177459 2017-09-25 13:17:10 -0700 [CMake] ENABLE(DEVELOPER_MODE) should enable logging support for Release builds. 2018-06-18 00:45:14 -0700 1 1 1 Unclassified WebKit Tools / Tests WebKit Local Build Unspecified Unspecified NEW https://bugs.webkit.org/show_bug.cgi?id=177441 https://bugs.webkit.org/show_bug.cgi?id=180913 P2 Normal --- 1 clopez clopez achristensen annulen aperez benjamin bugs-noreply calvaris cdumez cmarcelo dbates ddkilzer ews-watchlist jfbastien joepeck lforschler mark.lam mcatanzaro tonikitoo oldest_to_newest 1352373 0 clopez 2017-09-25 13:17:10 -0700 Currently if one wants to enable WebKit logging support it has to use Debug builds or pass manually the flag with something like CXXFLAGS="-DLOG_DISABLED=0" Its actually documented here: https://trac.webkit.org/wiki/WebKitGTK/Debugging#Loggingsupport For debug builds, logging is enabled at build-time but disabled by default at run-time. To enable it at run-time the environment variable WEBKIT_DEBUG should be set to the right value And for release builds I think it makes sense to do the same when ENABLE(DEVELOPER_MODE) is set (enable it at build-time). Its bit of a pain not being able to have logging support in release development builds. 1352397 1 mcatanzaro 2017-09-25 13:57:53 -0700 It should also enable asserts (but we're not quite ready for that yet) 1389993 2 331491 clopez 2018-01-17 06:56:41 -0800 Created attachment 331491 Patch 1389994 3 annulen 2018-01-17 07:00:44 -0800 This may result in different performance of developer and non-developer release builds. Using -DRELEASE_LOG_DISABLED=0 should be safer 1389996 4 clopez 2018-01-17 07:18:31 -0800 (In reply to Konstantin Tokarev from comment #3) > This may result in different performance of developer and non-developer > release builds. I don't think this will impact performance in any sensible way. The feature is disabled at runtime by default, you have to set the environment variable WEBKIT_DEBUG to enable it. > Using -DRELEASE_LOG_DISABLED=0 should be safer RELEASE_LOG_DISABLED is another complete different thing not supported in the CMake ports (GTK and WPE currently). It uses os_log() which is only supported on Mac. Check https://developer.apple.com/documentation/os/os_log and bug 160969 that introduced it 1390027 5 331491 mcatanzaro 2018-01-17 09:01:22 -0800 Comment on attachment 331491 Patch View in context: https://bugs.webkit.org/attachment.cgi?id=331491&action=review > Source/cmake/OptionsCommon.cmake:99 > +if (DEVELOPER_MODE) > + # Explicit enable logging support (ensures its also enabled on release development builds). > + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DLOG_DISABLED=0") > +endif () Can you find a better way to do this? What is setting LOG_DISABLED in the first place? 1390028 6 annulen 2018-01-17 09:02:48 -0800 // wtf/Assertions.h #ifndef LOG_DISABLED #define LOG_DISABLED ASSERTIONS_DISABLED_DEFAULT #endif 1390030 7 annulen 2018-01-17 09:07:12 -0800 (In reply to Carlos Alberto Lopez Perez from comment #4) > RELEASE_LOG_DISABLED is another complete different thing not supported in > the CMake ports (GTK and WPE currently). It uses os_log() which is only > supported on Mac. Check https://developer.apple.com/documentation/os/os_log > and bug 160969 that introduced it We could define RELEASE_LOG macros to use "normal" logging. Though for me personally it would be really nice to see ltt-ng's tracelog() support built-in :) 1390033 8 annulen 2018-01-17 09:10:07 -0800 (In reply to Michael Catanzaro from comment #1) > It should also enable asserts (but we're not quite ready for that yet) Are you really ready to fix all WebCore assertion as they pop up when sites change their content? 