170358 2017-03-31 14:03:51 -0700 IndexedDB keys are gone or nonworking after upgrade to iOS 10.3 2017-04-04 07:32:16 -0700 1 1 1 Unclassified WebKit WebCore Misc. Safari 10 iPhone / iPad iOS 10 NEW InRadar P2 Major --- 1 qoregexp webkit-unassigned andrew beidson webkit-bug-importer oldest_to_newest 1293290 0 qoregexp 2017-03-31 14:03:51 -0700 IndexedDB data retrieval by key stops working if an existing database is upgraded to iOS 10.3. For example, for a table that has a numeric primary key and a unique string secondary key, attempts to retrieve records either by the primary key or the secondary key return undefined. I was able to determine that the data itself is still there by doing a retrieval of all data. The fields that are used as keys have values, but I don't know what the state of the key structures themselves is, whether they are gone or pointing to the wrong place now. I'm using Dexie so I can't give you the native indexedDB syntax, but it's easy to reproduce. If the database is created under iOS 10.3, there isn't a problem. It's something happening in the upgrade. 1293545 1 webkit-bug-importer 2017-04-02 11:16:10 -0700 <rdar://problem/31393946> 1293570 2 beidson 2017-04-02 18:04:08 -0700 (In reply to Meredith Anderson from comment #0) > IndexedDB data retrieval by key stops working if an existing database is > upgraded to iOS 10.3. For example, for a table that has a numeric primary > key and a unique string secondary key, attempts to retrieve records either > by the primary key or the secondary key return undefined. I was able to > determine that the data itself is still there by doing a retrieval of all > data. The fields that are used as keys have values, but I don't know what > the state of the key structures themselves is, whether they are gone or > pointing to the wrong place now. I'm using Dexie so I can't give you the > native indexedDB