136259 2014-08-26 10:48:57 -0700 Prevent decoded images from being destroyed when they're in use. 2014-09-02 08:54:47 -0700 1 1 1 Unclassified WebKit WebCore Misc. 528+ (Nightly build) Unspecified Unspecified RESOLVED FIXED P2 Normal --- 1 daewoong.jang webkit-unassigned burg commit-queue darin dbates ggaren japhet kling msaboff psolanki simon.fraser svillar oldest_to_newest 1031405 0 daewoong.jang 2014-08-26 10:48:57 -0700 CachedImage shouldn't destroy its decoded image while it still has clients. 1031406 1 237157 daewoong.jang 2014-08-26 10:55:09 -0700 Created attachment 237157 patch 1031668 2 237157 darin 2014-08-27 09:38:17 -0700 Comment on attachment 237157 patch Bug fixes need to come with test cases. Please construct a test case that demonstrates the problem this is fixing. 1031705 3 daewoong.jang 2014-08-27 11:41:22 -0700 Darin, The problem I try to fix here is a regression caused by patch r172790(https://bugs.webkit.org/show_bug.cgi?id=135939). Before the patch, CachedImage does not release its Image object if it still has clients and this behavior is done by a check to !hasClients(), which had been called by deleted isSafeToMakePurgeable(). The problem can be seen on WinCairo build with cURL enabled. And I'm afraid test cases can't be provided since original patch has no new tests because no functional changes. Could you please see r172790 I mentioned above? I think it would be nice if Psolanski could take a look. Thank you. 1031711 4 psolanki 2014-08-27 11:57:45 -0700 isSafeToMakePurgeable() used to do return !hasClients() && !m_proxyResource && !m_resourceToRevalidate; Should this patch be checking for the other two conditions as well? 1031712 5 darin 2014-08-27 11:58:25 -0700 Please mention that you are fixing a regression in the change log and how you are restoring the old behavior; if you had done that, I might have just done a review+. I think it might be OK to not add a regression test if you don’t know how to make one, but we really do want one if we can possibly do it. The whole point of the tests is to prevent this from happening again. 1031713 6 darin 2014-08-27 11:59:34 -0700 (In reply to comment #3) > And I'm afraid test cases can't be provided since original patch has no new tests because no functional changes. That comment doesn’t make logical sense. If we can’t provide a test case, there must be some other reason. The presence or absence of test cases in the original patch that was meant to be a refactoring/removal of feature only tells us nothing about whether a test case can be created. 1031864 7 daewoong.jang 2014-08-27 17:58:15 -0700 Pratnik, I don't think checking of other two conditions is needed since a proxy resource is not likely to hold an image object. How do you think? Darin, All right, I get what you meant. The thing I tried to tell you was it seemed to me that this bug fix doesn't need a test case since the original patch didn't need one. >> The whole point of the tests is to prevent this from happening again. I totally agree to what you've said. However I don't know how a test case can be provided for this kind of bug. 1031871 8 daewoong.jang 2014-08-27 18:34:32 -0700 And I will not forget to mention about the details next time. Thanks Darin. 1031916 9 psolanki 2014-08-27 23:16:56 -0700 Patch seems reasonable to me. But can you add more information in the ChangeLog? Please reference the commit/bug that caused this issue. 