1390042 9 clopez 2018-01-17 09:32:04 -0800 (In reply to Michael Catanzaro from comment #5) > Comment on attachment 331491 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=331491&action=review > > > Source/cmake/OptionsCommon.cmake:99 > > +if (DEVELOPER_MODE) > > + # Explicit enable logging support (ensures its also enabled on release development builds). > > + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DLOG_DISABLED=0") > > +endif () > > Can you find a better way to do this? What is setting LOG_DISABLED in the > first place? I don't think there is a better way to do this. On debug mode DLOG_DISABLED=0 gets set on the header because cmake defines -DNDEBUG=1 Can you please remove the r-? 1390044 10 clopez 2018-01-17 09:33:10 -0800 (In reply to Konstantin Tokarev from comment #7) > (In reply to Carlos Alberto Lopez Perez from comment #4) > > RELEASE_LOG_DISABLED is another complete different thing not supported in > > the CMake ports (GTK and WPE currently). It uses os_log() which is only > > supported on Mac. Check https://developer.apple.com/documentation/os/os_log > > and bug 160969 that introduced it > > We could define RELEASE_LOG macros to use "normal" logging. Though for me > personally it would be really nice to see ltt-ng's tracelog() support > built-in :) That would be a different thing that I want to do here. Please open a new bug for that if you wish. 1390045 11 clopez 2018-01-17 09:33:30 -0800 (In reply to Carlos Alberto Lopez Perez from comment #10) > (In reply to Konstantin Tokarev from comment #7) > > (In reply to Carlos Alberto Lopez Perez from comment #4) > > > RELEASE_LOG_DISABLED is another complete different thing not supported in > > > the CMake ports (GTK and WPE currently). It uses os_log() which is only > > > supported on Mac. Check https://developer.apple.com/documentation/os/os_log > > > and bug 160969 that introduced it > > > > We could define RELEASE_LOG macros to use "normal" logging. Though for me > > personally it would be really nice to see ltt-ng's tracelog() support > > built-in :) > > That would be a different thing that I want to do here. > Please open a new bug for that if you wish. I mean "That would be a different thing than what I want to do here." 1390048 12 clopez 2018-01-17 09:37:40 -0800 (In reply to Carlos Alberto Lopez Perez from comment #9) > (In reply to Michael Catanzaro from comment #5) > > Comment on attachment 331491 [details] > > Patch > > > > View in context: > > https://bugs.webkit.org/attachment.cgi?id=331491&action=review > > > > > Source/cmake/OptionsCommon.cmake:99 > > > +if (DEVELOPER_MODE) > > > + # Explicit enable logging support (ensures its also enabled on release development builds). > > > + set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -DLOG_DISABLED=0") > > > +endif () > > > > Can you find a better way to do this? What is setting LOG_DISABLED in the > > first place? > > I don't think there is a better way to do this. > On debug mode DLOG_DISABLED=0 gets set on the header because cmake defines > -DNDEBUG=1 > > Can you please remove the r-? Well, it is possible to set it on wtf/Assertions.h if DEVELOPER_MODE is set. I can do that if you prefer. 