196395 2019-03-29 04:16:52 -0700 REGRESSION(r243512): Change locale of test in intl-datetimeformat.js 2019-04-02 13:44:09 -0700 1 1 1 Unclassified WebKit New Bugs WebKit Nightly Build Unspecified Unspecified RESOLVED FIXED InRadar P2 Normal --- 1 dpino webkit-unassigned andy ap commit-queue mcatanzaro webkit-bug-importer oldest_to_newest 1522425 0 dpino 2019-03-29 04:16:52 -0700 REGRESSION(r243512): Change locale of test in intl-datetimeformat.js 1522426 1 366263 dpino 2019-03-29 04:18:33 -0700 Created attachment 366263 Patch 1522429 2 dpino 2019-03-29 04:20:51 -0700 r243512 modifies Intl.DateTimeFormat to obey 2-digit hour. A test was added to test 2-digit hour in Russian using 12-hour setting as true and false. In the former case, the expected string appends the string 'AM' in Russian. This localized string made GTK-based ports fail since according to Glib the 'am_pm' values in Russian are not localized (https://github.molgen.mpg.de/git-mirror/glibc/blob/master/localedata/locales/ru_RU#L152). This patch keeps the same test but changes locale to 'en' to prevent the test from failing. 1522432 3 dpino 2019-03-29 04:38:08 -0700 FWIW, the Chromium ticket that fixes this issue features the original test using the 'ru' locale, but 'am_pm' string is not localized. https://chromium-review.googlesource.com/c/v8/v8/+/1529260/3/test/intl/regress-527926.js 1522544 4 andy 2019-03-29 12:35:02 -0700 The original issue was about the ru locale, but I understand not relying on ru data being complete. I believe the '2-digit' hour was also sometimes ignored, so this is probably still a good enough test for us. 1522545 5 andy 2019-03-29 12:35:43 -0700 (In reply to Andy VanWagoner from comment #4) > The original issue was about the ru locale, but I understand not relying on > ru data being complete. I believe the '2-digit' hour was also sometimes > ignored, so this is probably still a good enough test for us. *was sometimes ignored in 'en'. 1522554 6 ap 2019-03-29 12:57:46 -0700 Perhaps the subtest can be split into a separate file, to have custom results for platforms that need them? 1522896 7 dpino 2019-04-01 04:28:36 -0700 I run the test in an old version of WebKit (WebKit v2.22.6), and the 2-digit issue happens also if 'en' locale is used: > Intl.DateTimeFormat('en', {minute: '2-digit', hour: '2-digit', hour12: true, timezone: 'UTC'}).format(1e7) < "3:46 AM" (See attachment) Regarding isolating the test in a different file (so there's different outputs per platform), the problem is the test is valid for GTK but it fails because of the localized string. So this is more about