1031921 10 daewoong.jang 2014-08-27 23:30:28 -0700 (In reply to comment #9) > Patch seems reasonable to me. But can you add more information in the ChangeLog? Please reference the commit/bug that caused this issue. OK. I'll update patch with the details. Thanks Pratik. 1032038 11 237320 daewoong.jang 2014-08-28 11:41:33 -0700 Created attachment 237320 updated patch 1032164 12 237320 darin 2014-08-29 09:23:08 -0700 Comment on attachment 237320 updated patch We should not give up on trying to cover this with tests. The reason we have the “add a test whenever we fix a bug” policy is so we don’t keep breaking the same things over and over again. I understand that it’s much harder to make a test here than it is to simply undo the regression we caused. But the discussion we’ve had so far about why it’s hard to make a test have not been satisfactory -- it’s simply logically incorrect to say that because Pratik didn’t add a test when he refactored and broke this by mistake that we don’t need a test when we fix it. I know it’s frustrating, but when we make a mistake like that it is our policy to add coverage after the fact so we don’t make the same mistake again. So it’s not the person who made the mistake, but the rest of us after the fact who have to try to figure out how to add test coverage. I’m going to reluctantly say review+ here but I would like someone to make a test. 1032186 13 daewoong.jang 2014-08-29 10:27:46 -0700 (In reply to comment #12) > (From update of attachment 237320 [details]) > We should not give up on trying to cover this with tests. The reason we have the “add a test whenever we fix a bug” policy is so we don’t keep breaking the same things over and over again. I understand that it’s much harder to make a test here than it is to simply undo the regression we caused. But the discussion we’ve had so far about why it’s hard to make a test have not been satisfactory -- it’s simply logically incorrect to say that because Pratik didn’t add a test when he refactored and broke this by mistake that we don’t need a test when we fix it. I know it’s frustrating, but when we make a mistake like that it is our policy to add coverage after the fact so we don’t make the same mistake again. So it’s not the person who made the mistake, but the rest of us after the fact who have to try to figure out how to add test coverage. > > I’m going to reluctantly say review+ here but I would like someone to make a test. Darin, Thanks for the review and I'll try to add test for this bug soon. As you said, to prevent this from happening again is important. 1032244 14 237374 daewoong.jang 2014-08-29 13:58:22 -0700 Created attachment 237374 test case 1032245 15 daewoong.jang 2014-08-29 14:07:42 -0700 I've done the test case, but where should I put this? Please take a look and give me some advice. 1032534 16 burg 2014-09-01 15:26:41 -0700 (In reply to comment #15) > I've done the test case, but where should I put this? Please take a look and give me some advice. I would put it in LayoutTests/loader/. 1032550 17 daewoong.jang 2014-09-01 17:16:54 -0700 (In reply to comment #16) > (In reply to comment #15) > > I've done the test case, but where should I put this? Please take a look and give me some advice. > > I would put it in LayoutTests/loader/. I've already uploaded: https://bugs.webkit.org/show_bug.cgi?id=136417. If you could, please check this out. I'll move the test case to LayoutTests/loader. Thanks! 