZBqAJEkSZJ0t4AgKQElCpWk khSSVuG0QY31uXIGRJIkSdIxEdQIkkorl7BRC60Ual33B1uFvd6eWAGRJEmSdEhGUmqhxGqzqgGF pETSArZtzYhstmBJkiRJulspRECtUErQWqXVoLWgbitsRIV40YJlAJEkSZJ0SAIz1zX0CKgt2Gpd G7FyVURaW4PoVwYQSZIkSYdsFXJCEoyZ9D7puULJZAWU3mE/394YQCRJkiQdM5JgMsYgE0af9D7o Hfq+LqBnspLIhVuwJEmSJB3SO0CBuLRjRaHPpE8YrN/2HZ767Y0BRJIkSdIhmYOZk5EJM+i9M2eu 1qsd9gH7ZO3jvbAFS5IkSdIhMWEkZAZPozAmjAk91/cYqwoyxu2NAUSSJEnSMXNtwNr3ZM7B3ifn fdDPnZE8z4I8OYQuSZIk6V4VOI+EgD5gEEwqexbOA/YO50uV5MoAIkmSJOmQOSYF2GeSmfQBex/0 vioggxVMzg6hS5IkSbpXlgIlOAG9BVuFVoOtFbZYBwgfTjAfb2+sgEiSJEk6JqAFRLKun0dyqrCV FUA2oCacXqQOA4gkSZKkQ6IVogSlrT2722mjbpXagtKglFUFqQYQSZIkSXcLCJISUKJQSSpJIWkV ThvUWJ8rZ0AkSZIkHRNBjSCptHIJG7XQSqHWdX+wVdjr7YkVEEmSJEmHZCSlFkqsNqsaUEhKJC1g 29aMyGYLliRJkqS7lUIE1AqlBK1VWg1aC+q2wkZUiBctWAYQSZIkSYckMHNdQ4+A2oKt1rURK1dF pLU1iH5lAJEkSZJ0yFYhJyTBmEnvk54rlExWQOkd9vPtjQFEkiRJ0jEjCSZjDDJh9Envg96h7+sC eiYriVy4BUuSJEnSIb0DFIhLO1YU+kz6hMH6bd/hqd/eGEAkSZIkHZI5mDkZmTCD3jtz5mq92mEf sE/WPt4LW7AkSZIkHRITRkJm8DQKY8KY0HN9j7GqIGPc3hhAJEmSJB0z1wasfU/mHOx9ct4H/dwZ yfMsyJND6JIkSZLuVYHzSAjoAwbBpLJn4Txg73C+VEmuDCCSJEmSDpljUoB9JplJH7D3Qe+rAjJY weTsELokSZKke2UpUIIT0FuwVWg12Fphi3WA8OEE8/H2xgqIJEmSpGMCWkAk6/p5JKcKW1kBZANq wulF6jCASJIkSTokWiFKUNras7udNupWqS0oDUpZVZBqAJEkSZJ0t4AgKQElCpWkkhSSVuG0QY31 uXIGRJIkSdIxEdQIkkorl7BRC60Ual33B1uFvd6eWAGRJEmSdEhGUmqhxGqzqgGFpETSArZtzYhs tmBJkiRJulspRECtUErQWqXVoLWgbitsRIV40YJlAJEkSZJ0SAIz1zX0CKgt2GpdG7FyVURaW4Po VwYQSZIkSYdsFXJCEoyZ9D7puULJZAWU3mE/394YQCRJkiQdM5JgMsYgE0af9D7oHfq+LqBnspLI hVuwJEmSJB3SO0CBuLRjRaHPpE8YrN/2HZ767Y0BRJIkSdIhmYOZk5EJM+i9M2eu1qsd9gH7ZO3j vbAFS5IkSdIhMWEkZAZPozAmjAk91/cYqwoyxu2NAUSSJEnSMXNtwNr3ZM7B3ifnfdDPnZE8z4I8 OYQuSZIk6V4VOI+EgD5gEEwqexbOA/YO50uV5MoAIkmSJOmQOSYF2GeSmfQBex/0vioggxVMzg6h S5IkSbpXlgIlOAG9BVuFVoOtFbZYBwgfTjAfb2+sgEiSJEk6JqAFRLKun0dyqrCVFUA2oCacXqQO