expecting a different actual result per platform. In fact, the original issue that r243512 solves states that valid outputs for this test are: 03:00 03:00 AM (or a Cyrillic alternative to "AM") https://bugs.webkit.org/show_bug.cgi?id=195974 (In r243512, 10e7 is used instead of a number value equivalent to '03:00 AM'). So an alternative solution would be to modify the test to expect all these possible valid outputs. For instance: shouldBeTrue("['02:46 ДП', '02:46 AM', '02:46'].includes(Intl.DateTimeFormat('ru', { minute:'2-digit', hour:'2-digit', hour12: true, timeZone: 'UTC' }).format(1e7))"); Whether to change the locale to 'en' or adapt the test to expect different valid outputs works for me, although I have a preference for changing the locale to a language where am_pm string is normally translated (as in English), so the output of the test is different than the same test where `hour12: false`. 1522899 8 366394 dpino 2019-04-01 04:33:04 -0700 Created attachment 366394 Screenshot of test executed in JavaScript console using 'en' locale 1522902 9 mcatanzaro 2019-04-01 05:41:32 -0700 (In reply to Alexey Proskuryakov from comment #6) > Perhaps the subtest can be split into a separate file, to have custom > results for platforms that need them? I think this would be a good idea if we had more platform-specific failures. This is a sensitive test where it's possible to test the thing we want to test in ways that work cross-platform and ways that don't. E.g. Andy has previously rewritten checks that failed for newer versions of ICU to not be sensitive to the ICU version. It could have been done with subtests, like you suggest, but making the test work more generally was fine too. I suspect it's easier to maintain as just one test, though, so r=me. 1522905 10 366263 mcatanzaro 2019-04-01 05:51:12 -0700 Comment on attachment 366263 Patch It might be more interesting to test a locale other than English, but like Andy says, the goal of this test was to show that hour:'2-digit' and hour:'numeric' give different output, so it seems fine. Why not just test 'pt-BR' though? Since it worked above, we already know it's localized properly on all the platforms WebKit cares about, right? 1523306 11 366263 commit-queue 2019-04-02 07:31:33 -0700 Comment on attachment 366263 Patch Clearing flags on attachment: 366263 Committed r243740: <https://trac.webkit.org/changeset/243740> 1523307 12 commit-queue 2019-04-02 07:31:34 -0700 All reviewed patches have been landed. Closing bug. 1523435 13 webkit-bug-importer 2019-04-02 13:44:09 -0700 <rdar://problem/49532466> 366263 2019-03-29 