syntax, but it's easy to reproduce. If the database is > created under iOS 10.3, there isn't a problem. It's something happening in > the upgrade. There's definitely a new storage format for keys on disk, but we also definitely handle the only key format. I can demonstrate with a simple local test case that we do handle legacy key lookups fine. Even though you saw this with Dexie and therefore don't have the raw IDB, can you please include a minimal Dexie test case? I can find out the raw IDB from that. 1294031 3 306173 andrew 2017-04-04 07:32:16 -0700 Created attachment 306173 Test case using db.js indexeddb adapter This same issue has happened to me, using the db.js adapter. Based on the test case (attached), it looks like all of the table data is wiped if the table is set to NOT autoincrement ids. The other tables DO keep their data across the udpate. I have not found a way to retrieve the lost data after the update. In earlier versions of Safari, the autoincrement functionality was buggy, so many app developers worked around it by using their own IDs. The next step would be to test a non-autoincrement table using vanilla js. 306173 2017-04-04 07:32:16 -0700 2017-04-04 07:32:16 -0700 Test case using db.js indexeddb adapter 170358 - indexeddb.zip application/zip 8806 andrew UEsDBAoAAAAAAFlKhEoAAAAAAAAAAAAAAAAKABAAaW5kZXhlZGRiL1VYDABaq+NYSavjWPUBFABQ SwMEFAAIAAgA3EiESgAAAAAAAAAAAAAAAA8AEABpbmRleGVkZGIvZGIuanNVWAwA16njWH+o41j1 ARQA7T1bc9s21u/5FbQfImrCMLbytd2V62baJJ3Jdtt0N+nug0ZjUyIksaFIhaRsq67/+3cuAAmS 4EW2025bMzOxCAIHwMG5AgeAvdhG8yyII3sxvA4WdrbbiHhhiatNnGTp6enpYTz7Wcyzw8eP5ad1 7G9DcQBftpEvFkEk/MPhNae6qtzCHt6IMBVWAZLzIkRVJcDkRNdb+8Nr/m1Pps6CC19feIm1PClA XAaRH19Wq16ecnq1wmUYz7ywnpvTq7lTES7qeTGV27I8zVZBerN0/Zl1amH/boZ2jrwhtZVLOowL R+LiJBHZNomsPLMl7MyJnGR4naekduxsCf8H0SSe8q+MfiFc71Q2MhEft0EiyjiUiYing+3jx95Q VugB0IOjIaYHKi2QaQh1cRqJS+t1ksSJffjSi6I4s6D9vhxha3D4JH5yODgcnmSrJL60Fu489sXp 