A4gkSZKkQ6IVogSlrT2722mjbpXagtKglFUFqQYQSZIkSXcLCJISUKJQSSpJIWkVThvUWJ8rZ0Ak SZIkHRNBjSCptHIJG7XQSqHWdX+wVdjr7YkVEEmSJEmHZCSlFkqsNqsaUEhKJC1g29aMyGYLliRJ kqS7lUIE1AqlBK1VWg1aC+q2wkZUiBctWAYQSZIkSYckMHNdQ4+A2oKt1rURK1dFpLU1iH5lAJEk SZJ0yFYhJyTBmEnvk54rlExWQOkd9vPtjQFEkiRJ0jEjCSZjDDJh9Envg96h7+sCeiYriVy4BUuS JEnSIb0DFIhLO1YU+kz6hMH6bd/hqd/eGEAkSZIkHZI5mDkZmTCD3jtz5mq92mEfsE/WPt4LW7Ak SZIkHRITRkJm8DQKY8KY0HN9j7GqIGPc3hhAJEmSJB0z1wasfU/mHOx9ct4H/dwZyfMsyJND6JIk SZLuVYHzSAjoAwbBpLJn4Txg73C+VEmuDCCSJEmSDpljUoB9JplJH7D3Qe+rAjJYweTsELokSZKk e2UpUIIT0FuwVWg12Fphi3WA8OEE8/H2xgqIJEmSpGMCWkAk6/p5JKcKW1kBZANqwulF6rACIkmS JOnV8q/+43/hr97+7T//H/7+P337298BKyCSJEmSvkQGEEmSJElfiogwgEiSJEn6/6+1RinFACJJ kiTpdUp5fYx455132LbNACJJkiTpdSLi1W++9rWv8c477xhAJEmSJL1Oa69fpvszP/MzfPWrXzWA SJIkSXqdIy1YH374IV/5ylcMIJIkSZJep5TCf/uzP/vO/+vf/9fvfe87P/dzP8ebN28MIJIkSZJe p5RCa43//ud//n8NIX/2p3/6nW9+85v8xE/8BNu2eQldkiRJ0jEPDw/8j7/8y+/03jmfzzw9PTHG oLXGmzdveP/99/kP77/Pu+++S2uN3rsBRJIkSdLrjDGotRIRtNZ4fHxcK3YvlZHT6URrjYeHB1pr ZCZjDAOIJEmSpNfLTHrvzDnJTGqt1Fqfg8e2bWzbRkQw52SMwb7v7PtuAJEkSZL0OhHxHEIiglLK 8/f132EFlYhg33fO5zOAAUSSJEnS67TW+Pzzz4EVNrZtew4jAHPOtyok5/OZTz75hJ/92Z81gEiS JEl6ndYaX3zxxXPoqLUy53xut6q1kplkJk9PT3z66ac8Pj7y/vvvG0AkSZIkvc6ck5/+6Z/mH/7h Hzifz/zwhz/kzZs3z/Mf16rH9d9qrfzCL/zCGlL/t/6flyRJkvTvS2by9a9/ndPpxN/8zd/wySef 8ObNGx4eHp7/5unpic8//5wPPviAX/7lX+aDDz7gs88+M4BIkiRJep1SCqUUvvnNb/KVr3yFv/3b v+Vf/uVf+OSTT/jiiy949913+frXHK3J5AAAA5BJREFUv84v/uIv8ku/9Eu8efOG3jvgELokSZKk V7rOe7TW+PDDD3nvvff47LPPOJ1OPD4+cjqd3vrvl8q/0f+zJEmSpH+nIuJ549W2bfzUT/0U3/jG N6i1/ti//ygrIJIkSZJe7Rowrvc/vvrVrxIRb63nfXkP5MoKiCRJkqRX+9EqRymFd955B1iB43oD 5Pq3z3/35f5vSpIkSfr3LjMZYzzf+rh6eHjgr//6r9/6u5cfMIBIkiRJusPL6sbHH3/M7//+7/Px xx+TmZRSnisl178zgEiSJEl6tZeVj6s/+IM/AOAP//APnyskP/q3BhBJkiRJr3K9A/LSX/zFX/Dx xx8D8Pd///d8//vfdwZEkiRJ0v1+dPbj888/57vf/e5bv/3RH/0Rn3766Y+t4zWASJIkSXq1l8Hi