04:18:33 -0700 2019-04-02 07:31:33 -0700 Patch bug-196395-20190329111832.patch text/plain 5919 dpino U3VidmVyc2lvbiBSZXZpc2lvbjogMjQzNjQzCmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggYTM5MDQwMmE4YjE2Y2MwNGIyMDNi MWRkZWNlZjg5ZDdlZjlhNDM0My4uYjFkNTA0ZWY3YmNhMjZjOTg2YzRhNWYwZmJlZTZiNjVjYTFh NGZmNiAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDIzIEBACisyMDE5LTAzLTI5ICBEaWVnbyBQaW5vIEdhcmNp YSAgPGRwaW5vQGlnYWxpYS5jb20+CisKKyAgICAgICAgUkVHUkVTU0lPTihyMjQzNTEyKTogQ2hh bmdlIGxvY2FsZSBvZiB0ZXN0IGluIGludGwtZGF0ZXRpbWVmb3JtYXQuanMKKyAgICAgICAgaHR0 cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE5NjM5NQorCisgICAgICAgIHIy NDM1MTIgbW9kaWZlcyBJbnRsLkRhdGVUaW1lRm9ybWF0IHRvIG9iZXkgMi1kaWdpdCBob3VyLiBB IHRlc3Qgd2FzCisgICAgICAgIGFkZGVkIHRvIHRlc3QgMi1kaWdpdCBob3VyIGluIFJ1c3NpYW4g dXNpbmcgMTItaG91ciBzZXR0aW5nIGFzIHRydWUgYW5kCisgICAgICAgIGZhbHNlLiBJbiB0aGUg Zm9ybWVyIGNhc2UsIHRoZSBleHBlY3RlZCBzdHJpbmcgYXBwZW5kcyB0aGUgc3RyaW5nICdBTScK KyAgICAgICAgaW4gUnVzc2lhbi4gVGhpcyBsb2NhbGlzZWQgc3RyaW5nIG1hZGUgR1RLLWJhc2Vk IHBvcnRzIGZhaWwgc2luY2UKKyAgICAgICAgYWNjb3JkaW5nIHRvIEdsaWIgdGhlICdhbV9wbScg dmFsdWVzIGluIFJ1c3NpYW4gYXJlIG5vdCBsb2NhbGlzZWQuCisgICAgICAgIAorICAgICAgICBU aGlzIHBhdGNoIGtlZXBzIHRoZSBzYW1lIHRlc3QgYnV0IGNoYW5nZXMgbG9jYWxlIHRvICdlbicg dG8gcHJldmVudAorICAgICAgICB0aGUgdGVzdCBmcm9tIGZhaWxpbmcuCisKKyAgICAgICAgUmV2 aWV3ZWQgYnkgTk9CT0RZIChPT1BTISkuCisKKyAgICAgICAgKiBqcy9pbnRsLWRhdGV0aW1lZm9y bWF0LWV4cGVjdGVkLnR4dDoKKyAgICAgICAgKiBqcy9zY3JpcHQtdGVzdHMvaW50bC1kYXRldGlt ZWZvcm1hdC5qczoKKyAgICAgICAgKiBwbGF0Zm9ybS9ndGsvanMvaW50bC1kYXRldGltZWZvcm1h dC1leHBlY3RlZC50eHQ6CisKIDIwMTktMDMtMjkgIENhdGhpZSBDaGVuICA8Y2F0aGllY2hlbkBp Z2FsaWEuY29tPgogCiAgICAgICAgIEltcGxlbWVudCBSZXNpemVPYnNlcnZlci4KZGlmZiAtLWdp dCBhL0xheW91dFRlc3RzL2pzL2ludGwtZGF0ZXRpbWVmb3JtYXQtZXhwZWN0ZWQudHh0IGIvTGF5 b3V0VGVzdHMvanMvaW50bC1kYXRldGltZWZvcm1hdC1leHBlY3RlZC50eHQKaW5kZXggZDViMGIx OTkzMzAyMzZiYjNkNjBhNjRhNWYwZmFlZDg2MDc1Mzg4ZC4uNTVkYmI5OGFmMGMyZjlkZDI1MTll NDBmZGVlZjAzZDYyMGQ2NDFlNSAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvanMvaW50bC1kYXRl dGltZWZvcm1hdC1leHBlY3RlZC50eHQKKysrIGIvTGF5b3V0VGVzdHMvanMvaW50bC1kYXRldGlt ZWZvcm1hdC1leHBlY3RlZC50eHQKQEAgLTMyMSw4ICszMjEsOCBAQCBQQVNTIEludGwuRGF0ZVRp bWVGb3JtYXQoJ2VuJywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOidudW1lcmljJywgdGltZVpv bmU6ICdVVAogUEFTUyBJbnRsLkRhdGVUaW1lRm9ybWF0KCdwdC1CUicsIHsgbWludXRlOicyLWRp Z2l0JywgaG91cjonbnVtZXJpYycgfSkucmVzb2x2ZWRPcHRpb25zKCkuaG91ckN5Y2xlIGlzICdo MjMnCiBQQVNTIEludGwuRGF0ZVRpbWVGb3JtYXQoJ3B0LUJSJywgeyBtaW51dGU6JzItZGlnaXQn LCBob3VyOidudW1lcmljJyB9KS5yZXNvbHZlZE9wdGlvbnMoKS5ob3VyMTIgaXMgZmFsc2UKIFBB