5337 2013-10-11 09:33:38 +0000 [QA] Optimize test case for better use of time and resources 2013-12-10 09:39:24 +0000 1 1 10 QA/Testing Test Plans/Suite test-plans/suite 1.6 x86 Multiple VERIFIED FIXED Medium enhancement 1.6 1 corneliux.stoicescu ionutx.chisanovici alexandru.c.georgescu corneliux.stoicescu ionutx.chisanovici jessica.zhang --- --- No (bug/feature does not impact docs) oldest_to_newest 37826 0 corneliux.stoicescu 2013-10-11 09:33:38 +0000 Looking over many of our test cases for oe-core, meta-yocto and bitbake, we can identify many tests that can be optimized by smartly selecting the build targets. An example is using 'core-image-minimal' instead of 'core-image-sato' as the build target. Another example is using a package with few build time dependencies instead of one with many build time dependencies(example: 'binutils-cross' instead of 'man') Another thing we can do is use buildhistory to analyze image content instead of booting the image. This is a proposal and I believe we can save both man/hour and autobuilder resources. 38116 1 corneliux.stoicescu 2013-10-21 14:53:59 +0000 Started work on feature. 38194 2 corneliux.stoicescu 2013-10-28 10:26:28 +0000 The modifications are done. I have sent them to QA team for review. Affected test components: bitbake, oe-core, meta-yocto, HOB. The optimizations made consist of changing the build target for the test case. Also in some cases I had to modify the steps a little bit in order to accomodate the new targets (like replacing "1. build core-image-sato" with "1. add ssh-server-openssh to EXTRA_IMAGE_FEATURES in local.conf; 2. build core-image-minimal"). Except for HOB, I re-run all modified test cases in order to assure results integrity. For HOB, the changes made are not that complex and the result is theoretically the same. The new build targets have been chosen as follow: * m4-native - build time: 1 min - chosen as the fastest build-able recipe/target - used in test cases that don't require a rpm/deb/ipk output package * xcursor-transparent-theme - build time: 8 min - chosen as the fastest build-able recipe/target that has a rpm/deb/ipk package output - is architecture-independent * m4 -build time: 17 min - chosen as the fastest build-able recipe/target that has a pm/deb/ipk package output and is architecture-dependent * core-image-minimal - build time: 40 min - chosen as the