T/7kT/jBD37w1r9/+umn/PEf//GPtWp5B0SSJEnSIdcQ8q1vfYvf/u3fpvfOP/3TP/Hw8MC2bTw8 PFgBkSRJkvRvxwAiSZIk6UtjAJEkSZL0pTGASJIkSfrSOIQuSZIk6VX+8R//kc8++4wf/OAH/PCH P2Tfd8YYzDn55JNPOJ1OnE4n3rx5w/vvv8+bN2947733AAOIJEmSpFf64IMP+MY3vsHP//zPs20b EUFE/C+3YD0+PlJKoffOZ599ZguWJEmSpC+PAUSSJEnSl8YAIkmSJOlLYwCRJEmS9KVxCF2SJEnS IZkJwHe/+12+973vkZk8PT1RSqGUwq//+q/ze7/3e2+9sQIiSZIk6dWu4QPgN37jN/jJn/zJt/79 vffe4zd/8zeJiLd+N4BIkiRJepXr2t2rd999l29961tv/fZbv/VbvPfee28FFTCASJIkSXqlOSdz zrd++7Vf+zU++ugjAD788EN+9Vd/9TmQvAwhBhBJkiRJr/ajrVUAv/M7vwPAt7/9bWqtz3/z8m8N IJIkSZIOe1nd+Oijj/jd3/1dPvroIyKCOSeZ+fwBA4gkSZKkO7ysbmQmv/Irv0JmPrdpvQwgc04D iCRJkqTXGWP8WGUDYN/3/23wuH4MIJIkSZJe5Uc3W2UmYwz+9V//9Tlw9N4ZYzx/zzkZY3iIUJIk SdLrXAPFnJPz+cz5fObTTz/l888/p5RCRFBKofdOKeU5oBhAJEmSJL1a752/+7u/45//+Z/54osv 2PediODx8ZHT6cS2bZxOp+f/rrUyxuCLL77gfwKLOs1HD90r4gAAAABJRU5ErkJggg== 76930 2010-12-17 17:48:48 -0800 2010-12-30 20:06:31 -0800 Patch bug-51151-20101217174847.patch text/plain 7452 mihaip ZGlmZiAtLWdpdCBhL1dlYkNvcmUvcGxhdGZvcm0vY2hyb21pdW0vU2Nyb2xsYmFyVGhlbWVDaHJv bWl1bU1hYy5tbSBiL1dlYkNvcmUvcGxhdGZvcm0vY2hyb21pdW0vU2Nyb2xsYmFyVGhlbWVDaHJv bWl1bU1hYy5tbQppbmRleCBiNDdlOTk4NWIxMWE1ODcyMjk1NDhlN2Q4OTFjMjQ5NmYyZjQ0ZGI4 Li40ZDhmNjc2ZTZhZDIyZTMyZmFlNGU3OTc2ZmU1MmNhYjMyMjZmOTk2IDEwMDY0NAotLS0gYS9X ZWJDb3JlL3BsYXRmb3JtL2Nocm9taXVtL1Njcm9sbGJhclRoZW1lQ2hyb21pdW1NYWMubW0KKysr IGIvV2ViQ29yZS9wbGF0Zm9ybS9jaHJvbWl1bS9TY3JvbGxiYXJUaGVtZUNocm9taXVtTWFjLm1t CkBAIC0zOSw2ICszOSw0OSBAQAogdXNpbmcgbmFtZXNwYWNlIHN0ZDsKIHVzaW5nIG5hbWVzcGFj ZSBXZWJDb3JlOwogCitAaW50ZXJmYWNlIEZha2VBY3RpdmVXaW5kb3cgOiBOU1dpbmRvdyB7CitA cHJpdmF0ZQorICAgIEJPT0wgaGFzQWN0aXZlQ29udHJvbHM7Cit9CisrIChOU1dpbmRvdyopYWx3 YXlzQWN0aXZlV2luZG93OworLSAoaWQpaW5pdFdpdGhBY3RpdmVDb250cm9sczooQk9PTClfaGFz QWN0aXZlQ29udHJvbHM7CistIChCT09MKV9oYXNBY3RpdmVDb250cm9sczsKK0BlbmQKKworQGlt cGxlbWVudGF0aW9uIEZha2VBY3RpdmVXaW5kb3cKKworc3RhdGljIE5TV2luZG93KiBhbHdheXNB