1390091 13 331516 clopez 2018-01-17 10:48:04 -0800 Created attachment 331516 Patch Patch v2, set logging on wtf/Assertions.h 1390129 14 331516 clopez 2018-01-17 11:31:14 -0800 Comment on attachment 331516 Patch The are more than LOG_DISABLED that should be set to 0. Will upload new patch 1390149 15 mcatanzaro 2018-01-17 12:02:26 -0800 (In reply to Carlos Alberto Lopez Perez from comment #12) > Well, it is possible to set it on wtf/Assertions.h if DEVELOPER_MODE is set. > I can do that if you prefer. That sounds better! (In reply to Konstantin Tokarev from comment #8) > Are you really ready to fix all WebCore assertion as they pop up when sites > change their content? Nope, but this is a proposal for DEVELOPER_MODE. Problem is, some of our developers like to use release builds, and don't notice when their code is broken because they aren't running assertions. Anyway, this a topic for another bug, another day.... 1390170 16 331534 clopez 2018-01-17 12:51:36 -0800 Created attachment 331534 Patch Patch v3, enable also error and fatal logs on developer builds 1390173 17 331534 clopez 2018-01-17 12:57:11 -0800 Comment on attachment 331534 Patch Will try to do this in Assertions.h rather than in the cmake file. 1390184 18 annulen 2018-01-17 13:30:42 -0800 (In reply to Michael Catanzaro from comment #15) > Problem is, some of our developers like to use release builds, and don't notice when their code is broken because they aren't running assertions. Same problem here. I think the only working solution (if WebCore's assert rate didn't decrease significantly between my branch and trunk) is having non-fatal assertions, just print message + backtrace and proceed. 1390189 19 annulen 2018-01-17 13:35:50 -0800 Also, I can name yet another problem: JSC in general and Inspector in particular are too slow in debug, so everyone tries to avoid debug builds. I didn't research it yet, but I guess there may be unneeded assertions on the hot paths(which could be removed or put under special ifdef guards), and there may be hot paths which badly lack compiler optimizations (maybe we need to build JSC with enabled optimization in all build types, in old times Qt port was using this trick) 1390193 20 clopez 2018-01-17 13:43:38 -0800 (In reply to Konstantin Tokarev from comment #19) > Also, I can name yet another problem: JSC in general and Inspector in > particular are too slow in debug, so everyone tries to avoid debug builds. I > didn't research it yet, but I guess there may be unneeded assertions on the > hot paths(which could be removed or put under special ifdef guards), and > there may be hot paths which badly lack compiler optimizations (maybe we > need to build JSC with enabled optimization in all build types, in old times > Qt port was using this trick) That is because Debug gets compiled with -O0. WebKit/JSC is painful slow without compiler optimizations. 331491 2018-01-17 06:56:41 -0800 2018-01-17 10:47:37 -0800 Patch bug-177459-20180117155640.patch text/plain 1396 clopez U3VidmVyc2lvbiBSZXZpc2lvbjogMjI3MDY2CmRpZmYgLS1naXQgYS9Tb3VyY2UvY21ha2UvT3B0 aW9uc0NvbW1vbi5jbWFrZSBiL1NvdXJjZS9jbWFrZS9PcHRpb25zQ29tbW9uLmNtYWtlCmluZGV4 IDE1OGFjYmQ0NmI3YTZlMzIxMWQzNjlmZWI3ZmU2YTI2MGRhZTEyYWQuLmNiNTE0NWIyZTAxYzJk OGZiMjcyZjMwMzA4NzgwMDhmZmYxYmZkNmEgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9jbWFrZS9PcHRp b25zQ29tbW9uLmNtYWtlCisrKyBiL1NvdXJjZS9jbWFrZS9PcHRpb25zQ29tbW9uLmNtYWtlCkBA IC05Myw2ICs5MywxMSBAQCBpZiAoREVCVUdfRklTU0lPTikKICAgICBzZXQoQ01BS0VfU0hBUkVE X0xJTktFUl9GTEFHUyAiJHtDTUFLRV9FWEVfTElOS0VSX0ZMQUdTfSAtV2wsLS1nZGItaW5kZXgi KQogZW5kaWYgKCkKIAoraWYgKERFVkVMT1BFUl9NT0RFKQorICAgICMgRXhwbGljaXQgZW5hYmxl IGxvZ2dpbmcgc3VwcG9ydCAoZW5zdXJlcyBpdHMgYWxzbyBlbmFibGVkIG9uIHJlbGVhc2UgZGV2 