4fdvX/30z9dnP7x9f/bt259+eHXoLG4QXniKbT+9lv0eX9/cnGAfJkdTd+6FoR0q8nByvAnuYHRK GY+nEzFV+Ert6EU0FsMbJ3SKksJh3N3IXFil+kitCHrhagF9xtzx6dFJ/GXihiJaZquT+MmTYWon iPi8GTDM18fjSd5kCaIyysPrR4MtEFSaJcE8G5w8eoTgz9IwmAv/ffx1kng7pBs14vbQui7eKNub TCReBg3zksSxAsxAMOAVSk6mJ/wawUuWbIV89RGqBxQq3wW85zR8Ajl3WFGcWNTfswA+A8DJu916 FoduIOuc2kMHWntiHdhUgX2Wwv9ngRuJq8weDl0/jsTwpKgdW4ctczfbdAW53Qsv3GKOYGHZgfX4 MX9lvFrA9dihWSK8DyfWDfybe9l8ZdkiSQiSn3eKOgDJmA36AHSDHeBuIOwDaAJCDyaHPEKH02Hp zR6Wi2KhM39oMRGfCa5fji428qR4LYanGAMsT8PnBin9xW9D/KL4jEFIeWLZFdRaQWS9JRlaLWgc 9RzUtWwxMul7IGnJqG8iwHTgW16WifUms7IYxA5Q3XYOMIUVxdFTqnkGHBxEaeZFc3FIUG/gP0CO pEzJJYB2/sGtppEqGAYxXfpe9IoyppR4aL3QMAf6QuujLA2J2IRxYz54q1VmaAxkAykUcX9lK2Tm F0VzxuVqoct5tWdZ/BLKb9eIoGI4W8c55x5kniMHB3wEv/JMksqR+K0vrSIB3kGeAAAsMAmmkvWC qUZx+EkbcpnK7Vgk8ZoaQaP36FGhQewwBokKoB9Z8FQEDyZha9+8+uY7sfu3Fy2RpaiEq6f9+qtM vBSzD0GmfTrJYWSJF6UeVfo9CH8UClwnPsDNIBbC3dgaqJ8Dp/T1MgF6kZ/p94A+3xQVrLz07SUK FWYRd5PEWYyj5/KXH5N4I5JsV5QA0eZtw+x7bwMfdJFa+mBfDUstJbRenajqc2ig78SV8F99k6Oo SKkiqP5hHf9iSI1NOVNDYroK1qXkknrIW1/IAZYBg6KMVEb+YCi7Rgyeo2r20puvcPSvNZSnIrkQ yesLEWU4npOBNwMNNnCsgUDw+AO+p9CM+QqpYTCVIPPWBakUaDCiawOeMRm5VdlW9I6sOsg10wD4 VX8bK4nEMIecnW3PgexbpRXrOFrGr755Hyuy/TpZgiG1AWWsNQl7/EHs0oLE8I2zuSl03Jaowwdl AH7W9daxDk2DCAAxK1g3J7XPIKLhM9YxgTyGDJG3xmG5iEGSHzmlz9SOaB5u0+CiyFMGkV4GpD4B eLV1+Mw9EAgD8XEwrn3CR1Y+IIY9YaXcAGOZdcAI40uRfBPDSA7qMKpdIQ1vzNXaiLCrEVvk+E/c iCUIr/6o6OhPFyi9Qw2gpLQzwzFZDoPzgfWECPeJNTgHHrbQ4PcsNicgfTAs13JTepOsPcEmOtYE SjkFWqcaid+UWO9K4wPkFS0jft6VPh9XP2/QykkiyWruz3EQ2YOnAyXj8FGsILNW2UEN3lMkIu1F FG+i9E08NQ+P6v9gRsMCKLhyrJ2jZADLqyWKUdkVTgHQ0wr/N45c86gpPNDIeWDmgQ20AOsxEzRy qT50N4+K/8uyMljsQFRWhYaUebq8JunWU1wjtLK0RkgHBFLZnwBSsy6GJoF6ZpLmMO5mIY+VajTQ CgRNtbIfZhvzOdbIBBO7iRCMoI3C34MWNpc4nlYqkYSlIWiCdU5dsGPCna2lO2brdZkOa8OvAQZc lfQnmYZoQPxrK5KST1qk2hlW4IAB4bCF9ANx/SYR4ipIsyBaEnVW9ewZTpGgiIU/Wi/xE3jLQH6g