1032638 18 237320 commit-queue 2014-09-02 08:54:42 -0700 Comment on attachment 237320 updated patch Clearing flags on attachment: 237320 Committed r173172: <http://trac.webkit.org/changeset/173172> 1032639 19 commit-queue 2014-09-02 08:54:47 -0700 All reviewed patches have been landed. Closing bug. 237157 2014-08-26 10:55:09 -0700 2014-08-28 11:41:33 -0700 patch 136259.patch text/plain 1388 daewoong.jang SW5kZXg6IFNvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2Vi Q29yZS9DaGFuZ2VMb2cJKHJldmlzaW9uIDE3Mjk1NykKKysrIFNvdXJjZS9XZWJDb3JlL0NoYW5n ZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE1IEBACisyMDE0LTA4LTI2ICBEYWV3b29u ZyBKYW5nICA8ZGFld29vbmcuamFuZ0BuYXZlcmNvcnAuY29tPgorCisgICAgICAgIFByZXZlbnQg ZGVjb2RlZCBpbWFnZXMgZnJvbSBiZWluZyBkZXN0cm95ZWQgd2hlbiB0aGV5J3JlIGluIHVzZS4K KyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTEzNjI1OS4K KworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBDYWNoZWRJ bWFnZSBzaG91bGRuJ3QgZGVzdHJveSBpdHMgZGVjb2RlZCBpbWFnZSB3aGlsZSBpdCBzdGlsbCBo YXMgY2xpZW50cy4KKworICAgICAgICAqIGxvYWRlci9jYWNoZS9DYWNoZWRJbWFnZS5jcHA6Cisg ICAgICAgIChXZWJDb3JlOjpDYWNoZWRJbWFnZTo6ZGVzdHJveURlY29kZWREYXRhKToKKwogMjAx NC0wOC0yNiAgS3J6eXN6dG9mIEN6ZWNoICA8ay5jemVjaEBzYW1zdW5nLmNvbT4KIAogICAgICAg ICBbRUZMXSBVdGlsaXplIGVzcGVhayBhcyBhIHN5bnRoZXNpemVyIGJhY2stZW5kIGZvciBXZWJT cGVlY2gKSW5kZXg6IFNvdXJjZS9XZWJDb3JlL2xvYWRlci9jYWNoZS9DYWNoZWRJbWFnZS5jcHAK PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PQotLS0gU291cmNlL1dlYkNvcmUvbG9hZGVyL2NhY2hlL0NhY2hlZEltYWdlLmNw cAkocmV2aXNpb24gMTcyOTU3KQorKysgU291cmNlL1dlYkNvcmUvbG9hZGVyL2NhY2hlL0NhY2hl ZEltYWdlLmNwcAkod29ya2luZyBjb3B5KQpAQCAtNDU2LDcgKzQ1Niw3IEBAIHZvaWQgQ2FjaGVk SW1hZ2U6OnJlc3BvbnNlUmVjZWl2ZWQoY29uc3QKIHZvaWQgQ2FjaGVkSW1hZ2U6OmRlc3Ryb3lE ZWNvZGVkRGF0YSgpCiB7CiAgICAgYm9vbCBjYW5EZWxldGVJbWFnZSA9ICFtX2ltYWdlIHx8ICht X2ltYWdlLT5oYXNPbmVSZWYoKSAmJiBtX2ltYWdlLT5pc0JpdG1hcEltYWdlKCkpOwotICAgIGlm IChjYW5EZWxldGVJbWFnZSAmJiAhaXNMb2FkaW5nKCkpIHsKKyAgICBpZiAoY2FuRGVsZXRlSW1h Z2UgJiYgIWlzTG9hZGluZygpICYmICFoYXNDbGllbnRzKCkpIHsKICAgICAgICAgbV9pbWFnZSA9 IDA7CiAgICAgICAgIHNldERlY29kZWRTaXplKDApOwogICAgIH0gZWxzZSBpZiAobV9pbWFnZSAm JiAhZXJyb3JPY2N1cnJlZCgpKQo= 237320 2014-08-28 11:41:33 -0700 2014-09-02 08:54:41 -0700 updated patch 136259-update.patch text/plain 1744 daewoong.jang SW5kZXg6IFNvdXJjZS9XZWJDb3JlL0NoYW5nZUxvZwo9PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2Vi Q29yZS9DaGFuZ2VMb2cJKHJldmlzaW9uIDE3MzA2NSkKKysrIFNvdXJjZS9XZWJDb3JlL0NoYW5n ZUxvZwkod29ya2luZyBjb3B5KQpAQCAtMSwzICsxLDE5IEBACisyMDE0LTA4LTI4ICBEYWV3b29u ZyBKYW5nICA8ZGFld29vbmcuamFuZ0BuYXZlcmNvcnAuY29tPgorCisgICAgICAgIFByZXZlbnQg ZGVjb2RlZCBpbWFnZXMgZnJvbSBiZWluZyBkZXN0cm95ZWQgd2hlbiB0aGV5J3JlIGluIHVzZS4K KyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTEzNjI1OS4K KworICAgICAgICBSZXZpZXdlZCBieSBOT0JPRFkgKE9PUFMhKS4KKworICAgICAgICBUcnkgdG8g Zml4IGEgcmVncmVzc2lvbiBpbnRyb2R1Y2VkIGJ5IHIxNzI3OTAuIEJlZm9yZSB0aGUgcGF0Y2gs CisgICAgICAgIENhY2hlZEltYWdlIGRvZXMgbm90IHJlbGVhc2UgaXRzIEltYWdlIG9iamVjdCBp ZiBpdCBzdGlsbCBoYXMgY2xpZW50cy4KKyAgICAgICAgSG93ZXZlciB0aGlzIGJlaGF2aW9yIHdh cyBjaGFuZ2VkIGR1ZSB0byByZW1vdmFsIG9mIENhY2hlZFJlc291cmNlOjppc1NhZmVUb01ha2VQ dXJnZWFibGUoKSwKKyAgICAgICAgd2hpY2ggaW1wbGllZCBhIGNhbGwgdG8gQ2FjaGVkUmVzb3Vy Y2U6Omhhc0NsaWVudHMoKS4gVGhpcyBwYXRjaCByZXN0b3JlcworICAgICAgICB0aGUgYmVoYXZp b3IgYnkgYWRkaW5nIGEgY2hlY2sgdG8gaGFzQ2xpZW50cygpIGluIENhY2hlZEltYWdlOjpkZXN0 cm95RGVjb2RlZERhdGEoKS4KKworICAgICAgICAqIGxvYWRlci9jYWNoZS9DYWNoZWRJbWFnZS5j cHA6CisgICAgICAgIChXZWJDb3JlOjpDYWNoZWRJbWFnZTo6ZGVzdHJveURlY29kZWREYXRhKToK KwogMjAxNC0wOC0yOCAgTWlobmVhIE92aWRlbmllICA8bWlobmVhQGFkb2JlLmNvbT4KIAogICAg ICAgICBbQ1NTUmVnaW9uc10gSW5jb3JyZWN0IHNlbGVjdGlvbiBjbGVhcmluZyBvbiBhIGRvY3Vt ZW50IHdpdGhvdXQgcmVnaW9ucwpJbmRleDogU291cmNlL1dlYkNvcmUvbG9hZGVyL2NhY2hlL0Nh