fastest build-able image type - in tests that require booting I added the 'ssh-server-openssh" image feature to facilitate automated runtime testing * core-image-basic - build time: 53 min - takes longer to build than core-image-minimal; has more packages included - contains open-ssh by default - used in tests that require removing packages from the base image or additional package functionality (syslogd) * core-image-sato - build time: 80 min - smallest image that contains an x server and a desktop environment - used in tests that have graphical requirements (ex: emgd-driver-bin) * core-image-sato-sdk - build time: 148 min - I did not use this image in my optimizations - this image is still used in some unmodified test cases that require its -sdk functionality There are a few other chosen build targets for the special context of the test case. After I correct any issues signaled by my team upon review I will set this to RESOLVED. 38195 3 1583 corneliux.stoicescu 2013-10-28 10:27:58 +0000 Created attachment 1583 Optimizations table Here is the optimizations table. 38431 4 corneliux.stoicescu 2013-11-01 16:57:01 +0000 The changes are made and review complete. 38678 5 alexandru.c.georgescu 2013-11-08 08:42:41 +0000 Hi Ionut, Please verify this! 39058 6 ionutx.chisanovici 2013-11-25 10:19:24 +0000 Till now the oe-core weekly TC's were verified and they are ok. We still need to verify the full pass TC's for oe-core and bsp's but this will be done in the next full pass run. Ionut C 39313 7 ionutx.chisanovici 2013-12-10 09:39:24 +0000 All the test cases were verified. The spent time on a full pass has decreased considerable. Cheers, Ionut C 1583 2013-10-28 10:27:58 +0000 2013-10-28 10:27:58 +0000 Optimizations table Test Optimizations.xlsx application/vnd.openxmlformats-officedocument.spreadsheetml.sheet 12532 corneliux.stoicescu UEsDBBQABgAIAAAAIQCkBM/pcwEAAJgFAAATAAgCW0NvbnRlbnRfVHlwZXNdLnhtbCCiBAIooAAC AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADE lM9OAjEQxu8mvsOmV7Nb4GCMYeGAelQS8QFqO8s2dNumUxDe3tnyJ4asECKJl2227XzfbyadGY7X jclWEFA7W7J+0WMZWOmUtvOSfcxe8geWYRRWCeMslGwDyMaj25vhbOMBM4q2WLI6Rv/IOcoaGoGF 82DppHKhEZF+w5x7IRdiDnzQ691z6WwEG/PYarDR8AkqsTQxe17T9pYkgEGWTbYXW6+SCe+NliIS KV9ZdeSS7xwKikx3sNYe7wiD8U6H9uR3g13cG5UmaAXZVIT4KhrC4GvDv1xYfDq3KE6LdFC6qtIS lJPLhipQoA8gFNYAsTFFWotGaLvnPuGfLiNPS//KIG1+SfhCjsE/cUR6d8DT9++lSDJnEse4MYBX