geFztA3DyudkG9Wao9KIKh1Cq2PNtwko7feU4oPlQ1kdKwzWQSaRtAiAgxPIuyZLsEptauIA2AyM y3WQCs2uTkQahxcAJBFIyyblKtW/sqtN6hkf8liNwlkrTMT14o5jD8zIGK1WVHjZpn4wLsiME0Oz QmRUGQDXUiTOcfjkL7BmJ9N60WKg0OAuXiB7hSzUQ+QhUpCYKU+DGHPMQSpn5A8YxiI39L9m1pyo IajyosqsuT2Q3Z+5WoLijYKiX9TcJDd3eNAXNX1E28uAdS2vG0dkj5dmjVpGkqi6ZUBvumojN+A3 q20erzehyER1Uqy9PuJOW5KDuVrjgKZZnLDxp7WB1NU7/MJjCn72s2eoD2ee71RG3boMwlB2GGg8 SbNmKismVXLZyQoSffRoibqU2sOupp1nQk6mD4ZOoJYuZJ91gOCI5gdNOMsJnqfocmGJrDbAab2q vUf4q9cMKHkJFAsj5RUjla28jFGSigxnuDJwMwuO2LDXHogUvmWxCSYW2HhpChYMNCxOfNeIUIb5 HTtwOsvpzlyejhicNDE15/o3VaZTnZ5uc1uacFq0xl3EyWtwrjXd0eCM6U1gvzBvrsE51B8cc0lI VBDHXFXXOOxFdVgV/K+61FxPXZyrh4tSMxmYGchNq/6QUPoxK1HtpKB0pQop3SmIeuiiLpzPRZr2 E1nI2bkKQBPaC1EO79j/AfL2jcVIt1BrcFrazUDlisxlAdTgaxajpgriwEXb9UwkrcNWaDku2IDs YqaZs7WBxFyalkWxgWoWTfQiGd2Xr5QG7aKrQtGWADRTFxWihrqef4HugF0qyVL3vyhL2AGCYTxy rNk2Qy/HD2hcQTeB8w72J89apeB5irwpXnjpgXw4etJM4QXOGrChYH1V6Re4XNr78bQLPfhAd95C 64EAqKiVYFnXda30Q7BpLa3mgTurwIcsbmCxXhq0CcAaLcRvyWZrnRtpKs80i3MWtAA0QF0gyf6F Gnb6ApJZvpJD0ruSZju66ZEWqEFC85d9UKQejakZSO7y9xbITU95BPQ3IMu8MmnztEjxpmcvmtqr cdykSd5Gxoh8Pe4QCcam7lWiSesY85YXHPeqBw2geBv61kzIzqGIwoU2tiAtoHW21IH2ZQZFFXtV pMxpXPHZsyDp2f2H+GIs/df9EL8H3vszOrKYRmD7jpIU4E+e7Ic6GNw3C+tSDMAi8GOwzsHKZfvM wQkIC8R3FF/uBZI6Ulik+4/K/iJPPbk0rti0txQd+Eihvd2gjZSDQryd88u5la6IPdClRR7xrMN8 Pdq35iH4sYdPaT2a+eUWEuz2zKueW/NWAeCWPKae2/GaevbgubzIfky9V+7bkag0cNkh5dk5RVF7 jspdSeJO5HAnUrg9GexJAv0HtH9OKQzALs+CaCvsnsgrVqW7HpSdytBi27bnMgRmlusQKmJxmAf5 wF/3orupZjzUUysDodshN5V57dqkdn1G++NWbIVfm09XEIqZ1erUJ63w57M6p3JSp55HmzbCRZqN iF5SiiFraXa2PqnMDoOMuyiFW9TzbaMAepYHp9W+q7nN6poCTk9pKDGsyIgrMd+WJw5lktH7QZIS JuyqR1KJXAxwC9EQm4RDnRzULEfXcoXEyAtt0J4gNWMqknLP5YwK8RkQRKaQjh5KMCJHpx/zIkKZ enjK0cDMrTJRDs9Y/ejkKEOnZAxH3iea+DMGP9TI/Tuxyyl+v4b7Ip2P6f96jAZ9J8KdZ+P8lzlf FQHmXDzSY/nXnIcoY8x/zDmATMb4322QzPMVGpYpwU4zL8kcS0TGidGS0DiATEDikyPHolJTnJEt yhucQiUICjBuGq/1RbmLItSt+pSjDHHO80CbbDNw7xAap0VScf+GyK3bNU0trbdphlYsgBBJMB/g