Y3RpdmVXaW5kb3cgPSBuaWw7CitzdGF0aWMgTlNXaW5kb3cqIGFsd2F5c0luYWN0aXZlV2luZG93 ID0gbmlsOworCisrIChOU1dpbmRvdyopYWx3YXlzQWN0aXZlV2luZG93Cit7CisgICAgaWYgKGFs d2F5c0FjdGl2ZVdpbmRvdyA9PSBuaWwpIHsKKyAgICAgICAgYWx3YXlzQWN0aXZlV2luZG93ID0g W1tzZWxmIGFsbG9jXSBpbml0V2l0aEFjdGl2ZUNvbnRyb2xzOllFU107CisgICAgfQorICAgIHJl dHVybiBhbHdheXNBY3RpdmVXaW5kb3c7Cit9CisKKysgKE5TV2luZG93KilhbHdheXNJbmFjdGl2 ZVdpbmRvdworeworICAgIGlmIChhbHdheXNJbmFjdGl2ZVdpbmRvdyA9PSBuaWwpIHsKKyAgICAg ICAgYWx3YXlzSW5hY3RpdmVXaW5kb3cgPSBbW3NlbGYgYWxsb2NdIGluaXRXaXRoQWN0aXZlQ29u dHJvbHM6Tk9dOworICAgIH0KKyAgICByZXR1cm4gYWx3YXlzSW5hY3RpdmVXaW5kb3c7Cit9CisK Ky0gKGlkKWluaXRXaXRoQWN0aXZlQ29udHJvbHM6KEJPT0wpX2hhc0FjdGl2ZUNvbnRyb2xzCit7 CisgICAgc2VsZiA9IFtzdXBlciBpbml0XTsKKyAgICBoYXNBY3RpdmVDb250cm9scyA9IF9oYXNB Y3RpdmVDb250cm9sczsKKyAgICByZXR1cm4gc2VsZjsKK30KKworLSAoQk9PTClfaGFzQWN0aXZl Q29udHJvbHMgeworICByZXR1cm4gaGFzQWN0aXZlQ29udHJvbHM7Cit9CisKK0BlbmQKKwogLy8g VGhpcyBmaWxlIChhbmQgaXRzIGFzc29jaWF0ZWQgLmggZmlsZSkgaXMgYSBjbG9uZSBvZiBTY3Jv bGxiYXJUaGVtZU1hYy5tbS4KIC8vIEJlY2F1c2Ugd2Ugd2FudCB0byBkcmF3IHRpY2ttYXJrcyBp biB0aGUgc2Nyb2xsYmFyLCB3ZSBtdXN0IG1haW50YWluIGEgZm9yay4KIC8vIFBsZWFzZSBtYWlu dGFpbiB0aGlzIGZpbGUgYnkgcGVyZm9ybWluZyBwYXJhbGxlbCBjaGFuZ2VzIHRvIGl0LgpAQCAt MzQ4LDcwICszOTEsNDkgQEAgYm9vbCBTY3JvbGxiYXJUaGVtZUNocm9taXVtTWFjOjpzaG91bGRD ZW50ZXJPblRodW1iKFNjcm9sbGJhciosIGNvbnN0IFBsYXRmb3JtTW8KICAgICByZXR1cm4gZXZ0 LmFsdEtleSgpOwogfQogCi1zdGF0aWMgaW50IHNjcm9sbGJhclBhcnRUb0hJUHJlc3NlZFN0YXRl KFNjcm9sbGJhclBhcnQgcGFydCkKLXsKLSAgICBzd2l0Y2ggKHBhcnQpIHsKLSAgICAgICAgY2Fz ZSBCYWNrQnV0dG9uU3RhcnRQYXJ0OgotICAgICAgICAgICAgcmV0dXJuIGtUaGVtZVRvcE91dHNp ZGVBcnJvd1ByZXNzZWQ7Ci0gICAgICAgIGNhc2UgQmFja0J1dHRvbkVuZFBhcnQ6Ci0gICAgICAg ICAgICByZXR1cm4ga1RoZW1lVG9wT3V0c2lkZUFycm93UHJlc3NlZDsgLy8gVGhpcyBkb2VzIG5v