znYres65FgHUewzUoVcH+Kl9hkMKIyc1PdUrF+Gge8qf+mcanEeaJAEuB9iPijY69yQEIWo4DIuu pjs40hS63PCo66GdcwpUhzdPc3X0DQAA//8DAFBLAwQUAAYACAAAACEAtVUwI/UAAABMAgAACwAI Al9yZWxzLy5yZWxzIKIEAiigAAIAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAIySz07DMAzG70i8Q+T76m5ICKGlu0xIuyFUHsAk7h+1jaMkQPf2 hAOCSmPb0fbnzz9b3u7maVQfHGIvTsO6KEGxM2J712p4rZ9WD6BiImdpFMcajhxhV93ebF94pJSb Ytf7qLKLixq6lPwjYjQdTxQL8exypZEwUcphaNGTGahl3JTlPYa/HlAtPNXBaggHeweqPvo8+bK3 NE1veC/mfWKXToxAnhM7y3blQ2YLqc/bqJpCy0mDFfOc0xHJ+yJjA54m2lxP9P+2OHEiS4nQSODz PN+Kc0Dr64Eun2ip+L3OPOKnhOFNZPhhwcUPVF8AAAD//wMAUEsDBBQABgAIAAAAIQD+aepXCgEA AMwDAAAaAAgBeGwvX3JlbHMvd29ya2Jvb2sueG1sLnJlbHMgogQBKKAAAQAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAC8k09LxDAQxe+C3yHM3aatuohsugdF2KuuHyCk06Zsm5TM+Kff3lCxdWGp l+Jx3pD3fjwy291n14p3DNR4pyBLUhDojC8bVyt4PTxd3YEg1q7UrXeoYECCXXF5sX3GVnN8RLbp SUQXRwosc38vJRmLnabE9+jipvKh0xzHUMtem6OuUeZpupHhtwcUJ55iXyoI+/IaxGHoY/Lf3r6q GoOP3rx16PhMhOTIhdFQhxpZwTh+i1kSQUGeZ8jXZPjw4UgWkWeOSSI5bvIlmOyfYRab2awJY3Rr Hqxu3NzMJC01crsmBFkdsHzhEC+AZpATeQnmZlUYHtp4cNOHpXH+iZcnN1h8AQAA//8DAFBLAwQU AAYACAAAACEAJpXwGWMBAABVAgAADwAAAHhsL3dvcmtib29rLnhtbIySy27CMBBF95X6D5b3xYkJ iCISpKqtyqZCKoW1G0+IhWNHtiHw9x0H8ZC66WqeOTNzndn82GhyAOeVNTlNBwklYEorldnm9Hv1 /jShxAdhpNDWQE5P4Om8eHyYddbtfqzdEQQYn9M6hHbKmC9raIQf2BYMVirrGhEwdFvmWwdC+hog NJrxJBmzRihDz4Sp+w/DVpUq4dWW+wZMOEMcaBFwfV+r1tNiVikN6/NFRLTtp2hw76OmRAsf3qQK IHM6wtB2cEtklLh9+7JXGqvPw2RIWXE9cukIUgO4pVMHUZ5QKUokVGKvwwoPvszDPM84H8dvozhr BZ2/YWJIjhtlpO1yyjMU+3SJ0gRX6vrSRslQIyqb3HIfoLZ1yOkkSZNIZ3f4XlIc01ti+nu/oswc 3y7aBZ6EvpsqdNxCppHwp3t4143+tZv38/p2HFIKXaIc0URsmo3GvOexyx9R/AIAAP//AwBQSwME FAAGAAgAAAAhADCbtd4QAwAANQgAAA0AAAB4bC9zdHlsZXMueG1srFXfb9owEH6ftP/B8jskQGgJ SlKNUrRKW1WpnbRXkzhg4R+R7XSwaf/7zk4CVKVbN/aS2Of77rvvfLaTq63g6Ilqw5RM8aAfYkRl rgomVyn+8rjoTTAylsiCcCVpinfU4Kvs/bvE2B2nD2tKLYIQ0qR4bW01DQKTr6kgpq8qKmGlVFoQ C1O9CkylKSmMAwkeDMPwIhCESdxEmIr8LUEE0Zu66uVKVMSyJePM7nwsjEQ+vV1JpcmSQ6rbQUTy LrafvAgvWK6VUaXtQ7hAlSXL6css4yAOIFKWlEpag3JVS5viEYR2DNONVN/kwi1BAVuvLDHf0RPh YBngIEtyxZVGFioDiXmLJII2HteEs6Vmzq0kgvFdYx46gy9m6ycYSHPGwOXRZJMlS+fVcY3c8jMu b/lfXPUx1wldoaM/n8zrMyCQcb4v99BVFgxZAttuqZYLmKB2/LiroK4SOrSpj/f7g/dKk91gOD4C BJ4QSqp0ASei22jH3JiyhNPSgkbNVmv3t6qC71JZC+2TJQUjKyUJh2HQIbr/b5BwwOAspdiuWb4B smcbGGFkmeu3sB/FcTyJLqPwMhoPL3x/AI3jPk3d5gCVzCnnD47ka7mXNQKmbYlkLRbC3hZAgJFr rG4INWyHjYRmAtJeAw0AfwoEdlJVfHdXiyXVC38feDZvddt4mH3gbCUF9Xpb2L1WlubW30++wV7j