lLIwSCKt9k9LO32oAplhTP/fhf4U5TR8JZdiLP/ehroWah4hH1pOsXG5wLEahllN55DROMmzTk0m xR+MmbuGrT+zf6ohQ2yVYyzTeaciGeA87f6y9gH9BvOOMVEaAplmHIbc6jLP5LYOAOn2Mf9p4JAO DvorDQ2XrK8j2jxTYhocPQaGc+lBVzKlr8Mjq6mEXr1QYJpiQj65nQYoLyHF29iLWlQeYSMP4I72 FRV3p9Q/pLC5E802IrQdmW2I7IPEbgR2Ia8Nce1IaxYBLajU0KijcMLR0o6Vx3/qQbpOKfp3alh6 iii+o4x9CpejgNpp5xqeHD72MDkGN7dZ6y5h/lM3krA2N5F2urYPpxIwj4/uoJsC3CoxcBOUNQ5J nIqBb54n1Iq2BPFXetUazKYaW9F5xhBmwqEMK8hFqzG+TrVziDNeUL09oXqmw4YZH8Jw3cqtDSii 76OcGZIjmiOw0m3ZZsotQcumU5KZCmpt8sKwtUHKjSrab5fB6VCx9e9oD4kOk1NsjBhl8pR7SOA1 5m/fi2wV+7XNGcQFI3PU6DyMMUAon0Qq92optD05Pbrnz2p9ymEF6cu8ri443KoWdH+szvqpmEGK QGljNtnhko/6yoPSGJ2+8lLwTIVqwMAc+anFt7YE9DqWnODClZxXseDdAByAMdtZQYY7n1sIyvfr /at2rNh1GoqINr9JUpXbaxwVKi2jsjHXV9Yx9J5+PoWfY+sIN6bydpvjE/71JX3nF9pgVxMHCHdC eQHIVK+aUqdVKdGEwFsFCFNkDo1QR9htrzHeOzS3lsRhGrgSmDImMALV387L3aLvMCLBpjMIk7Iq kYgFDM1wrInRQ+8bXdscUXufUbNA+zKmiEmf139f0gZTL4gynpH2It+i4Fj8bK+8xB82QVtl2SYd P3s22y5/CcLQwx2C9DdOls/SVXx5Bl/c+TJ4Efinf/ti9LcvnpvtTle26xXPsjettPxV4n2J5O4/ 3tdgZjFxV2Yg26NCmcE+tuznUw/LsaYtA/jQ9nm1vVJVC14i74blcxQ4GWxO3Co5qG2nqVcoA2zz nSCmJ1FxsTIzwm7OrTYNHfCyQdcSbL/lYm5sdatS+7pd9x4H/enY71DOyiZRZ0Rc8+pp03qiMbkd RSBa3nkLgbv9UfVy/GMQFZus72ukmI45PB1qymmtfSR6xZiZQe8dl9xvxHuMdOcIN4wU9GPf4GN8 6PCKKmt3D4dZzRdNbPpCipailbXA5WZA0qP70cNdz7IgSMFtMheuTO+UBlz4tB+hFZUNzs4C/+ys Yetuey+xZpNolNA/IX7l/gN2H9UpBdX6ne4g/a0YW6Wo8mbtgY+gdSnUXmOi3OaWN1C+QVFXszYF NJDlr6Yp9zL+R/2s/5Fm/o8q9v+ocABG0gMYydcOH2BkdgJGD17AgxdAj9ELeLD3/4z2/oPBX+n5 g8FfZP1kBr8ptdPef78SoAnmXiRn5VA5kJ5wQFosXQeHJCdYuZMyaN5wcAc/YLPNPpUfUIC+gx/Q HgLcK9T3wReoZH/wBfao+a/qCwD39l5WYr9Brl1hSs+lq0Ss4wuDv8Gi6J7P9vidjfZqSusZIr00 5yc8AuSPaNQ/TOT/voZ9A5He8yR+M9uw0eELbH8nxzSbFa0mRZmNW+UnN6S3CGVRuL8InYfCS7pn bP5kwvN/Sxz9pc7n+Q14nDmZKLtaXSvLEXn14bg/CxsA0RMs06Dk4TVmv0eKJ3mKLAeqTWRYj2Ej R8Potw7HsnJ8xYNp9wc27R7Mt/8JuX4nOWssrM917mNu6TM9wOm/jcUlK953WqaiwipHPXWPa7vW qe7IfBB0D4KOnwdBZ67mUws6U9m7yTk+NZlPFFMnPKqd1+PS+19WDnq+T/cO/DNIMxFVouSJ/Pig 4RVwQFg/A4imzfXrC1w6gRyctqKwcZq3doPCTxHOzi6j4BewfyX34Y5xPGy7AFXpWa0teU7ehMXX Jxh3R85qXde722NYDPyJ7P+DED70YLazvg0SsYivMEJMyRMUHCxHrMsgW1XjWo31SMybWc4YMFmX rWVEdfRdDXXnFPCfhXAAH4YO9UdJqRv1jUbi4gfDViPaacSfynuNGhCW494wdE2NLJOE3M5RwoPW IYe+6wYnjlHLVpECsgalvFtEJKb9SpMpLe7zmhAgX1MG2JNUZ79UpZoIiDeqTPI8xrMguQ0awbxf CamTeFvMIVJKDgNPADl0rMtVAIJ/F28BrRdC3e0EFAbcHMZ4GrE68D9lVoZ+/Jxaa0ITAU7dQ9N5 vtzm5hmAxnXPMs3Wum4VV8roI2C4aIWLVqqm2xzkRqPOA3RlGw9gIGW0kc4CjjUZUPdwE16VWjHN wGyD6bAQAHVlfDN0adtqz4bVkKPOy+2eNysFyT3vFyT3XAbEYfajE/mTA+Key1dzQJx68sC45/Wg uOcNpzY2Ln/rGMBuy9lpTnEsDSt5qFjlwoK8ihY7Qvstq8Qb4h7JYvSXtovROdXv5iux9krn32jp NjlPH7cizRwrpSTeFsVEle9b0zHI6oPB/vqrLKXfEXS0h6Qq3ybGd4YZxFLlDrEaq8mrYgwl8YKx uvbTV8ZVt6MmtQeWBYmvlFrL0ulS0LHJuE0MR1IkEcgn9FQuV7v6XSZ4PqC8qu5dBsgncOfWU+u/ DEYeGb7dLBPPFyVr3QjrfZHhTYR/LxRIG13WXQkaCFFyo/iEZxM8etQUt48neKKOAHcp97uG0FYO s78UA5DJP+O5MHSQDXe7Eao8DXEmSFLH28xKgvQDdzhIQZzH9aN5sIc/xNm3uH04x5SqnfFF9+5h r7Cj/g6GLpjTDYe8sN7cnvMqfZybzFOe632rOUpRmyFT0bx8CnFV8+pqgEluWI2zI2/MpHA5ZAUP jTyV/DbJ8xrOQVNen8l3Q/I3MRieYgfGcKo1wsQLCrQUGm6rY8nWH23XIDKIBtKtL+OxLUzq7rwn odwb/yl4Ff9Bg1YwB81YYMs2STwDjOzAFsKdwlkMlgx6ktDsHR7XPaTueeQ1L1Ep462Se9WNlhWf C16cEy6Zr5hRYbMrP+xdHTXXVFFT/T/GaRogNZKPlzbjvFlOPUVCLEspP28XSR3hm+N9FCdz2B/e Cgl4wy4G8Ta1zv+1jTPv9dWcXEFZGZMNiJt0u9nQIg9YkjQS0IIQJwNYQ7XWl+WmKx7GDm+R9Cbx Qy+KebcDFrvSEeBFQHp85dg5cQyKBxVVdq5aHAbCcB+Egirp+mua1dCBaxC9bRa/iUDxo21DcM/R yD0nrmmo2WtHB5rkO4vv8LBoWyi8oVBAMW3zV0iOATvJZZCKobkH7YGfStrgsh1ZLW9NIsax8h40 hXb1jo6kmMCk4bIBmaUrNrKeUkrSFQE3nC/ZpFuCrm9qWoG+fmc2uruCO7Q7Vb679TRb6SoXvcET BbjBTlaF+K9+BSZ/6HsHJp0TUDvbh4GMa66Xen5v1UFwmYIXaMpI8SyFMjaB3Vsp9GihgQ2PFm43 a50SIGBJ6NRSZLmBJM84SGM8wDvNVXFJDTyogH4qgKlOnnV/T5rgDePhPBe/4NOjFs/d0SEL5FxZ oJmwgZx7KgWo53y9DbPgdQSC4xwP7j+n/lRUQisSTE9Xy3PNcFsdIMX/m5I4k4d3KK2F4jNPyF+K fD10w6iXchjds3YovbU7+omKVNcd/TM81lX38CkBXfuKF995+ow/s1ru9VGhLwy1CH4hBV05qyaV 