dCBtYWtlIG11Y2ggc2Vuc2UuICBGb3Igc29tZSByZWFzb24gdGhlIG91dHNpZGUgY29uc3RhbnQg aXMgcmVxdWlyZWQuCi0gICAgICAgIGNhc2UgRm9yd2FyZEJ1dHRvblN0YXJ0UGFydDoKLSAgICAg ICAgICAgIHJldHVybiBrVGhlbWVUb3BJbnNpZGVBcnJvd1ByZXNzZWQ7Ci0gICAgICAgIGNhc2Ug Rm9yd2FyZEJ1dHRvbkVuZFBhcnQ6Ci0gICAgICAgICAgICByZXR1cm4ga1RoZW1lQm90dG9tT3V0 c2lkZUFycm93UHJlc3NlZDsKLSAgICAgICAgY2FzZSBUaHVtYlBhcnQ6Ci0gICAgICAgICAgICBy ZXR1cm4ga1RoZW1lVGh1bWJQcmVzc2VkOwotICAgICAgICBkZWZhdWx0OgotICAgICAgICAgICAg cmV0dXJuIDA7Ci0gICAgfQotfQotCiBib29sIFNjcm9sbGJhclRoZW1lQ2hyb21pdW1NYWM6OnBh aW50KFNjcm9sbGJhciogc2Nyb2xsYmFyLCBHcmFwaGljc0NvbnRleHQqIGNvbnRleHQsIGNvbnN0 IEludFJlY3QmIGRhbWFnZVJlY3QpCiB7Ci0gICAgSElUaGVtZVRyYWNrRHJhd0luZm8gdHJhY2tJ bmZvOwotICAgIHRyYWNrSW5mby52ZXJzaW9uID0gMDsKLSAgICB0cmFja0luZm8ua2luZCA9IHNj cm9sbGJhci0+Y29udHJvbFNpemUoKSA9PSBSZWd1bGFyU2Nyb2xsYmFyID8ga1RoZW1lTWVkaXVt U2Nyb2xsQmFyIDoga1RoZW1lU21hbGxTY3JvbGxCYXI7Ci0gICAgdHJhY2tJbmZvLmJvdW5kcyA9 IHNjcm9sbGJhci0+ZnJhbWVSZWN0KCk7Ci0gICAgdHJhY2tJbmZvLm1pbiA9IDA7Ci0gICAgdHJh Y2tJbmZvLm1heCA9IHNjcm9sbGJhci0+bWF4aW11bSgpOwotICAgIHRyYWNrSW5mby52YWx1ZSA9 IHNjcm9sbGJhci0+Y3VycmVudFBvcygpOwotICAgIHRyYWNrSW5mby50cmFja0luZm8uc2Nyb2xs YmFyLnZpZXdzaXplID0gc2Nyb2xsYmFyLT52aXNpYmxlU2l6ZSgpOwotICAgIHRyYWNrSW5mby5h dHRyaWJ1dGVzID0gMDsKLSAgICBpZiAoc2Nyb2xsYmFyLT5vcmllbnRhdGlvbigpID09IEhvcml6 b250YWxTY3JvbGxiYXIpCi0gICAgICAgIHRyYWNrSW5mby5hdHRyaWJ1dGVzIHw9IGtUaGVtZVRy YWNrSG9yaXpvbnRhbDsKLQotICAgIGlmICghc2Nyb2xsYmFyLT5lbmFibGVkKCkpCi0gICAgICAg IHRyYWNrSW5mby5lbmFibGVTdGF0ZSA9IGtUaGVtZVRyYWNrRGlzYWJsZWQ7CisgICAgTlNTY3Jv bGxlciogc2Nyb2xsZXIgPSBbW05TU2Nyb2xsZXIgYWxsb2NdIGluaXRXaXRoRnJhbWU6c2Nyb2xs YmFyLT5mcmFtZVJlY3QoKV07CisgICAgaWYgKHNjcm9sbGJhci0+Y2xpZW50KCktPmlzQWN0aXZl KCkpCisgICAgICAgIFtbW0Zha2VBY3RpdmVXaW5kb3cgYWx3YXlzQWN0aXZlV2luZG93XSBjb250 ZW50Vmlld10gYWRkU3VidmlldzpzY3JvbGxlcl07ICAgIAogICAgIGVsc2UKLSAgICAgICAgdHJh