H77CD7rezD/zW/93+QTH1W1qfXaZ0bY8p94N2p/Qf6hWg276y2kDNUct9KyB9tKRO/Yp/gj3Ozwd CPaiywEta8ahhZ2ewaW/oLpu7EBwejVnctOBAH0EmpzC3Lk24h0AtvgI4LvksCuQfLE99L1fte6J 8CdiLwdiFLQkNbeP+8UUH8afacFqAam1XvfsSVkfIsWHceMVu5Tp1n4ycE/DH9WapfjHzewynt8s hr1JOJv0ohEd9+LxbN4bR9ez+XwRh8Pw+idocu/pFB6XM94r/67CWR1EU8PhVdOt2Db5h4MtxUeT T+5i830TQNpwjXUiArN/77NfAAAA//8DAFBLAwQUAAYACAAAACEAFdQiyToCAAAxIgAAIwAAAHhs L3dvcmtzaGVldHMvX3JlbHMvc2hlZXQxLnhtbC5yZWxzzFrBjpswFLxX6j8g3xuwTTBUS3LZVtpD L9X2HLngALsEI9vbbvr1takqdaVV23ebW1DCs4d5w5uB3ByfL3P2zTg/2aVlfFewzCyd7adlaNmX +4/vapb5oJdez3YxLbsaz46Ht29uPptZh3iSH6fVZ7HK4ls2hrC+z3Pfjeai/c6uZonfnK276BAP 3ZCvunvUg8lFUVS5+7MGO7yomd31LXN3fVz//rrGlf9d257PU2dubfd0MUt4ZYl8jJXcPC2Psah2 gwm/yvq4569Pw49pnvXuartgV2cfTBe2LQd38qP9fuq0N7tumI7pw2nqW6Gq33U+2T5u8cNzMG7R M8tfx8IlLhjJSyIYYGZUkza3Mfx/zIjEJGib8YIIRjbIYASNGWDJiJooGcGBianSlSZIRibwoJJR exqWMjUlKBbRKBqY9HNQLKpMCiA0GQcmRvJ0n6WAAWZGFkQwIkkMtM34pmcCMxJ4yigqMxLYmDVE xwwsf9EQpz8H9ph1QQQjYlCElX9FNMwC2TCTvQyyyeRELwOMRTREkwksf7UnpjIOTIwiDxngJMNL aixDHpk10WRKYMdMDf8l8sTcvAnBMAObf7UnKoYDg5E8dQ2BGAGcZHhJNZnASYYXxDaTwPO/JnbZ 1pWg4V/VRGK2xzigYERNNMwS2JgJ6lsZYF+mqnShCTdm4IgpauJTme1VIahi5BZNCMQIYMPMS6L8 BbBh5gWxzSSyY1bExCyBXab6SyzLX/zR4/ATAAD//wMAUEsDBBQABgAIAAAAIQB+wYrlYAEAAHQC AAAYAAAAeGwvd29ya3NoZWV0cy9zaGVldDIueG1sjJLBasMwDIbvg72D8b1x2q3bGpKUQSnrYTDG trvjKIlpbAXbXdu3n5KQMuilNwlJn3/9cro+mZb9gvMabcbnUcwZWIWltnXGv7+2sxfOfJC2lC1a yPgZPF/n93fpEd3eNwCBEcH6jDchdIkQXjVgpI+wA0uVCp2RgVJXC985kOUwZFqxiOMnYaS2fCQk 7hYGVpVWsEF1MGDDCHHQykD6faM7P9GMugVnpNsfuplC0xGi0K0O5wHKmVHJrrboZNHS3qf5o1QT e0iu8EYrhx6rEBFOjEKvd16JlSBSnpaaNuhtZw6qjL/OucjTwZwfDUf/L2a91wXivi/syozHfau4 6t0OXn84VkIlD234xOMb6LoJdNglae9XSMrzBrwi7wgTLZaXRzcySKJ2soZ36WptPWuhGrqeOXMj Jo4oDtj1s8+ELDAENFPW0HWBrhhHD5xViGFKerWX/5L/AQAA//8DAFBLAwQUAAYACAAAACEA+2Kl bZQGAACnGwAAEwAAAHhsL3RoZW1lL3RoZW1lMS54bWzsWU9v2zYUvw/YdyB0b20nthsHdYrYsZut TRvEboceaZmWWFOiQNJJfRva44ABw7phlwG77TBsK9ACu3SfJluHrQP6FfZISrIYy0vSBhvW1YdE In98/9/jI3X12oOIoUMiJOVx26tdrnqIxD4f0zhoe3eG/UsbHpIKx2PMeEza3pxI79rW++9dxZsq JBFBsD6Wm7jthUolm5WK9GEYy8s8ITHMTbiIsIJXEVTGAh8B3YhV1qrVZiXCNPZQjCMge3syoT5B Q03S28qI9xi8xkrqAZ+JgSZNnBUGO57WNELOZZcJdIhZ2wM+Y340JA+UhxiWCibaXtX8vMrW1Qre TBcxtWJtYV3f/NJ16YLxdM3wFMEoZ1rr11tXdnL6BsDUMq7X63V7tZyeAWDfB02tLEWa9f5GrZPR LIDs4zLtbrVRrbv4Av31JZlbnU6n0UplsUQNyD7Wl/Ab1WZ9e83BG5DFN5bw9c52t9t08AZk8c0l fP9Kq1l38QYUMhpPl9Daof1+Sj2HTDjbLYVvAHyjmsIXKIiGPLo0iwmP1apYi/B9LvoA0ECGFY2R