Y2hlZEltYWdlLmNwcAo9PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09PT09 PT09PT09PT09PT09PT09PT09PT09PT09Ci0tLSBTb3VyY2UvV2ViQ29yZS9sb2FkZXIvY2FjaGUv Q2FjaGVkSW1hZ2UuY3BwCShyZXZpc2lvbiAxNzMwNjUpCisrKyBTb3VyY2UvV2ViQ29yZS9sb2Fk ZXIvY2FjaGUvQ2FjaGVkSW1hZ2UuY3BwCSh3b3JraW5nIGNvcHkpCkBAIC00NTYsNyArNDU2LDcg QEAgdm9pZCBDYWNoZWRJbWFnZTo6cmVzcG9uc2VSZWNlaXZlZChjb25zdAogdm9pZCBDYWNoZWRJ bWFnZTo6ZGVzdHJveURlY29kZWREYXRhKCkKIHsKICAgICBib29sIGNhbkRlbGV0ZUltYWdlID0g IW1faW1hZ2UgfHwgKG1faW1hZ2UtPmhhc09uZVJlZigpICYmIG1faW1hZ2UtPmlzQml0bWFwSW1h Z2UoKSk7Ci0gICAgaWYgKGNhbkRlbGV0ZUltYWdlICYmICFpc0xvYWRpbmcoKSkgeworICAgIGlm IChjYW5EZWxldGVJbWFnZSAmJiAhaXNMb2FkaW5nKCkgJiYgIWhhc0NsaWVudHMoKSkgewogICAg ICAgICBtX2ltYWdlID0gMDsKICAgICAgICAgc2V0RGVjb2RlZFNpemUoMCk7CiAgICAgfSBlbHNl IGlmIChtX2ltYWdlICYmICFlcnJvck9jY3VycmVkKCkpCg== 237374 2014-08-29 13:58:22 -0700 2014-08-29 13:58:22 -0700 test case 136259-test.zip application/zip 12496 daewoong.jang UEsDBBQAAAAIAAUvHkVTLvj0zgIAACoGAAAcAAAAcHJ1bmUtbGl2ZS1jYWNoZWQtaW1hZ2UuaHRt bKVVX1PiMBB/d8bvkOuLdQ4oOqAOUmZKQYfxUFH895g2aRtpk06SWjs3fvdLSqVC7x5u3GHIsn+y u7/sLsMfkxt3+XI7He3v7e8NI5nEI31iiNQJFA2FLGIMZJFi25D4XVq+EAZIMCLQNoTPMaaGtq2I JCH4rX/WhIhIY1gMgBczf3W+rUwgDwkdgONu+r6jijAJIzkAR71eQ5cTJCPl1T9pqFKIEKGh8mte 6UF/FXKWUdT2Wcz4APDQg+Zxv98C3fJz1OkefnH62N+rcLBKID5RUYWTVH6F5RW+wbXUAH4EucDS NjIZtM+24AmAmROKWN6RWMi7jFLMD3cRq1WdHBL5QCWJJ4xic5Na0xBlSeqIpUrFlDzDW0XUbJBR XxJGAWJL5Ws2Qr9Brt4QhhjYysbPEkxlJ8RyGmPNjosZMo3SwNAhtp2rylTpS5JglklzE6+KtEOW BdwIUhVNRhj4jEoVQwDJQKgbq9P0KEN3BPdVfgaCEg5KiZXS8NyDAp/0WuRxfHOXd68uQ+Your5/ iKYPoWan+mvsOjN1uAt0dIsdzT2PZ0/Pc8WdaoObd+fy8SrveVo3mcbTxeNdj15ehz/nZ0oyc53F +NR9ueg55ZXxxXJ1ny0S1zXW2X4LEEU17F7o6ibVleYRkdg414BdMO6XeG0MNV4cIw7zHcD+r+u+ nfhfupIySYKi7t1/0Ieavaa+7t6meb+2r4XN0d2pR62G6ZtC7BcREqv8zIOYQXTQqsahBQIYi83s DK31QOuVaFU7cegxVHxugR/tNnCEUI8ggA9T6BNZABaAJ+y5jOPBYI4TxgsX+uq1bBuczc2niPgR IAIEnCW1oT/BAcxiWZq61VWHHdBujzYbSK9W1fi2wbFgmWoCYXkkbK8n5TUNjdEXQ4Ls9ZhupOlo yVIAKQIek1IFFxHLYgRixlZlOwmY4M7QSstydZklU/0n/AFQSwMEFAAAAAAAXC0eRQAAAAAAAAAA AAAAAAoAAAByZXNvdXJjZXMvUEsDBBQAAAAIAK4kHkUym98tIiwAAMYzAQAXAAAAcmVzb3VyY2Vz L2JpZy1ncmVlbi5qcGfsXAdYE8sW3iQQem9K0QgqqJQkECAgIIQqVZqKBUMSIBKSkEIREbteG4qK igXFgr0DitfuFbsgCgJ2L9IVHk2R8maTQPCivl6/Hfg3Z845M3Om7bTd7X/e/wZSn+rp4wmhUCjI G/xB/W/lTnokM6MhyN8fmghBkDykioIgDKBQE/HgghLThjhwQcO0CAUorJgPocFFDb5CmgC/otRF tBLAI5SmRAcFLjoSGgMu+pAsTANUoQwG6XcosyFxEkTxjAJoQBFFtDbgyq9iQjKwDmwSMOUCJA// Bfj4BgbgKIHBQYHBrqE+gQGQmONujccDbcWexRBkqj7w60rnRDFwQbEcAYcfy+HiKCEknD+VxmTD foiIJ1g74G0diEQc3s7BhuyAJ0OQ2dLsATsx6KW7B2nF1cbSfK02ldB3IGgjBMkBWgZPJBI2Ygbs x2hulBmkx2eiIEVJPA6Z6ME43TJlBvlTM2UH+cGZ2EE6IlNOWlaZ8hIauEyFIbTioH5U5kZRuhg8 CfA3DaEzpfRelLS+9qIl5QxM2YsZoGX35upJ090vziMKZuxHiWkUTKNFtDyIdv4BlNSeA+ghNEaa 1gGZQTt5B2TFfPQMwMcO0ZcbQssPoRWG0IpDaOUhNGhLEnsWHCCK0pIB9NIDNiK+DqBzYT2xPnox LIdpcUsgOBCscQQ7BxLegUSCvseDNjjdgdTcpgyE6+/v7wO/2AE/QBbcCgG5UZQ3UQMHP+AKg4TH E2ztSHD6d2D/gN5AeKk9eEu8hQ18xcXH46KtbCzxcDqogbLEDu1bMlhp35JRHdq30KiBeGV00YP1 jMmQlIE37B/4BfeMaogCyWOxclhZeTk5OQUFeUVlbRVlJSXlkZpaatpG+qNHGekbGuLGWprijM1N DA3N7CaY4wk2NjajTclO9sTJltY2RDgSlIKCgrKS8ggVlRHEMYZjiH+z678GachDEaA9osZAaA0U RgPVfwsaDayURYmcyHxRCWNkZLFy8gqKSkAhXx3kGINBy2BkZWVkgDQNyCEZDVnNMQRXrNY0qpxx gjZxyaZ98iZuZ67rBJe2jLWO4i1VUNTVGzFSf9x4U7MJE21Itnb2ZAeKu4enl7fP1JDQsPDpM2ZG 0OiM6JhY5ny+QJiYlJyyYNnyFStXrf5lTebmLVuztm3fkZ27/8DBQ3mHjxw9e+58fkHhhYtFN27e +u128Z27956UPX1WXvG8surtu/e/13yoratvaP1TW3tHZ9fnL91wvkDPQQ247+ZLA+QLLSODkZED +QLCJFhBQ0Z2DAGr6TpNjpqgZUxcIq/ttmnfmesKJtbBLTpRvFJF3bE2b8e1wlkT5eyvy9jSvytn gxmT5qsKUsagQOVhNCAXqDz3qo0LTtEMyl2MXP7yBdNfCfU3j3wnHbmswQ3A3s3HH1Q/7BwMUREQ xmCc1+Av6O3g3xBtB8nDdy48gWBtiJ4MKRHsSHgSaRIeOEO0IhhtRbGME8Wi/nuk6cVd3lMXrdDZ WtVwapyeSOYgkm2G5QP3VUiJywODZqBQwBUKgFcW5lFYvBAGWxg/QMO/wV5uFPjXhy1gwLIBWqIj 8vtz3VhRHA4LjllNFG8IMxlWd2MKYL40PQYvgBrPCPWYESoxRGSfo6iDnP3GPs0h9gVyBUwOmw9z RclyBWxRtGIboniDnuAYvr9UwmNTpB62QOrxi2LxBz0BMYLEQY9HPMtd5JFkUxq1Gy0uJjBqPk1i oLRkMCKajsPROcKoKZw30IDz4rGH8dxYuOE8Hj00jC3wNAlmwTUxRJeO+x4/mM8SiPhByawpwYNs xUQGTcDhuVMF1IG6gIJigvhwpQ3Qkl+KqBAY0YLvRR/K4X432RAaS8wP4tGmzAQMSe1hmsUtTzL2 oAd+RTU7XiRTlcbislgS6k9io8EsVuQkLUFFxIV7xCiRf+S3fkybyK8wOFiIQpmqi+MaqBhx2pgO yYhoFR0NJCxwxYoDSTj7dmQPcoiiawS4DnBIoquFlCOK87OI5oqtHu4wdRD0r5CJykJeXM6Smc4U MSSyUeJyGvTrwxRmv8j6QSe7Fy4k4OygSMgdCgFzAxuICECGcJAvxIR4UBxkCvGBjw4xAW0h/gV+ M8gfogIfB0qAhIDCAa5Ynwn+JgC/NWQLrskAcIw4UEY8IGFDAkDDFv/BfctTFFk3XGG4U5fy2UIW C65sOJNRHCGbzv9D/6QJCHChDDRqFocdA9MDjX+o303AiYf9A2YEx8SK5JICw/JZTBqDH87y48Px f5OOrEgmuTfJiTw+7kPilovhcYTcb1hYDo8Zw2RLeqayRwgcKFDEA34VqlDA8WKwGTyqgEEXWZ/C Hbj3KoqVYY7oFhUfg/s35B8j5LEkN21p4f+B48+PYXzDwVJZglBqzDc8VRoDhGMkC3z43qH+fgM3 KvkB9jfKCrEc3gJXFjNmoKTUxJn3HmDDpQsmO1Sh6E6lkMjgCb6jHg6zh6srRcVQOCwOb0jhaogD uHkNCkQDBIcN/yoIOFwwHPEZQwtOkQUKchhXOYojAAU6jK8Eqjh2qPpgfzUTh4Obs0sLJOVrD/Zr lMivDHs1xkqKCK6qCECCseTwNQieYI8QzfqbIWXROjmS4g/8HyFVkQ+iL4bDgen8ckhZXl5eQV4Z TOo1lBSVNHTUlJXVdEZoaeloaY3QUBY58PNjh1JRUlJRVVFXVVXXVlVV1QYXcNUQB/xrIhiyHjAe WA+8BBlV7L+Fchm2KsCA2fHA7Bn1RyEKkk6t1SGUDAosF9Cyclh5WYyytWi+KjNGkyDrOo2qZZyw hIjVBnNqN5OxOsHXo6xteEtLKXLjMkNa3rbS+CTd3LPLxrtvDqV73NgvsNV7EvaO8adzy2+WCd+3 eZpuObDi/NZbT39vP5j/27OajvDoxJVZhwpul3/otPOaHpO0alteYXFFbReY2Yum9iKbwMKLJDJh DEFTBliQYKwlC5Yr2rAF18FqxXps1Fve0kxKiA6Nb9M6DgsbIDeedOMJMGK/Ht0jzFbAeDdowo8t MJWa8O003fXk6/MFkxXzQyLKjV60Tlib/RbXeD2EcG1msFPo9RO6RoGlb2cdundMU0Znbpbu18Ul f9p0/PJtP9Q1/2Pe7FmZSonWm+/OrNxs4Lb5LKGBf8yxH3LXZ3GpFgbXU7EftjnWFfbq74z0b7Y8 8fC0igJz+x7TrpANZMicnRk43flo05laDso64UWXQ8oenfurKpl2hBUf+csonqtCHncuyrTyT6pq rV70OHBhfYtj0IiHkaVh1MOPsm3vGo5xUbZ54Noa0Ul/H9z5pGtJ3KOqhlir1P27t4YbfnrU86m0 H7KYnlu6vVYjgxF2eqMRk35h2ceqKqP4VJeE5qKX9e8Nj63NfaVVZJjoXKD65lj8K5p7eNl57n4L pv+bZ3M0qkOi3PI2ztCv30h7Lpe06OPrR9N9HHPJbvduxT/TnY/J9Q3Idaoij1JyOLbi4M74l+R1 t7gdkXG+2yvOo0LurpqBPXGx2jmj2+aMbDXGJM4bVXqorhsUTDrnw7GkICvmXRXoCa5wY/nNpAvk 1/XO2WHPMXQ34lPmXoWe7BQnpywF7jbt9rdlM7ZR9LEH6tC/Wp0qS05jfOy5cQOVU8Puh7AHPvmf fK68aV9JM6FvfXFVu70gjNRa3xt9NTBu30ODjBJmUhs79WpDmHXk0eIgVnyX87sK0pIXV9Kae1+u 77W6qH/v5ebU1u1K81PiLOJ5RnEHnIsKeuVYMZ89gjut8+iXt1bGLTMoajusNm8733z0nscfciIW 0gx0ilwbFp3qL+FRVB/9blz+7kPdokmkbeyEhG1aPWFV5rGEa0q202T9XwetjDdEHV0RKFzF1Q8U GvVDO5/vP19w5pRGbuwa5o3umz5Hu8/uccqyCf1T8wylgA8uVAvantv9UIWTy4fNVAZ9uuuV7IcF bzfVnzmRYlfSlb8j+dX+vt1bZ8bbTevp2tftmMYJuB/e2Dln1tEz688W1n2YLGeTYXsx+GH+blPN X7MbzjU9V7VVPPrr03IlzBkW36CuJs2Wzuq51w+lBW5IlBnrctQl6vOdMJ2iwuDiwpEn99XkFnwc 14QrfGowX9euntAP+T7rffz5N99Dwu7Tp4/fyt9AnlGwQBhe4n16xpfXDQq0NIvHHWHHP27tDNhb +/RP3D9NOtkUo/FR5QK7K/zTPgbLIHFR1G8PCktuJya8uLf+4TGd/JzUEy5eaZfi23utAp+da+ec /b3uhmnp87OPcFNGlxokFzRffMlpNCCY4d6VfJq6SrfXLDWHVhjh79jEv9x5tU+r3feoXnjjnHpo YszZC5mqBx/zfXu/Onk5lr3sU2G0v/plfdnhQztnNvz+ZMzj1PVLTo2qWKBPWB/Ty3k6veoOs83v 7pbRfp7rxkcEzzJ8XiM3D5M0YZlyY6Uv7bGhqpbbijPxHOvjmMdrrGdgvFuD4+QVrYmaIVNf2Om5 mtLWbVTNtM/KVY5YNgUtH+63ccfeJU4lwXJBk2KOzjt7ea59Vl/i65pz9QSXque3VxLPbV3FZzTN 257DZ8XVXFl29W1Sqnr3QcZTve7oJ+W3MLJnP3yt7DnUD01AXWDTzB8fIF1oP90t1NP0SdE52xDP Ke0wT3ztY28UkrZ189adZYkFRQbdEavHXxV2kGse9ZI2dFqXv2w+9+XpTZO8lyt7Jt2u1n8y4cS1 vo2YAv+dl1/fC2ZjZvm0NyV+Xbv+0+Zfnwde/GjZPesqoYE8x7L2fdlMvopaYWphs+xJTGp6YE8H J7Uv58PMRhxr0/P5iqRuQXp/6alw65v0OoWS5P2UEahlF+qOkw++Kq9bqP6x+HQ4rV71vs1Bk3E3 17s5heltVPi4uELXbjf0y2fV0Ee+DajaIPLnsToL7PSrns/blCfosta9HV909BNrKTN/X5NgEe15 