837aQTidjan1NtXvzOTTJ15UTyZOcTmpSON1GgVLxxL1Tesut2JsDY7c48/cI+2ubsSedjc7ITPe 4G/j0VCpOv1H5nE5oe6WySq1nCoFiPXY4MapqZscMCXUM1YQqZWofKmsEtJUSGVUzQs2pSyG6U7j 0VJ3WrRvIrYmDmWab+Zepulx564HJPlxI6k3Fm7YldOD6l1UIDWsmCy1FjdZ3X406jXRio92h4+i s70u8Om3py0nYf7RdUT+p97O1g0Tn/6XEnRcQrDv3jiqnFfecXJFWc54opdL5F2dlW25R0BN0Nxi PxYX57ns3pqkLxnnGGhToqrpe0SOqAdtUznfIc/kJ/P//BvP/w+3XXq9X6PJJk0jNmIdS2Rzs9Gm QeediMqLiATi1kuCcAft3pDRSrd+Ad7jkNyDlyuQgTxrADZqCi07AXQ270+S9dxbMEsVJoAEpTr/ AM5IsgjjS3cer58RxlFlPHv++Wj02Rdf/P0Zm+CImac4dxtETyU1+rOn0mSPyMR+SujGgp9//tnR qLURfSNr1NODx3tRUqnB+1FUsZjauXJSMZ3qyyjtnZWXM+wj+zqvbGmRQX3wNgvj+YfbYAwxQGsy jVYACYk+fcULEYWF5MqXCMvFdnDZPJxdwutX8WjdD3jZEtTWBx4beaAgeVkBva84CZZB5IW59EVn MQOG6w+PBEuQpVYhdPlWWASD9+TipYv9wJFfDNRjbXkIaFUHV3aVT0+p8NIPHIijiE8k70FZrYrj os+I4SNVx8U+uoNJolud97j8505KpACiAu96WRld5kAHCPJfmGn4x6cSjc03EFVvHBrtdeXQqO3O oZHp0qEW59cpWsizYppLdiY3rfkz02rZfdxpbzLBuFIVHawqN5wZajpY41bWWDX8sF+c6a1Iv0+M KcXYxdJawqtlElDkYX1J+Zbqg40rMNfCUEb++j50PQXUbryErpDdyyY6Ho2ORl/83ViXoIkOIJH/ KKdcBauCLy4ZAIjnamdgADpPnonraod2wLVFrmXeRYzuZ0NARhX0M7ajPHayaBfWR81Se+F541EL yzeKmXvSyveljUkvST6+mx6+J/17T3q3r769q56tcwu6JGzf5qyJrf8RBiSL1/94N87ZZwn2/HZG hj+4Ljvv2Ybz/Jw+C9IUmvXs+P/gX6exemCXSBWcHWh3H+UqLkq8V2LFLnu2T6vi0L9dq4qC1Kri 9U6tSii09nfwhnsaME12QkORnoZK/90RTYp/vt5oYhXebFAE3vrYsehvbbr+rpq/bVMDq+DCGjC0 5r6367RYR6XpbW0ubx3725Avc9eV1uPHVimDK65wriKtZtQbWcl5mp8Yr90fL4Fy8coMIlbKH1xv Xdp+xal280Sl6Yj6YanyIj/IYy90aXY/b+Cjm6GNR7ajmfvoxrm+meJ/Dt5Ybx8PHyGs/wdQSwcI zYzfHVwbAADrmwAAUEsDBBQACAAIAE9KhEoAAAAAAAAAAAAAAAAUABAAaW5kZXhlZGRiL2luZGV4 Lmh0bWxVWAwAT6vjWDar41j1ARQAtVVNbxMxEL3nV4z24l0pbEUFlzRZqYhKBEGpWm6Ig7t2ElPH u9jeNEub/87Y+1Fv2xQo6smesefNm+exPV3ZtcxG0xWnLBsBTE2uRWnB6HwWscv0h4my6UHjxG0H zb7pZcFqv52JDQg2i4QSNoJC5VLkV40ZJ1E2x1FQKX5x8GDTAwzIRmGk5qaS1kQAxtaSY1ZhSknr iSoUj1wSgJOtMFaoJTBq6cS7XHzmTeRcUuWxnO35oiNrc/WbaWWLuco1X3Nlwyi3ILqFveFCMb7l