U1MgSW50bC5EYXRlVGltZUZvcm1hdCgncHQtQlInLCB7IG1pbnV0ZTonMi1kaWdpdCcsIGhvdXI6 J251bWVyaWMnLCB0aW1lWm9uZTogJ1VUQycgfSkuZm9ybWF0KDApIGlzICcwOjAwJwotUEFTUyBJ bnRsLkRhdGVUaW1lRm9ybWF0KCdydScsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdp dCcsIGhvdXIxMjogZmFsc2UsIHRpbWVab25lOiAnVVRDJyB9KS5mb3JtYXQoMCkgaXMgJzAwOjAw JwotUEFTUyBJbnRsLkRhdGVUaW1lRm9ybWF0KCdydScsIHsgbWludXRlOicyLWRpZ2l0JywgaG91 cjonMi1kaWdpdCcsIGhvdXIxMjogdHJ1ZSwgdGltZVpvbmU6ICdVVEMnIH0pLmZvcm1hdCgxZTcp IGlzICcwMjo0NiDQlNCfJworUEFTUyBJbnRsLkRhdGVUaW1lRm9ybWF0KCdlbicsIHsgbWludXRl OicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhvdXIxMjogZmFsc2UsIHRpbWVab25lOiAnVVRD JyB9KS5mb3JtYXQoMCkgaXMgJzAwOjAwJworUEFTUyBJbnRsLkRhdGVUaW1lRm9ybWF0KCdlbics IHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhvdXIxMjogdHJ1ZSwgdGltZVpv bmU6ICdVVEMnIH0pLmZvcm1hdCgxZTcpIGlzICcwMjo0NiBBTScKIFBBU1MgSW50bC5EYXRlVGlt ZUZvcm1hdCgnZW4nLCB7IG1pbnV0ZTogeyB0b1N0cmluZygpIHsgdGhyb3cgJ21pbnV0ZScgfSB9 IH0pIHRocmV3IGV4Y2VwdGlvbiBtaW51dGUuCiBQQVNTIEludGwuRGF0ZVRpbWVGb3JtYXQoJ2Vu JywgeyBtaW51dGU6bnVsbCB9KSB0aHJldyBleGNlcHRpb24gUmFuZ2VFcnJvcjogbWludXRlIG11 c3QgYmUgIjItZGlnaXQiIG9yICJudW1lcmljIi4KIFBBU1MgSW50bC5EYXRlVGltZUZvcm1hdCgn ZW4nKS5yZXNvbHZlZE9wdGlvbnMoKS5taW51dGUgaXMgdW5kZWZpbmVkCmRpZmYgLS1naXQgYS9M YXlvdXRUZXN0cy9qcy9zY3JpcHQtdGVzdHMvaW50bC1kYXRldGltZWZvcm1hdC5qcyBiL0xheW91 dFRlc3RzL2pzL3NjcmlwdC10ZXN0cy9pbnRsLWRhdGV0aW1lZm9ybWF0LmpzCmluZGV4IGI0NDMw MjZlZWE0MDE5N2U5OTU3MmViNmVmNGY5MzEwOWMzYzk0MDMuLmZlOWY0ZmUxMjkzY2FlOTViMWJi ZTljMmQ3YjRiODkwMDUzMzAwYTYgMTAwNjQ0Ci0tLSBhL0xheW91dFRlc3RzL2pzL3NjcmlwdC10 ZXN0cy9pbnRsLWRhdGV0aW1lZm9ybWF0LmpzCisrKyBiL0xheW91dFRlc3RzL2pzL3NjcmlwdC10 ZXN0cy9pbnRsLWRhdGV0aW1lZm9ybWF0LmpzCkBAIC00MzUsOCArNDM1LDggQEAgc2hvdWxkQmUo IkludGwuRGF0ZVRpbWVGb3JtYXQoJ2VuJywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOidudW1l cmljJywgdGltZVpvbmUKIHNob3VsZEJlKCJJbnRsLkRhdGVUaW1lRm9ybWF0KCdwdC1CUicsIHsg bWludXRlOicyLWRpZ2l0JywgaG91cjonbnVtZXJpYycgfSkucmVzb2x2ZWRPcHRpb25zKCkuaG91 ckN5Y2xlIiwgIidoMjMnIik7CiBzaG91bGRCZSgiSW50bC5EYXRlVGltZUZvcm1hdCgncHQtQlIn LCB7IG1pbnV0ZTonMi1kaWdpdCcsIGhvdXI6J251bWVyaWMnIH0pLnJlc29sdmVkT3B0aW9ucygp LmhvdXIxMiIsICJmYWxzZSIpOwogc2hvdWxkQmUoIkludGwuRGF0ZVRpbWVGb3JtYXQoJ3B0LUJS JywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOidudW1lcmljJywgdGltZVpvbmU6ICdVVEMnIH0p LmZvcm1hdCgwKSIsICInMDowMCciKTsKLXNob3VsZEJlKCJJbnRsLkRhdGVUaW1lRm9ybWF0KCdy dScsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhvdXIxMjogZmFsc2UsIHRp bWVab25lOiAnVVRDJyB9KS5mb3JtYXQoMCkiLCAiJzAwOjAwJyIpOwotc2hvdWxkQmUoIkludGwu RGF0ZVRpbWVGb3JtYXQoJ3J1JywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOicyLWRpZ2l0Jywg