eso0snBuS3ns+8dLz0/kV1WMeqGm7V60Aq1Ry1DsnsO5c64567ztmeRYg1c3qhoZ6zDVPTV+/dDj tRvqHxS9eMHevdHrEO+yV8mXxFpyWUhTP7T6NV+PsSs+3f1u8c3D0/eUWRQYyL2odnqQVJi6s9vR vC/pSR+n/djW4gfzGj+Pv8HK31bO+9NoVLnTxy6n1t74e7F9W5boWaRjF2ZqLvc8j3n+xMmrWV/t zprOps/jj5w4cfJcXf3dGo9tBbWuDmo6s7GXT6/6+AK7unodd19jEGrVXf0pNZS0jG4Hm5apvMYJ LMvnXqWmPkFHLy8dP23SbynlgraJAbdT1ANeL+WuuzR+jfGHQsq+CV3zFUr3b/49rzOa9cE7RChY +TWDrfFmwgHd01erKY3KFjJk0O8W1b199Tp3cUPRipkkGsNplXxZI3bl+kf5p54/vKW3aOXVvOAV TcqMrBl2odet0Iy9ObXui95hsWkpO2wXOEau3XD7WUAFtmrSKWHWS2FAZ5Mwa5+mgknU6t1CFvmr muZCPRdQCumxBVl422r94ulUuVhz4sNyyvPa+nN9FoKOXbN6n9lPTuEd+SBX8rZxJT0qMubgebu4 xldl9UV2C0DJJjr2WNXmrS6/dGD96P4nhbtnRRWGezbOWzY54IjKu9KL1x84jGqcZHDYticymB5s W5xQGf51zgGXMsftJObqsutBNicnPV0VcnT2CI+wO6beq0vaFK6TSj6FPlhfQppp1+FyvryqSYPm vGRyP7Rre+SWe3dpvW6t6oEKBxIjkkovdX6oXB95cKYqtXva86lPrl0yKXvNXrV9ksoO9rER4fMc hBHCjPzwCmsc0YAwIUwlWGXsorgTxwgd9ucL46JdkllVVN7I1MZ7HXb+l3a1vpN33J7eoOdTll3m KdT6tTDul1aN6D6XT417GrBEXOcojfdfrrZveV5krZKQcepDZ1hRz9HGe3LtarFn9iZtuqhT7vLE kltp9N57ttHMHQJnszsF++6v4PD1088+yfvq5fTI4FPB7RyDUYs+8SKTKj70jLi2YK1DwTjPEUoZ HOhI4Ym5Peur72e/fmhovjJm9mN+/Iu3Vbfz36UWJrxY/Sifr5Dq4dpWVZDRk2miIMfdo7Vg2+Hi Rt3JkH7g7+HmDYm6U9jHMnsv73I6URjTMWV011MDh4qgUe83U43eNI4eu4Gc80TB7mUTNiCgsro6 cIJf9ankrsUxG5OdPRiz+wo9ZxbXd/aVFNMmJtw762TafZNvS65NS6/ry3bkO2xwyq94yj2+7zG7 loW9VqL22T+H0pdNXb5qlqumytEjL7cXr9JoGGn/PpyUnxP22i91Vv7q5JkPF5ml7MuyLb9Artmr x3E+1j56x8V7I42mFp1g+B3WY1x8WkZ9BdGWz4krJF3pPPe0V/Ni+HinUMemkx0jysbk7EJVsG+V jgxfNitMv6yo6U7K0x7PTV0Td8xM4Br4l/t6hN1efMHlqWVfcW1u2Zbq2ROP3LhkZMKv8o1wSakz Pyd8Np3BLv7qeP+uqpquXn4zbU+XXR0rrjfns7H+Zo+IDqv+sg/1p8srSOd0jVLReZ33BPuZpj7L 9a+fi3O3ubTsg/asDU1xizbyc/Y5K2btlM27/uTcqBPXy412Lq7G2rRMSb3de97+a0zhq2Up+TXH LV8Z3FuVl+q7fiHex4y789mcWZdddC52ORzfptv8fsbWdMXshE8Fa42unNT/UuszQ7tmxvuG0DrC Nc/9xLam5M4i+vJ+KLd1TGVdUWrrFBv8koavBgTZW30bZn969LpphlHkqNtV+oKqEmV3H8vuuc1X s58+Okx49rx4975mvl+XqoxlYaOeT/y2tdkQ7SXu1vUXzelJbZlFuHO0DvcNj2VcUwxIzotMMw72 Q5s8LFRiWzeY5pD77lUWHlxjmJbpm7dM/YlXYV8/1A/l6M6ds164nGhbXyCTkj6p+2gAwfuNSeKl mISe95qOW+XQrvxFv2R6v/XrNTGZ+363dgZm5iin2t70lElhxu1dsTYmc9ULNcy5xMk6ixP34No0 CC0+D2i33+8brfZmtdmYGY+udAfIW+pswwRVMJO7iHnF5/qsvsxflBekWxm3fj1fPv7BNluZlg+F bSPbN44U5l/IftHQDx317llTkza1kZlo63/LYPbLStWnm0wqmd2z0rdsolxhNZDTxuT//qrnQcNx 8roztYnl53XTdq8o6YfKd0SNcyt4v+bV5/xmtVTh+IuoV5tffilI3Vh0jFOfYn7n5mnGtMwJ7ArS 9TfkDfyI3ktX3Tv7IbszM6/XOXkdfbZp15xOc6LvDIuxsw58yiqMqf2a5js7t+NTbrVt+Uib6VzV M7+1qrwQmpzfsuCMw8q5zpSWxrPl7+8bPpFrW5yf0f6y/UrWp9LL0b+30zKeMX8T3rhOy676NJmd sTe9oOh4HLEoXCfP7OKWy9VV9TubErC32Ym1X12mZnsd3lxJOWLV/7T2DPfN6Vd1FknX3MLnTLl7 5uDm2deT7m+NccxqK74bdPXLmVaX0c0TYwSm9+JGeFX2rV3+eM1Wmd4dDxNXPjQMXKqT8xCsDS58 Gccp4Uzu9M0tOpdUOZEVWmL4Ui17/vHorsnr1c6r6QS8rl3VvWK/V+loyqZW7+ZnEdVWlUt/9VUp x25b+sHu2in3C2G7S3dsCyy6lFtuH9ecmhaQfrI2//7DBdvamroXqh288cn5ePo7ytW2gjSXfe0v ex736PU1df+Sco56Oe76w4bPTZyATOqj2G3p4bZggnsmwS62LsmpeGn75MdfLHK82q7u7p5d2Xdq bsq23gk9ba/fLeeEp5SvZm+sNNzg7GNz3/X3Wwszz3051xvyuLVvqn835nPCjZSoIJkcK/np+MZr 3PQq36bTu/OD70bvnNMdd7Kqk3QlL2+WS+OV3GORnLf3qrov9TS/frJod/qD9PDXeaB7BZefv7IT d3GOy6FZu3fhV34KWp+4wdB5Q8u7lydSr5KZ9N45eXkGszKrOXeZ7StG7TCP26Xf/fpSZee2fuha +Anh6LxNvZN61NrTay+4PHGp+FLj/EufXndM2/G+nNZ+6Hw/1NIP1bu8Kd2m59sX2Zbl27Dp4oVV J/Ivrk/TnvdCpdx+gT/fzIJEq8hon1lfPSdyb9qejksVC9arHrm4PbOrbB7/lc3RD96pTZWOaY0Z Me3HN6QfcKwH7bHn02e1z4wStS8T+iFK/R7nVS47enJe90PLpzQfqjhaVn7pfOSepj1nwudqT8/w Tms+eGNu4+tu07MuKa8KC7vXUkwte47vz8hsqS2vdb7z5sKliH5oc5iLx/HQ0y41ky7U5Gxru7oV LLpCFyT1HPzy2ue1b86j2i/pXR0ji1pCC7qv9leK9kvGQLALGzjdVoNcRZuYURADbEsGQbGAFgDw RRQXaGj9BQ0c2DQNkWwOi1LAijf5IHlxGv1vjSNjBQKug5UVm29JhfddLGmceKtkKteKYIm3gia7 JHOptDiGABfFAPt9TsafLl01xjHpTsbTSf54fy6FEcv0XsBjhCwICKUtiKOR6cYuziqKk5MdkuO5 