midkgn2I4i6ORoJizQBvElyYsUO+XBrSvJD0BU1U2/swwZARC3qvnn//6vlT9Or5k+OHz44f/nT8 6NHxwx8tLWfhLo6D4sKX337259cfoz+efvPy8RfleFnE//rDJ7/8/Hk5EDJoIdGLL5/89uzJi68+ /f27xyXwbYFHRfiQRkSiW+QIHfAIdDOGcSUnI3G+FcMQU2cFDoF2CemeCh3grTlmZbgOcY13V0Dx KANen913ZB2EYqZoCecbYeQA9zhnHS5KDXBD8ypYeDiLg3LmYlbEHWB8WMa7i2PHtb1ZAlUzC0rH 9t2QOGLuMxwrHJCYKKTn+JSQEu3uUerYdY/6gks+UegeRR1MS00ypCMnkBaLdmkEfpmX6Qyudmyz dxd1OCvTeoccukhICMxKhB8S5pjxOp4pHJWRHOKIFQ1+E6uwTMjBXPhFXE8q8HRAGEe9MZGybM1t AfoWnH4DQ70qdfsem0cuUig6LaN5E3NeRO7waTfEUVKGHdA4LGI/kFMIUYz2uSqD73E3Q/Q7+AHH K919lxLH3acXgjs0cERaBIiemYkSX14n3InfwZxNMDFVBkq6U6kjGv9d2WYU6rbl8K5st71t2MTK kmf3RLFehfsPlugdPIv3CWTF8hb1rkK/q9DeW1+hV+XyxdflRSmGKq0bEttrm847Wtl4TyhjAzVn 5KY0vbeEDWjch0G9zhw6SX4QS0J41JkMDBxcILBZgwRXH1EVDkKcQN9e8zSRQKakA4kSLuG8aIZL aWs89P7KnjYb+hxiK4fEao+P7fC6Hs6OGzkZI1VgzrQZo3VN4KzM1q+kREG312FW00KdmVvNiGaK osMtV1mb2JzLweS5ajCYWxM6GwT9EFi5Ccd+zRrOO5iRsba79VHmFuOFi3SRDPGYpD7Sei/7qGac lMXKkiJaDxsM+ux4itUK3Fqa7BtwO4uTiuzqK9hl3nsTL2URvPASUDuZjiwuJieL0VHbazXWGh7y cdL2JnBUhscoAa9L3UxiFsB9k6+EDftTk9lk+cKbrUwxNwlqcPth7b6ksFMHEiHVDpahDQ0zlYYA izUnK/9aA8x6UQqUVKOzSbG+AcHwr0kBdnRdSyYT4quiswsj2nb2NS2lfKaIGITjIzRiM3GAwf06 VEGfMZVw42Eqgn6B6zltbTPlFuc06YqXYgZnxzFLQpyWW52iWSZbuClIuQzmrSAe6FYqu1Hu/KqY lL8gVYph/D9TRe8ncAWxPtYe8OF2WGCkM6XtcaFCDlUoCanfF9A4mNoB0QJXvDANQQV31Oa/IIf6 v805S8OkNZwk1QENkKCwH6lQELIPZclE3ynEauneZUmylJCJqIK4MrFij8ghYUNdA5t6b/dQCKFu qklaBgzuZPy572kGjQLd5BTzzalk+d5rc+Cf7nxsMoNSbh02DU1m/1zEvD1Y7Kp2vVme7b1FRfTE os2qZ1kBzApbQStN+9cU4Zxbra1YSxqvNTLhwIvLGsNg3hAlcJGE9B/Y/6jwmf3goTfUIT+A2org +4UmBmEDUX3JNh5IF0g7OILGyQ7aYNKkrGnT1klbLdusL7jTzfmeMLaW7Cz+Pqex8+bMZefk4kUa O7WwY2s7ttLU4NmTKQpDk+wgYxxjvpQVP2bx0X1w9A58NpgxJU0wwacqgaGHHpg8gOS3HM3Srb8A AAD//wMAUEsDBBQABgAIAAAAIQBWWDTwwAoAAG02AAAYAAAAeGwvd29ya3NoZWV0cy9zaGVldDEu eG1snFtdc+I6En3fqv0PFE+7DxPwBxBSSW4FG3unau+tWzv37j47YIJrALPYk8zsr9/ulrGkluRC 8xKCdLqt091qH3/w+Mv342H0Xl6aqj49jYO76XhUnjb1tjq9PY3//CP7dD8eNW1x2haH+lQ+jX+U zfiX57/+5fGjvnxt9mXZjsDDqXka79v2/DCZNJt9eSyau/pcnmBmV1+ORQtfL2+T5nwpiy0ZHQ+T cDqdT45FdRoLDw+XW3zUu121KdN68+1Ynlrh5FIeihbW3+yrc3P1dtzc4u5YXL5+O3/a1MczuHit DlX7g5yOR8fNw+e3U30pXg/A+3sQF5urb/piuD9Wm0vd1Lv2DtxNxEJNzsvJcgKenh+3FTDAsI8u 5e5p/BI85PFiPHl+pAD9uyo/GuX/UVu8fikP5aYtt5Cn8aitz/8sd21SHg5gHI9H/6vr45dNgYu9 nylff8MMAAYHMWmvdf0VvX8GP1NYR0NecR3Fpq3eS+ExiSJI/H9pafg/rGvSL0z9/7rIjBL9+2W0 LXfFt0P7r/rjH2X1tm9htXBkCtnD9kdaNhtIHBz6Lpyh1019ABfwd3SssAIh8MV3+vyotu0ere/i