aG91cjEyOiB0cnVlLCB0aW1lWm9uZTogJ1VUQycgfSkuZm9ybWF0KDFlNykiLCAiJzAyOjQ2INCU 0J8nIik7CitzaG91bGRCZSgiSW50bC5EYXRlVGltZUZvcm1hdCgnZW4nLCB7IG1pbnV0ZTonMi1k aWdpdCcsIGhvdXI6JzItZGlnaXQnLCBob3VyMTI6IGZhbHNlLCB0aW1lWm9uZTogJ1VUQycgfSku Zm9ybWF0KDApIiwgIicwMDowMCciKTsKK3Nob3VsZEJlKCJJbnRsLkRhdGVUaW1lRm9ybWF0KCdl bicsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhvdXIxMjogdHJ1ZSwgdGlt ZVpvbmU6ICdVVEMnIH0pLmZvcm1hdCgxZTcpIiwgIicwMjo0NiBBTSciKTsKIAogc2hvdWxkVGhy b3coIkludGwuRGF0ZVRpbWVGb3JtYXQoJ2VuJywgeyBtaW51dGU6IHsgdG9TdHJpbmcoKSB7IHRo cm93ICdtaW51dGUnIH0gfSB9KSIsICInbWludXRlJyIpOwogc2hvdWxkVGhyb3coIkludGwuRGF0 ZVRpbWVGb3JtYXQoJ2VuJywgeyBtaW51dGU6bnVsbCB9KSIsICdcJ1JhbmdlRXJyb3I6IG1pbnV0 ZSBtdXN0IGJlICIyLWRpZ2l0IiBvciAibnVtZXJpYyJcJycpOwpkaWZmIC0tZ2l0IGEvTGF5b3V0 VGVzdHMvcGxhdGZvcm0vZ3RrL2pzL2ludGwtZGF0ZXRpbWVmb3JtYXQtZXhwZWN0ZWQudHh0IGIv TGF5b3V0VGVzdHMvcGxhdGZvcm0vZ3RrL2pzL2ludGwtZGF0ZXRpbWVmb3JtYXQtZXhwZWN0ZWQu dHh0CmluZGV4IDg3NjQ3Y2Q4Mzk0ZTE3NjE4N2ViNzE5ODk3ZDgzYTQ1ZGY5OWNmYWQuLjU1ZGJi OThhZjBjMmY5ZGQyNTE5ZTQwZmRlZWYwM2Q2MjBkNjQxZTUgMTAwNjQ0Ci0tLSBhL0xheW91dFRl c3RzL3BsYXRmb3JtL2d0ay9qcy9pbnRsLWRhdGV0aW1lZm9ybWF0LWV4cGVjdGVkLnR4dAorKysg Yi9MYXlvdXRUZXN0cy9wbGF0Zm9ybS9ndGsvanMvaW50bC1kYXRldGltZWZvcm1hdC1leHBlY3Rl ZC50eHQKQEAgLTMyMSw4ICszMjEsOCBAQCBQQVNTIEludGwuRGF0ZVRpbWVGb3JtYXQoJ2VuJywg eyBtaW51dGU6JzItZGlnaXQnLCBob3VyOidudW1lcmljJywgdGltZVpvbmU6ICdVVAogUEFTUyBJ bnRsLkRhdGVUaW1lRm9ybWF0KCdwdC1CUicsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonbnVt ZXJpYycgfSkucmVzb2x2ZWRPcHRpb25zKCkuaG91ckN5Y2xlIGlzICdoMjMnCiBQQVNTIEludGwu RGF0ZVRpbWVGb3JtYXQoJ3B0LUJSJywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOidudW1lcmlj JyB9KS5yZXNvbHZlZE9wdGlvbnMoKS5ob3VyMTIgaXMgZmFsc2UKIFBBU1MgSW50bC5EYXRlVGlt ZUZvcm1hdCgncHQtQlInLCB7IG1pbnV0ZTonMi1kaWdpdCcsIGhvdXI6J251bWVyaWMnLCB0aW1l Wm9uZTogJ1VUQycgfSkuZm9ybWF0KDApIGlzICcwOjAwJwotUEFTUyBJbnRsLkRhdGVUaW1lRm9y bWF0KCdydScsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhvdXIxMjogZmFs c2UsIHRpbWVab25lOiAnVVRDJyB9KS5mb3JtYXQoMCkgaXMgJzAwOjAwJwotRkFJTCBJbnRsLkRh dGVUaW1lRm9ybWF0KCdydScsIHsgbWludXRlOicyLWRpZ2l0JywgaG91cjonMi1kaWdpdCcsIGhv dXIxMjogdHJ1ZSwgdGltZVpvbmU6ICdVVEMnIH0pLmZvcm1hdCgxZTcpIHNob3VsZCBiZSAwMjo0 NiDQlNCfLiBXYXMgMDI6NDYgQU0uCitQQVNTIEludGwuRGF0ZVRpbWVGb3JtYXQoJ2VuJywgeyBt aW51dGU6JzItZGlnaXQnLCBob3VyOicyLWRpZ2l0JywgaG91cjEyOiBmYWxzZSwgdGltZVpvbmU6 ICdVVEMnIH0pLmZvcm1hdCgwKSBpcyAnMDA6MDAnCitQQVNTIEludGwuRGF0ZVRpbWVGb3JtYXQo J2VuJywgeyBtaW51dGU6JzItZGlnaXQnLCBob3VyOicyLWRpZ2l0JywgaG91cjEyOiB0cnVlLCB0 aW1lWm9uZTogJ1VUQycgfSkuZm9ybWF0KDFlNykgaXMgJzAyOjQ2IEFNJwogUEFTUyBJbnRsLkRh