LAxsXftCuoLWVC+FemWLcnL4ttwGMhm64e+Rs5PqAufgjB7lbtLpj/MN1WC43nB95MxFpXIrCuTj Fb/xyVHqouQqhsaENmAChJYlGbdO9BiMnMDrzmPyFV+joF0YNHoHNsAVr9HhhjNqVwgpGIEx3BN3 QaXhsOvjduN+em/nv2NbXe2B7o/ouaDjIK4BM0MwFFxoY0/pmjvQQQCWpsy1k/kGKiV+VvwBVwh5 j8JxB0lqV+7M+hONwUBy7whhBuao9bAirxz5dMntifR1vKvnLCbtjSJJarg9tlaLy8rymPh2JGNC kj9CuG7aGx9eUIfVgsmCsvMmddxmwKrcZDcatOpgY1NfU13aHmCK2uk6TlIqZZz4DSn6c8eiNR9o 1RYdKPZEcT4L1ieU4vrD18+fUNePF19OU4PFqqVY1HGHd3sL374Py+npDnrohUkPHqn/p+4u9gsz dimeTXQ36Ji7d/LRdqGMxR2d4H6SY6oL1b93kvYLZ8U1l7Jf6e7tm8OGQnsVgz59RLwwqcA3hziS jPwdwrBzQijLt/gcEcuNfRrLKxR+zc2fjF+0/9F/A1BLBwiV4abOfQIAANkHAABQSwMECgAAAAAA Y0qESgAAAAAAAAAAAAAAAAkAEABfX01BQ09TWC9VWAwAWqvjWFqr41j1ARQAUEsDBAoAAAAAAGNK hEoAAAAAAAAAAAAAAAATABAAX19NQUNPU1gvaW5kZXhlZGRiL1VYDABaq+NYWqvjWPUBFABQSwME FAAIAAgAT0qESgAAAAAAAAAAAAAAAB8AEABfX01BQ09TWC9pbmRleGVkZGIvLl9pbmRleC5odG1s VVgMAE+r41g2q+NY9QEUAGNgFWNnYGJg8E1MVvAPVohQgAKQGAMnEBsxMDCqAWkgnzGCgSjgGBIS BGGBdTwA4go0JUxQcX4GBvHk/Fy9xIKCnFS9kNSKEte85PyUzLx0iLr3QJzJwGCOUJObWpKYkliS aJXt6+KTmJSaE5+fXlaVnJljllhekJyWUZhibp6UlJWTnl+eCtRcWpKma2FtaGxiZGhuaWHySTKq aj/rZWWWNwVaXBeS+irEs62fSzuHCBo+tag88mvV8YQOJclK3c3L29of211lvGTxt/fh1vsvGH8e 2xD3XPiRlZNlzcxTzz8VnpKS13eKm8by63JcJ9e8pXK2LHsTPlYUtk/vetEnZi51pAYAUEsHCIRh KlUHAQAAWAEAAFBLAQIVAwoAAAAAAFlKhEoAAAAAAAAAAAAAAAAKAAwAAAAAAAAAAEDtQQAAAABp bmRleGVkZGIvVVgIAFqr41hJq+NYUEsBAhUDFAAIAAgA3EiESs2M3x1cGwAA65sAAA8ADAAAAAAA AAAAQKSBOAAAAGluZGV4ZWRkYi9kYi5qc1VYCADXqeNYf6jjWFBLAQIVAxQACAAIAE9KhEqV4abO fQIAANkHAAAUAAwAAAAAAAAAAECkgeEbAABpbmRleGVkZGIvaW5kZXguaHRtbFVYCABPq+NYNqvj WFBLAQIVAwoAAAAAAGNKhEoAAAAAAAAAAAAAAAAJAAwAAAAAAAAAAED9QbAeAABfX01BQ09TWC9V WAgAWqvjWFqr41hQSwECFQMKAAAAAABjSoRKAAAAAAAAAAAAAAAAEwAMAAAAAAAAAABA/UHnHgAA X19NQUNPU1gvaW5kZXhlZGRiL1VYCABaq+NYWqvjWFBLAQIVAxQACAAIAE9KhEqEYSpVBwEAAFgB AAAfAAwAAAAAAAAAAECkgSgfAABfX01BQ09TWC9pbmRleGVkZGIvLl9pbmRleC5odG1sVVgIAE+r 41g2q+NYUEsFBgAAAAAGAAYAxAEAAIwgAAAAAA==