Y2tJbmZvLmVuYWJsZVN0YXRlID0gc2Nyb2xsYmFyLT5jbGllbnQoKS0+aXNBY3RpdmUoKSA/IGtU aGVtZVRyYWNrQWN0aXZlIDoga1RoZW1lVHJhY2tJbmFjdGl2ZTsKLQotICAgIGlmICghaGFzQnV0 dG9ucyhzY3JvbGxiYXIpKQotICAgICAgICB0cmFja0luZm8uZW5hYmxlU3RhdGUgPSBrVGhlbWVU cmFja05vdGhpbmdUb1Njcm9sbDsKLSAgICB0cmFja0luZm8udHJhY2tJbmZvLnNjcm9sbGJhci5w cmVzc1N0YXRlID0gc2Nyb2xsYmFyUGFydFRvSElQcmVzc2VkU3RhdGUoc2Nyb2xsYmFyLT5wcmVz c2VkUGFydCgpKTsKLQotICAgIENHQWZmaW5lVHJhbnNmb3JtIGN1cnJlbnRDVE0gPSBDR0NvbnRl eHRHZXRDVE0oY29udGV4dC0+cGxhdGZvcm1Db250ZXh0KCkpOwotCi0gICAgLy8gVGhlIEFxdWEg c2Nyb2xsYmFyIGlzIGJ1Z2d5IHdoZW4gcm90YXRlZCBhbmQgc2NhbGVkLiAgV2Ugd2lsbCBqdXN0 IGRyYXcgaW50byBhIGJpdG1hcCBpZiB3ZSBkZXRlY3QgYSBzY2FsZSBvciByb3RhdGlvbi4KLSAg ICBib29sIGNhbkRyYXdEaXJlY3RseSA9IGN1cnJlbnRDVE0uYSA9PSAxLjBmICYmIGN1cnJlbnRD VE0uYiA9PSAwLjBmICYmIGN1cnJlbnRDVE0uYyA9PSAwLjBmICYmIChjdXJyZW50Q1RNLmQgPT0g MS4wZiB8fCBjdXJyZW50Q1RNLmQgPT0gLTEuMGYpOwotICAgIEdyYXBoaWNzQ29udGV4dCogZHJh d2luZ0NvbnRleHQgPSBjb250ZXh0OwotICAgIE93blB0cjxJbWFnZUJ1ZmZlcj4gaW1hZ2VCdWZm ZXI7Ci0gICAgaWYgKCFjYW5EcmF3RGlyZWN0bHkpIHsKLSAgICAgICAgdHJhY2tJbmZvLmJvdW5k cyA9IEludFJlY3QoSW50UG9pbnQoKSwgc2Nyb2xsYmFyLT5mcmFtZVJlY3QoKS5zaXplKCkpOwot Ci0gICAgICAgIEludFJlY3QgYnVmZmVyUmVjdChzY3JvbGxiYXItPmZyYW1lUmVjdCgpKTsKLSAg ICAgICAgYnVmZmVyUmVjdC5pbnRlcnNlY3QoZGFtYWdlUmVjdCk7Ci0gICAgICAgIGJ1ZmZlclJl Y3QubW92ZSgtc2Nyb2xsYmFyLT5mcmFtZVJlY3QoKS54KCksIC1zY3JvbGxiYXItPmZyYW1lUmVj dCgpLnkoKSk7Ci0KLSAgICAgICAgaW1hZ2VCdWZmZXIgPSBJbWFnZUJ1ZmZlcjo6Y3JlYXRlKGJ1 ZmZlclJlY3Quc2l6ZSgpKTsKLSAgICAgICAgaWYgKCFpbWFnZUJ1ZmZlcikKLSAgICAgICAgICAg IHJldHVybiB0cnVlOwotCi0gICAgICAgIGRyYXdpbmdDb250ZXh0ID0gaW1hZ2VCdWZmZXItPmNv bnRleHQoKTsKKyAgICAgICAgW1tbRmFrZUFjdGl2ZVdpbmRvdyBhbHdheXNJbmFjdGl2ZVdpbmRv d10gY29udGVudFZpZXddIGFkZFN1YnZpZXc6c2Nyb2xsZXJdOworICAgIFtzY3JvbGxlciBzZXRF bmFibGVkOnNjcm9sbGJhci0+ZW5hYmxlZCgpXTsKKyAgICAKKyAgICBbc2Nyb2xsZXIgc2V0Q29u