dGVUaW1lRm9ybWF0KCdlbicsIHsgbWludXRlOiB7IHRvU3RyaW5nKCkgeyB0aHJvdyAnbWludXRl JyB9IH0gfSkgdGhyZXcgZXhjZXB0aW9uIG1pbnV0ZS4KIFBBU1MgSW50bC5EYXRlVGltZUZvcm1h dCgnZW4nLCB7IG1pbnV0ZTpudWxsIH0pIHRocmV3IGV4Y2VwdGlvbiBSYW5nZUVycm9yOiBtaW51 dGUgbXVzdCBiZSAiMi1kaWdpdCIgb3IgIm51bWVyaWMiLgogUEFTUyBJbnRsLkRhdGVUaW1lRm9y bWF0KCdlbicpLnJlc29sdmVkT3B0aW9ucygpLm1pbnV0ZSBpcyB1bmRlZmluZWQK 366394 2019-04-01 04:33:04 -0700 2019-04-01 04:33:04 -0700 Screenshot of test executed in JavaScript console using 'en' locale 2-digit-hour-en.png image/png 5372 dpino iVBORw0KGgoAAAANSUhEUgAAAycAAAAuCAYAAAAyTmb8AAAABmJLR0QA/wD/AP+gvaeTAAAACXBI WXMAAAsTAAALEwEAmpwYAAAAB3RJTUUH4wQBCxoKTZX+DQAAAB1pVFh0Q29tbWVudAAAAAAAQ3Jl YXRlZCB3aXRoIEdJTVBkLmUHAAAUYElEQVR42u3ceXyU1b3H8c9MZpIQyMISYtgihEVQVlGEKC1o QaBCi2xVqOvV4tpel9va1l6XtrRXX25XrUtVNF6pCyibCIIoyhY0EFnNhgiEEJJAliHLzJz7x4Rm nSWThED4vl8vXy+HyTzPc36/c85zznOe57EYYwwiIiIiIiKtzKoQiIiIiIiIJiciIiIiIiKanIiI iIiIyBk9OSktLeWDDz5gw4YNrX5wO7bCmNehy1Ow0nX2BtkUw1VPwPuuc7OSrVgJU1PBWeffK7Kh xz9gi7v59tUS2zQn4cE3YcgbMH4dlJxDeW+JeJ5SvBN6rYCKZohZsL9r1vK5YPbTsMyl3Cq3ovNt 22rTge7PXQJ3JMOlL8NPvoTiuj9ywl+T4f4D0NgHnvfvhlHPQ7+X4NJ3IdV95ubC27in2TUhnmcy W91/SE9PJyMjg4yMDBwOBxMnTvS/FTc88ja8cxyyHdA9GoaMhA9GQkgAnf5v3oEZ10FSnanS0Eth 43CY8WIjziGFcMUrcDgKQg306QN/GwdDQwM7AXk7Fm82fQG37AZ3ZXXZQy3wq+nw61jP31giYfm9 YA9phoy54OZn4eMwiK46RlsXWHwt9Lc0shMHLM1wPL5iVpQFvyuCpZPqVzZrGMSEQ4yl+Sp0S2yz 8hAs6gC7pkNkI37XrHkPsn62VDy/2wm3boQTFqgMhQenwI1dmr6/YGPm9Xd+YtYS9aXFB1/N0W6V 2zM+t2UO2LQPfvs5XP0LeCSuhXPUiP6luepgSxxns/e7Z5GG6nzlARi0Hj6ZB30s4MyBiz+G5Jug ZIPv8cvBDLj9M9jnBFs7uG8S/Eec7/1ZO8ALc8GUwKS3YMcYuNxae9R53wQYtxQ23QhjbIE3jn9t hcunwZM9Wrn/DHLc46tNex1eH4dffgh7qyZiphzSO0HWbOjSlHiebZOTAQMGsG3bNvLz89m+fTuA /wmKFf40Dx46CIPXwJIbYeCp7FfCrFeh9yDIyYedBXDTVLi7q2dS83AyvH4E1iyEDlaYNwXubGIn a42Et2+DJDcsWQ3XfgrfTIYoYOMmuH8vhFqhLBJeuAZG2P0fS/Ze+NVmKAvx/Hf/ZJgZA6PHwu6x UNlQ2YG1n8Hv90PmMXjxP2FGVYeZtQVuOAzxRWASIP4YfGaFj38OPSze93cq3o/dALe2r1/2rD1w +2Zw2sERAr+bAj+L8nz3yQp47Tzo9x18VAwRfeHL8fCDn2MJNmYYWLQVRl8K5zfQG1jCIKYdRNf5 zmvZfdUlP9ts0oDBgCUEas5v/eXPW959lsEFs5+HG+6EySHg2A290+HANAgLsn6eUr4XOq+GpXfC +EactBuKpymDRzfAL6+HW6MgcyskfQWzp0E7P9tL2QK3fQvdYuDCMHBb/bcVgNQUuCUVOkXDj2Ph 5RJInwphvn4XQP/S7PXFAus3wLPfQ3Y53PdzmB/rp236yruPdmtXbttsbq3H4LbP4PwBMC6mTn8U QI4aXSf8xNNrHfRTPn/9UqP5Oc5mP9+64OFFsK7q8ndBIdhHwNdjPYMnr3n3c646nf21L77GL6YU 7lkJF8+A5d3gaCZMXQ8/mVV9Pve1P4vdUx8bWsAK7Qp3d4Jn0mH0QP+TBedhmLYadhwF2yr4ygb2 eFg8EWLPknGPy0eb9lUnrDGQfGP1tt9aAqkXQ+cA4xlsXWr9KzUNKC0tNa+++qpZsGCBWbBggVm5 cqVxu93Gn4ofjBnwmjG7a/6p05iZfzfmt4c8HzM2G9NnlTEVVV+7i4yZ8KIxX7q8bdSYa58xZoXT BMRZYEzSC9Xbc50wZtwzxqx0GmNcxizZZMy2qp2/9Z4x1++t/q23Y3GXGjPjFWM+Pun5nLvLmH6L