8QwBFZccz2LzHZKdjEWxOwAaZlsZ40QqgjgnY/FDEjP8g3AUDo+BI4GTZxreFo+zJVgSrG3s7OzM ceBECW+FJ1oRiBbgLNya6IDH4yTOGCSnMJlHj3YIdveUpAZ8TsaSfCUlJVkmWVtyeDFWBDKZDMdC JFoADQt+CltATbZg801EcYgjcWfwaTym6AQHB/upYK9S4GRsLImZO3AgBsc/vNwGxOLSkwSC8ylV H17MUkV/Jo3H4XOiBaJzt0hC5NBg8QNCaUpw4IGwVGHy9xNhgGduRKkAjaEm+fv/3Kj4+EFtvsAj UfBzbT68/2gVzOBzhDwawyORwRaYSIMHM6L/luBAfTAwnTYYkgv2dUUVSadZMViMeJAGH4QmDNpJ 4/G/nwyNGg/2py141CQLsJkpYLJj+JKCH6wxhzBeDINNS3EyJgzluoNdbQqPAa6gzcMHmxZ4ggXB OnTgkQxLPJ40VN+PEUOlpfgEhVLcmTEMPig1DzzJ08bO3cbd2pXo7m5vh3clk22tKSQPWxt7GzyJ PDS0aCfXn0NnOBlbD+X7UChBPE40kwUEfHCSh/PxoNgSyLa2FkRLApx7kDA1isFyMg5m0MV+kdFg W5jDAdyfPock1gepMqNT4OwO5BNETgwdeDDJAm8DOt2QqBmDqsOLhCTWC6bCJe1kLAnmD/o9HZz3 /cU0hnWDoTGBVuwQwuAxqawAYXwUg+dkLHmoRSzyY7D5PuxoDpwA3oqAsxFfxZCqwOLhj7gMicId tAOitZjhE0+NYQykRsDbgD+xwJXL5XGSmfEgQ54cmpDvzuQLqGwaA2iRQLJAS9TPHNyBEG6tcKyA Y0ln0h08Pew8CQQ7G7wdEW9rTyDY4+2t3d2s8RR3ezdXVxJ5IKz4oIXKGhqHK5mEJ+M9bNwpNq62 RBIFT3L3AOfqBHDXJHnYkG0GwvqwxfYMpMsE6YLI7dxBoyXgxem6WtuQyBQ3kqe7HdHezROEpdMc ojk8kCknYyacc6v5XEaMMQ50LlAPSZ6gEYJTcZDHSPcQig3RhmwZ4OEpFoOzCz6TA0YJW0s7MQe0 WnASxR8UkAYE02OZAoYblSUuLoqQD04gxJJQRjwXPkwS8hiiusVL2Ew2MEji8UjmcvgiBbzlgEJI LOjvSXwQRux3Ex1fsBl8wJpEkihROOCUnAp3y0nWtpJwICmQnMg8aVQ8rjgkURKbnzCeyYaNDYnn cASxotZoTRTJxL02gMPkM0DvE9LEUVlL4gqHT3gEAoZrPEcoyYHU2m8yBQJ7CxlDvd8xzYsHjn8k alLGdxTdWEKGWE/q/44aqE6WH1xQcJsdsDiKB3IqDeoN5CwAcBunccCxFbhHkiU5Z1F5TAHwTyLa iznujGgeKBxpYGCDK30+qF+48YI8fV8QCFL8UaCZ4LSNk/R9mSj33xe5Jgip35fARfF9SZCQx2X9 QOZPjQFXaZTSwhyWv+Hy4dkcrjI8t8NVpJn+mYY078MVpEXwMwVpSQxXGV4gf+giw8tjuHh4cUg1 htX9zzSkhTFcQVoWw+XDi2K4XFoSwzV+1jK4LKYglMMG/UDczUE7+rFQWrTf0xnseyCOn8p/Ho3k TjsoCqLywD1cwGPSBm6bwyXuVF7c9/ii28X3BIO28H+Yjsgk6V1VKvZn0gXg5FoiJ/04bomGHWnI rZrBDqbSmUL4Rg9mdd8I3MGkg8mSpChle9BjGP5UfhwoGqmtHL6AwuNwf3DH9uJRmWwxb3hj+Obu L0lUmg3pADFcxeanMUnGq8EURZMTyVTQg02NYjG+kfhT2UIqC56GcHgg+B/NDQLDMJcBIk4UPWbA pFFZ35N5gwc+FoCEvy8N5gioAtHIO1wWAmKEZz8DY7IreDbED2jzqCyK67DRG0wBiaJ4/jA4w/zh gw9/kC0dPwd50jnFtyw3Fli9DrC+GbPEPGnKYFwDfQVUcgpfnAlPKovPkMw9QMukCHmJkomPPwM0 tnjcgLnDdSQ58wMPzVB5khREKxBJzQ1MyEPA7IxO5dG/ozKwegBTRHt7ki2JSKGQiZ5kohuR4uFq DVYTJAIBfspyWKsIYQiEYLkJc9zFT4/wJQVJ5YdI1j5OxqE8IWOQDTf6b/LrygIzfHqKKxf0NQZd rC1eIytMFs8svUVNLEXEk6ycQxgJsHfQz2LiROtGB6pkRkQHs/ZEsDyRsLnivs2A12w0cfkz6Lho HiceJ552gqU5M47DtmAzonECjoQpYEZHG1v9PCE+dUgyzD/Ogf+qubc4bFIsg/3HlQoRrFHsxCuV ATV4qZJE5TFc4WFg+GJLXK84MHfGgXkxzmxw4TVhIAJaLBgH4YK2+oez5mb38+n9z7JmbQ0WcH9b 1oauI//23Eiq/fttAm4J0koXNwHpcsTqn9vW/vYU/nJN2P8P1cRfzg3578+NaHH/35Wbv7gI/k/k ZrKV9D4q8v7xXiu9/7qL27cn3IJF22t/980ADkv/u/YnxGE5f8/+hJUkM98MnD8fTJzx5jj8ZCux 548yayLYIyb+SGprY44j2f5ISiDam+PA5YdyMhHIyT8MTySRzHHgMiD/UWV+m9fh+Ycnff9AEfxd ZoiSHG6JaE3177ZFkuhwa8AC7R9pG3+XMeI0h9sSFA5P7f7d1ohT/bE9/4GWM5jwj636j7SiIUn/ 2LL/RIv6JmVY+ofTJsAFTMk5FvBMtho8QwM+3L/BIYn8TQ5J5G9ySCJ/k0MS+Zscksjf5P41iUgf FmGw6U6mSaYuzv3vlGeAw+vIoOBATx8/D9FzLcreYGOMgwbvxsWDDTT4NBu8Bo6TewTejVMALwCC t8CoND7XP8QzVPTOmAcFBx95Q9+4znLxW1dlFt5BOBz0tzkNGpcnAM+/BAHamg6GYEAvBzQrScCF +S2A1o6Kg2m0DEzzgIGAHgHTMWLaXKQjpqfAND2eTQc0bDOXHk+H6VuAXp0oZAAa4wfolYlMRhKg nwJ6LAts+AL6Mxw2nkHlg/emlGG+gEGLBTQe0Mq80GAKoCeDV86UY4bQUUNoAXgPTfTKMIebIjrx xJnRJoD1Ctke581IYoEdP4sgUClglxFsWMZzqewUCBLlWew0hz1OAHP/GuFf6eC6FVNt08QfodB9 APN+rMfJBa+TdYCy2SjlRe2AoMIVEDSiSsobuxc8IgXqreDxkPzowu1lyKM/TAbNEhQocH+twl/h hqRnCUc3WDw4yZYrDi43Gtj1F/JwfNAnGDiLwUb8jwb8vh3mYC3O4DHYIEQ4aGVglxfen6YzRQ8y MdnDKvEfC/YHJ2rXsNM60AdpR1pC6o+1IUzzA0hGSwnCzN4D1/pgvfkphENwz5s+uhZu9z9+exad AV/4zBhROEpwKI4GJtKQ9N1hWfC6pxr4wM1I8PEeE8gMvABMBC8LO0JTIA9oKhQIhUIzobkQDTwL Fw9e702CFkJLoJXQWvDFnK1QNpQDHYCOQCehs1Ah9Ct0HboN3YdKoGdQNfQGqoEaoRaoE/oK3tuT