cLa4DwA/ei2bNqvQ53i0+da09fE/HarzJbyEnRf47LzMl3eLYLqMFrc7gQDQUuCzcxLfbgwJIWP4 7IyjQOExsHY4BlnCZ2e5mHmvfd45gc/r4aO72WIaURQHjr7oDOHzajhV1j0c/4lIJtVJWrTF8+Ol /hjBvoZsNecCu0TwAJ7x85o/UTc0QLUC/abdV5uvq1pk2VI40QzKdoN+X9Ax2QK4gdH35+nj5B3q c9MhViYi0BGJiZjpiNRExJEOWZuQuY7ITEQc65DcApFLmUAs+4BCYasBFbvsDoubgvdHfaaoWIIH u+gaPHDShy3Ul7LCAzyNVQRjnKhzC906DSkV9ywXa9UkYJOZsInZcI7Du+c0/JTBEneYYQWixQQb llJkg9wTAPfc79nqI1p9ELPxtWozZ5OZMJqxKOU4vKND7YtLuR2LZp9H0OwhWU1FPTiNPmVgSOyW joxDI7mZ3QrBevZYqSWA6Pnz7MX27GkmrF4yYcNqPsdRjb0gDG3//Xkhl6SlEcNyaxpXCNaJLvVk Jrglrr2B11yKe0ErKOoqa82GtYtM2LD05zjqYhrJYGlMoTnfzhTBOtNAuqVlJwCRVFkhpnOiGjI2 a81GZoQcZsKGmeQ46qIayu2rUcWmf3NSxRlCaxZyUwiqAHFTXdipajacqrDhVHHUlyoK5pupIphl le2hBE+IfQHzrN7bqWo2jFQmbHgB46iLaiAPq2V16UMVwYwqazwJQNxUl3aqmg2nKmw4VRz1pRrA hdHtaSU0I8tWkSDGzRZmsTUZ+1W34nw7K3aonIa9GaMaubmQA6FdtE3LWzFiesZKp6AtncKsnbFm ZTAWVgZjHPZmDKu/mXESqFxNNkLOmPnTrAw2nXDSz2E5HOon2PjooVWAaL1iQ35+QcxA/oQCMhlr VgZjYWXkz9BNinJwtaPASyMRmjHmpxnESMbsnJHCLFWsPOdRJa91K9k7xUm1s2KByGnYWbFyK2kN GK+sb67YF0RLNuw8syJfLB68QeseGO8UZu07WDsuI551VkYFDCkqZwV4SarA1FQhW0aCmD5mLBwp TCJh4zpHM+KiM+uslMsYKoycxl0VoKD1CvBSVoEprUJZWrSMBDE948jIcSeuWG9Ya1a8B2QwS4Hi fW1IXwVyu+mMvQRWYCqsiFVggpiesZHjTi6xQKw1IzPHwkrJWpfjIZmloHXGXjorMIWWcgHS5VhV TQbjTjUZjFUjk7GwUjh0jIfUloLWGIdeaovQeueKWNdNEOPMMUxa97FmZDDurBQOgjGN++7j0Ett EZoxlttF5BgxPWPW1lKYRMbmPQ7NasF2StaZ8TsjOY27KAdTWUl6ln301ipENOPMz9eI6TlH7AyX wiySjpnVWrNasMrJHFY5jeOdKbw1Be2Q7t64mHppsdDUYjHLRIKYnmnMunEKs8iUq9K1biUXSxWT dVYBawg5jbuyK8Ol59ZHi72Eqs6K2PFXOMszzwo60T2wc1oKs/Z4qMflGznrrMx4DN3WcsXDR6mt 8FEHYxyzak4Q01cAC1lKd3ktykQzMk7TVyt2mqZhV/4Xsvb0AvCSYqEpxWLGKUGMm7BDimlGZgsX VjzzOVi5L7gUtM7YS4qFphTjbTVBjJtxp6nYLl7rRmw2g1nqgWw8p3FXjp2MvaRYaEqxmG3UBDFu xg4pphsxZhnM2hn/jBQLvaQYofVT1owtL0GMm7FDiulGzGUGs3bGPyPFIi8pRmjGmJ+6EONkDJO4 duOSSjMy9nFnpdSpkGI07lvVkZcUIzRjLFuikGKI6Rkr1640m8IsUubdeK1bsShmnRU7D+Y07GKs HFrrXJGXECM0Y8zFJ2IGGDtujulWBmP7zTEwcvdqJ2MvQUbvQjDG8jzf5VgVZCwrKTjAFJtyGz33 cZqzbph1ZmZZD94dm8qj61n2kWSryBRdMyacE8T0q58xgZLCrK0NrTUrzi3rrAzRReOuwg5lNnTK XqoL3ivgqmvGRQhiespsMqUXE2zNSzUynxyL22bmo+OhO2ChPLbO2Et2RabsUh7LdoU9JLvAgb1d q0Zmu3bILvLmyrFSKTpjH9n1EqmSaiZ3CnFd4SzT3TO2KRPNg8xC18sdokwzMuPhEGVg5W5tznh4 ibLIFGVzpicSxPQ1z8s3hVksAV6+a83KpOxQZeTNVQLObu6lyiJE6918zs41CWIGKAuBZVJWrUzK DlkGx3Jn2UU59pJlhGaUuUhBjJsyzFqzrFkZlDsrpVRpm+Q07pvl2EuXEZpRlmcJ0doQM0BZ6DIj y5qVSVlYmZRx3JuylzCLEc0oc2GGmJ4y714wae3mmhFvAFlnxeOU07iLsev8FXsJM0IzxlykIMbN