jSnwV/YaMZj3lDHv1ShD9hZjeq0wpuiwMRf9w5gtFcbc9aIxH1T62Z/TmJueNuaVkvq7cZcYM+05 YxYWeT5npRiT8KExJ6q+37fRmM7PGfPEkdq/83UswcbMGGNMmTEznjHm/fKGc+UuM+b/9lTXBb+x 9lOXvG2zqXZ+ZcyA1bW36TNmPvLuswxOY2bVqOulu4zp+qExZc1QPysPG3P7ai/104dA4rlvozED 1/iPubvYmEnPGvOuw/N5+TJj4pYbU+6nrbhLjZn2jDFvlhhj3MYkv29M56XVcfEa6wD6l2atL1W5 nZ9pjNsYs3NDdW59tk0/effWbpXbtp9b4zbmb28Y8/CRxuUomDrhK55ej9NH+QLpl4Lhd8zQnOfb muOiHGPGvWbMKkcA51sf/fzp7q8rvjem70JjMt3V2x7yT2PS3L7HLyf3GRO70Jhsd5B9SKUxN/+v Mc8XNvzbo6nG9KzbR/jiMuaPrxrzVOHZPe5pqE0HWgdzdhozckXDbchbPIOtS62twQWgiIgI5syZ wzvvvEN+fj5paWnEx8czbNiw4C44hUJS1fJVfCSUHIHTdaugNQJiKyHfCYRBvB0eXww2O+TlQmR/ /8t8zjzYWgSH3oMFVVdvbHY4ZKBjE67KdY+GsFCIioR4K0TawOHys7+qz39cCE9UXZkMS4AvJ0HY UdjeEZ7v4Pn3XgkQmwrpbrjYChF2CE2AO+ICPxZswcUMwFUKh8Ohp837VZ5fXNCIWAdQlxraZlPW f99aAvceg3fmVV+pDiRmp7M9BFI/bfHwj/gg2q6feGbtgRvS4bmZ9eNT7zgLID0GLgn3fL6sJ9gP BrCCXggZUXBphKfSTekPD+xvpovhzVlfqnI7KcHTNnp3gpJ8T24rfbVNP9v01W6V27adW3+85SjY OtGcx9lS583Ter6tsYpw70q45hqYWLU85S/v3vr51uyvG+NkGVS0850rn/uzwWNXwhX/hB1T4aV+ tb+OjoGQvXACz+pHsCrPsnFP0G3FCU9ugnkzqsaBBBbP5u4LWu22rn9XOosFq9Vac4WlScvhrfVa MHcJ5IRCnM2zNHhLCjx7E4wPg2VL4eUAj9/eGd6aC4nN+XyE9VSs68TH3/683NZVYWn42Gv+c8eI hpPu7ViCjlntQ2jUH/ss++msSxaYNx16r4Vfp8CVY2vHzmv+mqk9uE0zxaxFllxh9Xp45AS8MBuG h9WoM7kwaTFkG7DHwcfTq5e3axbJ5Q54V7gtATy/dibwdpwBtE1feffWbpXbczS3PnLUUgI5zlrl a41+qSXOty54dim4k+De2Ebk3Vs/f7rjYgGru8aD2QZcAbS5dhHQzgF5pvq2rQIHREcE2F7d8OQX cPNc+H1cy5avzY97gKJM+LAjbI7hnNBg23E4HCxatIi8vDwABg8eHPSqSSBHYHVCSRCXjp058Ks1 sMfbIM4FyzZCbn8YEwLuMjjRAfqFgnHA8kPgMv6PxRYLFxfBe8c8nwv3w/1pLfcWhmD3Z4+FEcfh s1LP5+/3Q8F50L8Jo/lgYwYQ0h7iy+CAq+XLHogtO+DJ9IbvgfXlkl6Qd7zxvwtGpAVyyz3/vz0H Kmme+um3rTTS8lWwwApLf1Z/YGSLgzXzIeMO2HNt9eA1pCMkHoeUMs/n9fvrlM9bnYiGhCLYXu45 sa5KD+x3Te1fgq0vwbRNn3lvaj+o3J61uQ0mR0HXiSDj6a18gfRLQeWhCXkP5pyzZi182AueG1B7 wBTs+fZ099chUdCtCL52eD5n/wAF0dDdz+g5rAdcVQQvH/RcRMg/AFcnw1eBnowrIcsFl3tZFjlx AlyREN3EunuujHu+yoB+idCJxsWzbD8kLYQVZWfX5KTeZLKsrKzexGTSpElYLBafM+Sab+ua8lLg b+uyRMB1PeH2V6BnJAwdAc8OAtch+PlaOO6GfQ7Ykwx/scFjc2Bc1UZdJyB5D8waDwNr7MhdDNe9 DKEu6N4b3h0H7QF6wO0WmPUe9IuGGaPgv7bB2z1hbifvx2JtB09MhJs/gmVhnqjdNdH/VSRnLsz5 BI4YSC+DlGR4OgQemgmDfMUkyP1Z2sOT4+H292GhHRx2eH4ydGhCBQkNNmYAoXBVLKz5Aa7tE+AF hCDLHkhH+eJnED09yKu0jThJ+Mr7ZF8FCYGbh8Pd78O6TtA3HKKM/7YSSP301laC4TwCv9sBjliY mFnVkXSCt6f5vhJojYQ/jYDb3oSXYuBHncFeFkDMOsBDw2D+m/BaJxjfCcIcnitTPn8X5qd+tmR9 aUzb9JN3f5Tbtpfbmm/oyc0H94ewIgz+MAd+etx/joKpE0HF00f5/PblQebB23G6W+B868yF+1LB xMPkrKrBcDd490roGOT51nKa+2trNDw+Eu54A/7bCrSDv15TfxBb7zjD4alpcPtq6FfheVvXf06D sc20hLsuE0ZfUP+W0caW/WwZ9/hq0z/zVycM7C2AfoO9r8h4i+c3GZAXD+PCz67JicXUuV8rNTWV 