dHJvbFNpemU6c2Nyb2xsYmFyLT5jb250cm9sU2l6ZSgpID09IFJlZ3VsYXJTY3JvbGxiYXIgPyBO U1JlZ3VsYXJDb250cm9sU2l6ZSA6IE5TU21hbGxDb250cm9sU2l6ZV07CisgICAgCisgICAgTlNH cmFwaGljc0NvbnRleHQqIHByZXZpb3VzR3JhcGhpY3NDb250ZXh0ID0gW05TR3JhcGhpY3NDb250 ZXh0IGN1cnJlbnRDb250ZXh0XTsKKyAgICBOU0dyYXBoaWNzQ29udGV4dCogbnNHcmFwaGljc0Nv bnRleHQgPSBbTlNHcmFwaGljc0NvbnRleHQgZ3JhcGhpY3NDb250ZXh0V2l0aEdyYXBoaWNzUG9y dDpjb250ZXh0LT5wbGF0Zm9ybUNvbnRleHQoKSBmbGlwcGVkOllFU107CisgICAgW05TR3JhcGhp Y3NDb250ZXh0IHNldEN1cnJlbnRDb250ZXh0Om5zR3JhcGhpY3NDb250ZXh0XTsKKworICAgIC8v IERlc3BpdGUgcGFzc2luZyBpbiBmcmFtZVJlY3QoKSB0byB0aGUgc2Nyb2xsZXIsIGl0IGFsd2F5 cyBkcmF3cyBhdCAoMCwgMCkuCisgICAgLy8gRm9yY2UgaXQgdG8gZHJhdyBpbiB0aGUgcmlnaHQg bG9jYXRpb24gYnkgdHJhbnNsYXRpbmcgdGhlIHdob2xlIGdyYXBoaWNzCisgICAgLy8gY29udGV4 dC4KKyAgICBbbnNHcmFwaGljc0NvbnRleHQgc2F2ZUdyYXBoaWNzU3RhdGVdOworICAgIE5TQWZm aW5lVHJhbnNmb3JtICp0cmFuc2Zvcm0gPSBbTlNBZmZpbmVUcmFuc2Zvcm0gdHJhbnNmb3JtXTsK KyAgICBbdHJhbnNmb3JtIHRyYW5zbGF0ZVhCeTpzY3JvbGxiYXItPmZyYW1lUmVjdCgpLngoKSB5 Qnk6c2Nyb2xsYmFyLT5mcmFtZVJlY3QoKS55KCldOworICAgIFt0cmFuc2Zvcm0gY29uY2F0XTsK KyAgICAKKyAgICBkb3VibGUgdmFsdWUgPSBkb3VibGUoc2Nyb2xsYmFyLT5jdXJyZW50UG9zKCkp L2RvdWJsZShzY3JvbGxiYXItPm1heGltdW0oKSk7CisgICAgW3Njcm9sbGVyIHNldERvdWJsZVZh bHVlOiB2YWx1ZV07CisKKyAgICBmbG9hdCBrbm9iUHJvcG9ydGlvbiA9IGZsb2F0KHNjcm9sbGJh ci0+dmlzaWJsZVNpemUoKSkvZmxvYXQoc2Nyb2xsYmFyLT50b3RhbFNpemUoKSk7CisgICAgW3Nj cm9sbGVyIHNldEtub2JQcm9wb3J0aW9uOiBrbm9iUHJvcG9ydGlvbl07CisgICAgCisgICAgW3Nj cm9sbGVyIGRyYXdLbm9iU2xvdEluUmVjdDpbc2Nyb2xsZXIgcmVjdEZvclBhcnQ6TlNTY3JvbGxl cktub2JTbG90XSBoaWdobGlnaHQ6Tk9dOworICAgIAorICAgIC8vIEZJWE1FOiBjaGVjayB0aWNr bWFya3MKKyAgICAvLyBGSVhNRTogY2hlY2sgdHJhbnNmb3JtcworICAgIC8vIEZJWE1FOiBjaGVj ayBoaWdobGlnaHRpbmcKKyAgICAKKyAgICBpZiAoaGFzQnV0dG9ucyhzY3JvbGxiYXIpKSB7Cisg ICAgICAgIFtzY3JvbGxlciBkcmF3QXJyb3c6TlNTY3JvbGxlckluY3JlbWVudEFycm93IGhpZ2hs