Q6mgtFAjUaNR41CTUESUPcoZ5YHyQwWjZqLmoWJQbJQQtRC1DLUWlYnKRu1DHUGdQV1EXUfdRZWi KlFvUfWoP6G60Ri0MlobbYQej7ZC26Nd0b7oUPQcdAw6Ab0AvRy9Ab0dnYs+ji5AX0ffRz9Dv0E3 ojvAt1aUMLqYMRgLjD2GggnERGCiMTzMIswaTBYmF3MSU4QpxpRh3mCaMF9ksDJaMjgZCxlHGW+Z MBmaTILMIpl1Mtkyh2UKZG7JlMm8lWmR6ZNVkTWUnSTrIOsjO0M2RjZJdqVsluwh2XzZ32SfydbI dmKxWF2sKdYO642diZ2PTcWuw+7GnsJew5Zi32M7wCdZRspNknOSC5SjygnkVsrtkDsud1XuiVyN 3Gd5JfnR8kR5T/kIebb8Uvks+aPyV+SfyNfKf1VQVxin4KAQqEBXSFHIUDigUKTwWKFG4auihqKp opNiqOJ8xSWK2xVPKv6m+EKxTUlJyViJrDRNiam0WGm70mmlO0pvlb4oaypPVKYoz1YWKm9QzlO+ plyp3KaiojJeZYpKhIpAZYPKEZWbKq9UPqtqqVqq+qjSVdNVd6oWqD5R/aimoDZOzVVtrtoCtSy1 c2qP1ZrUFdTHq1PUqeqL1HeqX1SvUO/Q0NIgaARqxGus0ziqcVejTlNOc7ymhyZdc7nmfs2bmu+1 MFomWhQtmtYyrQNav2nVaGO1TbV9tOdrr9U+of1Iu0VHU8dGJ1wnWWenzmWdN7oY3fG6Pros3Qzd s7rlut16Rnquegy9X/RO6j3R6xphMGLKCMaINSNOjXg2onskbqTHyLiRm0YWjnypL6M/UX+afpL+ Hv3f9JsMtA0cDWgGawzOGlQZog0nGgYbphruN3xg2GE0ysjLiGu0w+imUdMo3VFTRs0ftWXUlVH1 o7VGO49mjt4y+uroBpwOzhXHwm3H3cK1gO/feI8Rjtk35tGYr8amxmHGS41PGb80UTSxN4k22WJy w6Rl7Oix/mMXjj02tmqcwjj7cbHjto0rHtc13nT89PGrxheOrzMdYepjusD0mOkLMxUzF7MEs1yz pxOwE+wnxE3YPaFkInoiaWLsxJ0TH09CT7KdxJy0e1Kpuaw52ZxtnmteYaFs4WqRaHHM4q2lrqWf 5VLLQsuPVmOtIqw2WRVb9eFJeBb+AL6aoEmYSlhKKCL8iTiRSCPuJD61VrH2tE63vmDdajPJhmGz x+Y5SYvkT1pFukHqtbWz5dmetK23G2s3z26XXYW9tn2Q/Tr7O2RZshs5nfwr+YuDrYPA4azDJ0cL xzjHo451k00nMyYfmPzeydiJ6rTP6Y0zznme817nNy5jXKguuS7vpphMoU85NKXWdYLrfNfjrh/d 8G48t3y3LooDJY1yzR3j7uW+xv2Rh6ZHmEe2xytPY88Yz2OeLV4kr1Sva96y3r7em7wrfIx8aD5H fFqm2k1Nm3rLV9k3xDfb953fRD+eX5E/2n+q/2b/FwHjAtgBhYFQoE/g5sCXQaZBCUGXpmGnBU3b Oe1DMCF4YXBxiFZIZMjRkM5Qt9CM0OowszBh2I1wtfDZ4UfCu6a7T8+c/maG1Yy0Gfdn6s9kzrwQ IRcRHnEoomOWx6yts2pmk2avnF0+x3RO8py7c/XnsuZejlSLpEaemyc7b/q8o/N6qIHUXGpHlE/U rqgWGoW2jdZIn0LfQq9nODEyGbXRTtGZ0XUxTjGbY+pjXWKzYpuYFGY2s3W+9/yc+V1xgXF5cf2s 6axT8fLx8+IvsjXZcexbnFGcZE4pdxJ3JfdNgkPC1oQWni/vEB/Fn8O/INAGk6kHQjPhCuHbROfE nYmfk8KTziVrJLOTH6RMTPklpXaB54KDqTKptNQbC8csXLLwbZpr2r5FqEVRi26km6QvT69Z7LX4 8BLFJXFLHi7FL81c2r5s+rKi5UbLFy9/v8JrxbGVqit5KytWOa7KWS2zmrn60S/Wv+z4pW8Nfc29 tfi1WWt71tHW3VtPWL99ff+G6A2PMmwz9mzEbmRvLN/ksulwpkbmgsz3m/03F2zBbVmzpX1r5Na7 WTZZOdsUtwm3vdnut/3CjrE7Nu7oyY7NfrbTbeepXYa7ftnVtZu++8meKXtO5hjlrM3p3svc+3yf 176C3PG5Wfux+xP3fzgQfqD4oP3BI4f0D6091JvHzntzOPjwrSN2R44cNTyacQx9THis/vjs4yUn 3E9cOGlxct8p3VNrT0Onhacbzsw7U37W9+yNc/bnTp4fd35Xvlb+mgJUQUpBS2Fs4ZsLMy+UXpx6 8UaRY1H+JctLeb+O+XXnZZ3LGVcUryy/0n91wdWOa9xrTddjrr+/EXmj+uaMm09vTbv16Dff3+7c 9rx9s9i1+Oodpzu/3nW4e/Ge/b3C+7b3Cx6QHuQ/JD3Mf2T7qOCx3eMLJeSSotLJpVeeuDy5XuZe dvupz9P7zwKelZaHlT+vmF3x5jn9eV0lq7K1KrHqa/XiF7Iv1rxUf5n1yvBV7usJr0+9sX1z+a37 2wfvQt5Vv6e9b/yd/3tPzfIPKh+yakfXHqkj1v1a71lf0jCroaaR2/i1aWWzRvOuj2Yfz3+a8ulB y4yWmlZea/+f1rWNbMtrt2m/0RHU8aozvvNr15rPIz8f/mL/pbh7enft16QeuZ7tvRN6i/p8+170 x/f3iz4Whkah4H+Jw8hiYIcFb7PLyoG32WHIKygpyMsrgu+JKaooKiopAyenoKqmoqwK03AkcHA4 FPhXlpeTV/6bnejjALJ7FSHwcQDkY2H/Bx8L61hjlnv1XltPiKJZ7mIECBAgQIAAAQIECBAgQIAA AQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIE CBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAg QIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECBAgQIAAAQIECP7cPhzTAADAIACTNv8S pgY8cLdJVVVVVVVVVVVVVVVVVVVVVVXdX75QSwECPwAUAAAACAAFLx5FUy749M4CAAAqBgAAHAAk AAAAAAAAACAAAAAAAAAAcHJ1bmUtbGl2ZS1jYWNoZWQtaW1hZ2UuaHRtbAoAIAAAAAAAAQAYAOnS OqjLw88B6dI6qMvDzwENNM7JvMPPAVBLAQI/ABQAAAAAAFwtHkUAAAAAAAAAAAAAAAAKACQAAAAA AAAAEAAAAAgDAAByZXNvdXJjZXMvCgAgAAAAAAABABgARh4Mz8nDzwG2TF3PycPPAQOl18nJw88B UEsBAj8AFAAAAAgAriQeRTKb3y0iLAAAxjMBABcAJAAAAAAAAAAgAAAAMAMAAHJlc291cmNlcy9i aWctZ3JlZW4uanBnCgAgAAAAAAABABgAE3kbqsDDzwFf2drJycPPAU+y2snJw88BUEsFBgAAAAAD AAMAMwEAAIcvAAAAAA==