WAgz43JDMzJzLKzMHA/pMgWtnb9iL1lGaMaYCa8EMW7GQpWZjFUjM8fCyszx0K0wZ469VBm9J8cY s8JNEONmLPSVyVg1MnMsrJSsda1rSJUpaD3HXhoFXgNX2LB6TmEWGsnzlz9//VsaPqTx/O/dvedw rjwyo8VmCjYLHzKJDUJ+uYkvn1/95iG8ntj7DaKlbJyClngVXLziu/9xLi+H6vQVXuPu/xevOa7w gu7yUG2fxpfP2wBf0TYQ0IJ7RGhFwHmpR9CL6NwH3g3pEbHNB94h6hH02rnhA2LeI+ZWHyBpewS9 qG/4gILqEfdWH7DJesTSioBu1SPgvR1LyPBxhYRYo4pPcSTEGlZ8tiUh1rjik0QJsQYWn4JKiDWy +GRYQqyhxeflEmKNLb55ICHW4OK1tIRYo4tvZvQQiIAluviSi4RYo4vvMUmINbr4mpaEWKOrBhee NFiWosYW7sFbEGpo4Z61BaFGFu7xWhBqYGGfWBBa0KxxVSsS7rlZfOBFfx8QuEllg6grgc5hg6iE Id82iBo1OKoNooYertktEJS1crnWyKIMlBBraFE3SYg9tmq5wZWAbS1q6GBhNogaOliYBYKX+v1a YGEAmfQNDFr3uXgrfy0ub9WpGR3gh0X4Cx2MtvgJD/0PPzmiUTjYa93Cb2+u3/bws64Sfo4xvQM2 u7pur1/wIP0PxZ7/DwAA//8DAFBLAwQUAAYACAAAACEAQbz84icDAADSEwAAFAAAAHhsL3NoYXJl ZFN0cmluZ3MueG1srFjJbtswEL0X6D8Qurg9yFq8SA5sB0jQoD0UKQoXPQY0RVtsJFIlR4nTry9l J4VNZhFK3mxS8/iGM5x55Px8V1fojkrFBF8EyTAOEOVEFIxvF8GP1VWYB0gB5gWuBKeL4IGq4Hz5 /t1cKUDalqtFUAI0Z1GkSElrrIaioVzPbISsMei/chupRlJcqJJSqKsojeNpVGPGA0REy0GvO04D 1HL2u6WXh5FsGiznii3nsPwmRdESmEewnEfd0GF4RTWDS6yoOXHBYI1vreGOpdI01+32D6sqPHwQ BEQjxS9KYM8S5I0qxf0N0aBDsmXn3Y8bViySLDUXcUKbmGjfaVNhQgsEJVPm5E8G5bMT65ZV2gTL LQU0IELSkNV4S0OFQQxMmFe/DlVxa1m4+BiPeq9fM65pV9byp4R3pJVKyBAk5qrBknIIQSccteyc aE9fp11jbq13yrMehxwDu/PMKzd5uXg59prOYyvUTtzGXj3NDLSuZp3p/CG6lumipKi8o8HSCKEO MvoPu5dT1ERbuuzQ1NrvU/rrQ/1DIUEbCqREu72XMtxt8mlIBN+w7RtJ/ATRx9LJFSutv1LA4b4y GwFw2bJsZq3jwHoUzzxyy2Mz4fukqG5hjZWjLls0SmKPTo2SxCuauUVunvoMXzZ7o2McteWnRofu u44+eK6XXFN9QKVP7ZLmJsE++VVRpaz86mNYjy0zl2BluVns+pBYt+phLXY2ky+cAcMVeiyYrKYf dFA+mrl6xfjbH31SoGULaNG20jhI4TtaPIvmUGrSzAye02aOfZ7KbOJTRWQTM9BOnk59eprmPutP OvOKlvdTOC5JOJv41C+p3YxP9UtD5cbSJy78c6vgf76+MM+8wwKZV0GQWbdDF2qW6w5guU+J4PVM 2U3Ywc0k9lnYksRnCU8S87gbd5cjubHGihHrIJ2etaPPDXXSveIoVZr2fdrv66Bhh2oVlL60QIhK hfrVZsMqOkD6YQoNNGD4eMV5gbWxSWvGW2Aa53BTR//+Eym07jFLQ19u3eJ770wAh2RMZ1bAVwK0 OumkRgtUHYmXSL/JLf8CAAD//wMAUEsDBBQABgAIAAAAIQBvRcVhQAEAAFECAAARAAgBZG9jUHJv cHMvY29yZS54bWwgogQBKKAAAQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAB8klFLwzAUhd8F /0PJe5ukkzlD24HKnhwIVia+heRuKzZpSKLd/r1pu9UOhpCX3HPud08uyZYHVUc/YF3V6BzRhKAI tGhkpXc5ei9X8QJFznMted1oyNERHFoWtzeZMEw0Fl5tY8D6ClwUSNoxYXK0994wjJ3Yg+IuCQ4d xG1jFffhanfYcPHFd4BTQuZYgeeSe447YGxGIjohpRiR5tvWPUAKDDUo0N5hmlD85/Vglbva0CsT p6r80YQ3neJO2VIM4ug+uGo0tm2btLM+RshP8cf65a1/alzpblcCUJFJwYQF7htbZHh6CYurufPr