aWdodDpzY3JvbGxiYXItPnByZXNzZWRQYXJ0KCkgPT0gRm9yd2FyZEJ1dHRvblN0YXJ0UGFydF07 CisgICAgICAgIFtzY3JvbGxlciBkcmF3QXJyb3c6TlNTY3JvbGxlckRlY3JlbWVudEFycm93IGhp Z2hsaWdodDpzY3JvbGxiYXItPnByZXNzZWRQYXJ0KCkgPT0gQmFja0J1dHRvblN0YXJ0UGFydF07 CiAgICAgfQogCi0gICAgLy8gRHJhdyB0aHVtYmxlc3MuCi0gICAgSElUaGVtZURyYXdUcmFjaygm dHJhY2tJbmZvLCAwLCBkcmF3aW5nQ29udGV4dC0+cGxhdGZvcm1Db250ZXh0KCksIGtISVRoZW1l T3JpZW50YXRpb25Ob3JtYWwpOwotCisgICAgLy8gRklYTUU6IGFyZSB0aGVzZSBuZWNlc3Nhcnk/ CisgICAgR3JhcGhpY3NDb250ZXh0KiBkcmF3aW5nQ29udGV4dCA9IGNvbnRleHQ7CisgICAgYm9v bCBjYW5EcmF3RGlyZWN0bHkgPSB0cnVlOwogICAgIFZlY3RvcjxJbnRSZWN0PiB0aWNrbWFya3M7 CiAgICAgc2Nyb2xsYmFyLT5jbGllbnQoKS0+Z2V0VGlja21hcmtzKHRpY2ttYXJrcyk7CiAgICAg aWYgKHNjcm9sbGJhci0+b3JpZW50YXRpb24oKSA9PSBWZXJ0aWNhbFNjcm9sbGJhciAmJiB0aWNr bWFya3Muc2l6ZSgpKSB7CkBAIC00NDYsMTUgKzQ2OCwxNiBAQCBib29sIFNjcm9sbGJhclRoZW1l Q2hyb21pdW1NYWM6OnBhaW50KFNjcm9sbGJhciogc2Nyb2xsYmFyLCBHcmFwaGljc0NvbnRleHQq IGNvbgogCiAgICAgICAgIGRyYXdpbmdDb250ZXh0LT5yZXN0b3JlKCk7CiAgICAgfQotCi0gICAg aWYgKGhhc1RodW1iKHNjcm9sbGJhcikpIHsKLSAgICAgICAgdHJhY2tJbmZvLmF0dHJpYnV0ZXMg fD0gKGtUaGVtZVRyYWNrU2hvd1RodW1iIHwga1RoZW1lVHJhY2tIaWRlVHJhY2spOwotICAgICAg ICBISVRoZW1lRHJhd1RyYWNrKCZ0cmFja0luZm8sIDAsIGRyYXdpbmdDb250ZXh0LT5wbGF0Zm9y bUNvbnRleHQoKSwga0hJVGhlbWVPcmllbnRhdGlvbk5vcm1hbCk7Ci0gICAgfQotCi0gICAgaWYg KCFjYW5EcmF3RGlyZWN0bHkpCi0gICAgICAgIGNvbnRleHQtPmRyYXdJbWFnZUJ1ZmZlcihpbWFn ZUJ1ZmZlci5nZXQoKSwgQ29sb3JTcGFjZURldmljZVJHQiwgc2Nyb2xsYmFyLT5mcmFtZVJlY3Qo KS5sb2NhdGlvbigpKTsKLQorICAgIAorICAgIGlmIChoYXNUaHVtYihzY3JvbGxiYXIpKQorICAg ICAgICBbc2Nyb2xsZXIgZHJhd0tub2JdOworCisgICAgW3Njcm9sbGVyIHJlbGVhc2VdOworICAg IAorICAgIFtuc0dyYXBoaWNzQ29udGV4dCByZXN0b3JlR3JhcGhpY3NTdGF0ZV07ICAgIAorICAg IFtOU0dyYXBoaWNzQ29udGV4dCBzZXRDdXJyZW50Q29udGV4dDpwcmV2aW91c0dyYXBoaWNzQ29u dGV4dF07CisgICAgCisgICAgCiAgICAgcmV0dXJuIHRydWU7CiB9CiAK