1atXBz4xEfHhRCZckQLLZkNCK1ajyoMw4nP46DrP6xQboyIbLtgGX89s+F5PaXnfZUN+FxgdCWkb 4BZg8xUtdztQU+qLKLeiPJzr3EUwYRH8+VYYVWeGW5nnefXt329oO6++bc1xj9d4GvjLWxA5Fe4+ y24Hq1ctEhMT6d27N3FxcYwdO1YTE2mS6ET48164Zzt8MPw03jdfR8YBSBoV3EkwtAfM3wLjF0Js L1g8rmlXZaTx7CfhwX+BywblYbBgass+p9CU+iLKrSgP5ypTAvd8BNsc0H0kDK+79OaEJz+BUVfB aFvbjMFpHff4iqcLulwE152Fz6nUWzkRERERERFpDVaFQERERERENDkRERERERHR5ERERERERDQ5 ERERERERqcN24sQJRUFERERERFqd3tYlIiIiIiJnBN3WJSIiIiIimpyIiIiIiIhociIiIiIiIm1x cuLC8cwEvpq/AndeMmn9Z5KT25YeZWml8rkPkrLlAE7VUxERERE5lycnLpeLtLQ00tLS8P/MvIWQ qEhs0VFYOkRis0Vhi7JUfXeS0nd+w/bxY/n6J2PZNu03HPmuvBGHWE7pS9fyVeIvyS2qMW7PWkz6 zKv4+uor2XbNneTsLQt4onHyxTlsGX09h35wV29v1Xy+TLiS7/e4ACclT45jwyW/53ilv/K1xKSk iOyU9az7fA+5eXv5/NPP2X7IgScLBld5MXnZqaxbtobdJ/Q+AxERERFpG2zeJiYpKSkUFBQAkJCQ QHR0tM85ji06BltUJNgqsXXtSMipLect5/t/uun+r/XExhrKXp9F6vOf0/WZCf+eGblX3MbGP3Tk oq1/I8Zec7uGyk8fY/+hYcTEZNQYvGeT89tXCHlgMRdfGolr18ccKy4Cwv2X2LmL3LVxnH9vJblL Muh2T38804wQwvoaCj74hp4PGo5uNoR3sPgvnw+VOdv5Ymce7prTOHs3ho69kFgfa1buY5nsq+zN +LEh7NjmZthgNzu+zaG4WyKRxVl8s7OQiG5xxHbIVw0WERERkbY7Oak7MenatStRUVF+N2QdNpse PXthCYkl9i47ESFVX8TOZNCnM6tG60dwZJUQ3qcHNdcdLD3G0HVqBKF1Buzuva+SsWwgff7SnR8+ rjE5yfuc/MJEoj66ix2PH8XdZSy9HpsQUIHdKe9SmPBTek4uo2Du+5TMf4jIqgmR9bKf0n7zexSs heKLrqbDFyX+y+eDPX4YV8Y3PimW9tF0KDlEVk5HKkx7LBHnM3JU1ZdRiVwyBjDF7MsGl+qwiIiI iLQRVn8TkxEjRgBQWlrqe0DdZwJxl3QCa3c6zriCsNpbxvH6XLYMHs13300m8aaBtScnQ2+k/yOz ag/489eS/VQe8Y9eR7vQ2vsy+Uep+D4Drn2OoUuXMCApjYw/LqHS7x1OJRS+u5n211xOSNSPiE1Y R+7WGreYhV9G3Ji9ZP/PATpOH1L7GH2Wr3lZ2vdh1GXnE1JwjKKj37Ju3SZ2Hi6ttQIjIiIiItLW 1Fo52b17d72JidVqJS0tjYMHD3LhhReSkJAQxG5CiLgpmVE3lOBY9Gt23ZHM4DfnEe51gO+mfFUy BemHKbp+A5hiTublcXzuw9jefpTOEZHYulxMzGDPrWYRV40jLHkvJ11g93W7Vf4qjqw9TGnaFL5+ HEzxYZwh6+mdNLFqIhJK5Ow5RJe1J66fjewmBteZs4MNu4/VnlTY4hl6xSC6+JzcGEIiu9HvIjfH T7oZOtjC9g3b+b5TEr3DVWlFRERE5ByYnNR88N1qtWKxNP2hb7P7KVLvzqP3kj/TMao9oT26QN5R XDVWOcyON0hfHEGPh0+tnlgJu/51Lrm+6g8q1/PdmNeJTn6UzpFA+Bii293DsS3Hibm8A2WbNlKe MJlwn7dauSlf/i6Oaa8y8rExniWj4pXsnfge+YUT6FL1V5bE6+j/KFCxuunBjR/KuEbf1mUo2vMZ W51D+fGgqlyERRBudeJ0GcCiWisiIiIibX9yMmjQIEpLSyksLOTIkSOkpqYyfPhwBg8eTN++fYmI iGj0DiwDbyRx+gNkTrmcTBu4Q/tx3sMP0L7GRMIc3MjRpR3p+odZAT3LgX0IvZ6Yy76HppBSBkSN pvfT0wn1NW53ZZL7/mE6PXxJ9b1skT8idtSfOLziCF26nCkpsRDVdxgJ3+xg/foKTjrgi/WhtOs1 hAvaWzCl2aRsyaLYuCgrdsCWdeTYOzMwaRjdQlWhRUREROTsZTF13hPsdDpJSUmhsLAQgPPOO4/h w4c3yyqKNJL7ICkpboaP6oVN0RARERGRc21y0tAEJSkpyc+rhEVERERERFpgcnJqgrJz504AhgwZ gtVqVbREREREROT0T05EREREREROJy2HiIiIiIiIJiciIiIiIiKanIiIiIiIyBnFtmvnt4qCiIic eydAm52YjjF07Rqn1+WLiJwh9EC8iIickyrKy8nKzsJus9MlNlYBERHR5ERERKT1VFZWkpWZyYAL LlAwRETOAHrmREREzll2ux2ns1KBEBHR5ERERERERESTExERERER0eRERERERESkvv8H42JVCpQT +xYAAAAASUVORK5CYII=