sONtBfLxGPQrNSn6uAMEZBQCsCHuWdnMnp7LFSq6HcbkIabzkhDWn89u5EV/F2goqNPgf4l0FlMS p4uSEpbes/RuQjwDhtyXn6D4BQAA//8DAFBLAwQUAAYACAAAACEAi6Q+jxQBAADwAwAAEAAAAHhs L2NhbGNDaGFpbi54bWx0001ugzAQBeB9pd7Bmn1jZkzTH2GyaBUu0B7AAicggUEYVcnta1Uhkfra JY/n4WMkF7vT0KsvP8duDJZ4k5HyoR6bLhwtfX7sH55JxcWFxvVj8JbOPtKuvL8ratfXb63rgkoT QrTULsv0qnWsWz+4uBknH9KbwzgPbkmP81HHafauia33y9BrybKtHtIAKotazZbe8ydSnSUh1ScK 6Uu+/yevJDX/6Ff54yVfJ1R5DomBZJ12O8XQSev5+eK1Y14gSRv71Vn/63ZqCx0wGzAbMBswGzAb MAuYBcwCZgGzgFnALGAWEAoIGYQMQgYhg5BByCBkEDJslcHMYAYyiAEMXuCCFrBgxduir3e1/AYA AP//AwBQSwMEFAAGAAgAAAAhAOCMII+OAQAALAMAABAACAFkb2NQcm9wcy9hcHAueG1sIKIEASig AAEAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAnJJBb9swDIXvA/YfDN0buWkxDIGsokhX9NBh AZK2Z1WmY6GyZIiskezXj7aXxGkPA3Yj+R6ePlFSN7vGZx0kdDEU4nKWiwyCjaUL20I8be4vvosM yYTS+BigEHtAcaO/flGrFFtI5AAzjghYiJqoXUiJtobG4IzlwEoVU2OI27SVsaqchbto3xsIJOd5 /k3CjiCUUF60x0AxJi46+t/QMtqeD583+5aBtbptW++sIb6l/ulsihgryn7sLHglp6JiujXY9+Ro r3Mlp61aW+NhycG6Mh5BydNAPYDpl7YyLqFWHS06sBRThu43r20usleD0OMUojPJmUCM1dvGZqh9 i5T0S0xvWAMQKsmGcTiUU++0dtd6Phi4ODf2ASMIC+eIG0ce8Fe1Mon+RTwwjLwjzrrnG8/8y3ck HaSrT+jDNhjiw7HL2LQm7Fk4Vo8uvOFTu4l3huCw6fOhWtcmQcmPc9BPA/XAS06+D1nWJmyhPHg+ C/2/eB4/v768nuVXOT/5ZKbk6ZvrPwAAAP//AwBQSwECLQAUAAYACAAAACEApATP6XMBAACYBQAA EwAAAAAAAAAAAAAAAAAAAAAAW0NvbnRlbnRfVHlwZXNdLnhtbFBLAQItABQABgAIAAAAIQC1VTAj 9QAAAEwCAAALAAAAAAAAAAAAAAAAAKwDAABfcmVscy8ucmVsc1BLAQItABQABgAIAAAAIQD+aepX CgEAAMwDAAAaAAAAAAAAAAAAAAAAANIGAAB4bC9fcmVscy93b3JrYm9vay54bWwucmVsc1BLAQIt ABQABgAIAAAAIQAmlfAZYwEAAFUCAAAPAAAAAAAAAAAAAAAAABwJAAB4bC93b3JrYm9vay54bWxQ SwECLQAUAAYACAAAACEAMJu13hADAAA1CAAADQAAAAAAAAAAAAAAAACsCgAAeGwvc3R5bGVzLnht bFBLAQItABQABgAIAAAAIQAV1CLJOgIAADEiAAAjAAAAAAAAAAAAAAAAAOcNAAB4bC93b3Jrc2hl ZXRzL19yZWxzL3NoZWV0MS54bWwucmVsc1BLAQItABQABgAIAAAAIQB+wYrlYAEAAHQCAAAYAAAA AAAAAAAAAAAAAGIQAAB4bC93b3Jrc2hlZXRzL3NoZWV0Mi54bWxQSwECLQAUAAYACAAAACEA+2Kl bZQGAACnGwAAEwAAAAAAAAAAAAAAAAD4EQAAeGwvdGhlbWUvdGhlbWUxLnhtbFBLAQItABQABgAI AAAAIQBWWDTwwAoAAG02AAAYAAAAAAAAAAAAAAAAAL0YAAB4bC93b3Jrc2hlZXRzL3NoZWV0MS54 bWxQSwECLQAUAAYACAAAACEAQbz84icDAADSEwAAFAAAAAAAAAAAAAAAAACzIwAAeGwvc2hhcmVk U3RyaW5ncy54bWxQSwECLQAUAAYACAAAACEAb0XFYUABAABRAgAAEQAAAAAAAAAAAAAAAAAMJwAA ZG9jUHJvcHMvY29yZS54bWxQSwECLQAUAAYACAAAACEAi6Q+jxQBAADwAwAAEAAAAAAAAAAAAAAA AACDKQAAeGwvY2FsY0NoYWluLnhtbFBLAQItABQABgAIAAAAIQDgjCCPjgEAACwDAAAQAAAAAAAA AAAAAAAAAMUqAABkb2NQcm9wcy9hcHAueG1sUEsFBgAAAAANAA0AVQMAAIktAAAAAA==