ZWxvcG1lbnQgYnVpbGRzKS4KKyAgICBzZXQoQ01BS0VfQ1hYX0ZMQUdTICIke0NNQUtFX0NYWF9G TEFHU30gLURMT0dfRElTQUJMRUQ9MCIpCitlbmRpZiAoKQorCiBpZiAoVVNFX0FSTV9MTFZNX0RJ U0FTU0VNQkxFUikKICAgICBmaW5kX3BhY2thZ2UoTExWTSBSRVFVSVJFRCkKICAgICBTRVRfQU5E X0VYUE9TRV9UT19CVUlMRChIQVZFX0xMVk0gVFJVRSkKZGlmZiAtLWdpdCBhL0NoYW5nZUxvZyBi L0NoYW5nZUxvZwppbmRleCBhNjBhNzgxNjMxY2RjMjAzMmI3YWNiYjEyMzQ0NDAxYmIwMmZlNDE4 Li4yNjkyZTI3ZWUyYTg1ZjdkN2E2NWFmZTkxZWE4YzY5ZjMyMjMzZDljIDEwMDY0NAotLS0gYS9D aGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMTIgQEAKKzIwMTgtMDEtMTcgIENh cmxvcyBBbGJlcnRvIExvcGV6IFBlcmV6ICA8Y2xvcGV6QGlnYWxpYS5jb20+CisKKyAgICAgICAg W0NNYWtlXSBFTkFCTEUoREVWRUxPUEVSX01PREUpIHNob3VsZCBlbmFibGUgbG9nZ2luZyBzdXBw b3J0IGZvciBSZWxlYXNlIGJ1aWxkcy4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcv c2hvd19idWcuY2dpP2lkPTE3NzQ1OQorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09Q UyEpLgorCisgICAgICAgICogU291cmNlL2NtYWtlL09wdGlvbnNDb21tb24uY21ha2U6IEVuc3Vy ZSB0aGF0IC1ETE9HX0RJU0FCTEVEPTAgZ2V0cyBkZWZpbmVkIHdoZW4gREVWRUxPUEVSX01PREUg aXMgc2V0LgorCiAyMDE4LTAxLTE3ICBDYXJsb3MgR2FyY2lhIENhbXBvcyAgPGNnYXJjaWFAaWdh bGlhLmNvbT4KIAogICAgICAgICBVbnJldmlld2VkLiBVcGRhdGUgT3B0aW9uc0dUSy5jbWFrZSBh bmQgTkVXUyBmb3IgMi4xOS42IHJlbGVhc2UuCg== 331516 2018-01-17 10:48:04 -0800 2018-01-17 12:51:22 -0800 Patch bug-177459-20180117194803.patch text/plain 1271 clopez U3VidmVyc2lvbiBSZXZpc2lvbjogMjI3MDcyCmRpZmYgLS1naXQgYS9Tb3VyY2UvV1RGL0NoYW5n ZUxvZyBiL1NvdXJjZS9XVEYvQ2hhbmdlTG9nCmluZGV4IGQ4ZTA2NTZlYjFkMjE4MTkwYzI0NWY2 MjEwMzRhNDBmNDY0MTE4ZWMuLjU3YjEwMmM3MWNjMzZmYzQwZjUwZmFlYTk3MDhiNjM1Y2VjNWMz NDMgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9XVEYvQ2hhbmdlTG9nCisrKyBiL1NvdXJjZS9XVEYvQ2hh bmdlTG9nCkBAIC0xLDMgKzEsMTIgQEAKKzIwMTgtMDEtMTcgIENhcmxvcyBBbGJlcnRvIExvcGV6 IFBlcmV6ICA8Y2xvcGV6QGlnYWxpYS5jb20+CisKKyAgICAgICAgW0NNYWtlXSBFTkFCTEUoREVW RUxPUEVSX01PREUpIHNob3VsZCBlbmFibGUgbG9nZ2luZyBzdXBwb3J0IGZvciBSZWxlYXNlIGJ1 aWxkcy4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE3 NzQ1OQorCisgICAgICAgIFJldmlld2VkIGJ5IE5PQk9EWSAoT09QUyEpLgorCisgICAgICAgICog d3RmL0Fzc2VydGlvbnMuaDogRW5hYmxlIGxvZ2dpbmcgc3VwcG9ydCBhbHNvIG9uIFJlbGVhc2Ug YnVpbGRzIHdoZW4gREVWRUxPUEVSX01PREUgaXMgZW5hYmxlZC4KKwogMjAxOC0wMS0xNyAgUGVy IEFybmUgVm9sbGFuICA8cHZvbGxhbkBhcHBsZS5jb20+CiAKICAgICAgICAgW1dpbl0gQ29tcGls ZSBlcnJvcjogJ3d0Zi90ZXh0L2ljdS9UZXh0QnJlYWtJdGVyYXRvcklDVS5oJyBub3QgZm91bmQu CmRpZmYgLS1naXQgYS9Tb3VyY2UvV1RGL3d0Zi9Bc3NlcnRpb25zLmggYi9Tb3VyY2UvV1RGL3d0 Zi9Bc3NlcnRpb25zLmgKaW5kZXggODU5MWJiNDg0ZmRiNzY0NWY5NGRiYjVkNWIxNjY1YTllNzIx NjRhYS4uNzA3YzNiN2QwMTIxOWFjM2Q4MWYwNjIwZTkxNGMzNjkxOTY0ODU3YiAxMDA2NDQKLS0t IGEvU291cmNlL1dURi93dGYvQXNzZXJ0aW9ucy5oCisrKyBiL1NvdXJjZS9XVEYvd3RmL0Fzc2Vy dGlvbnMuaApAQCAtOTUsOCArOTUsMTIgQEAgZXh0ZXJuICJDIiB2b2lkIF9SZWFkV3JpdGVCYXJy aWVyKHZvaWQpOwogI2VuZGlmCiAKICNpZm5kZWYgTE9HX0RJU0FCTEVECisjaWYgRU5BQkxFKERF VkVMT1BFUl9NT0RFKQorI2RlZmluZSBMT0dfRElTQUJMRUQgMAorI2Vsc2UKICNkZWZpbmUgTE9H X0RJU0FCTEVEIEFTU0VSVElPTlNfRElTQUJMRURfREVGQVVMVAogI2VuZGlmCisjZW5kaWYKIAog I2lmbmRlZiBSRUxFQVNFX0xPR19ESVNBQkxFRAogI2RlZmluZSBSRUxFQVNFX0xPR19ESVNBQkxF RCAhKFVTRShPU19MT0cpKQo= 331534 2018-01-17 12:51:36 -0800 2018-01-17 12:57:11 -0800 Patch bug-177459-20180117215134.patch text/plain 2168 clopez U3VidmVyc2lvbiBSZXZpc2lvbjogMjI3MDc2CmRpZmYgLS1naXQgYS9Tb3VyY2UvY21ha2UvT3B0 aW9uc0NvbW1vbi5jbWFrZSBiL1NvdXJjZS9jbWFrZS9PcHRpb25zQ29tbW9uLmNtYWtlCmluZGV4 IDE1OGFjYmQ0NmI3YTZlMzIxMWQzNjlmZWI3ZmU2YTI2MGRhZTEyYWQuLjA5NmVlMzg1Mjk3ZDky NjExZWIyZmEwNmNhNjFiODViNWMxOGU1YTYgMTAwNjQ0Ci0tLSBhL1NvdXJjZS9jbWFrZS9PcHRp b25zQ29tbW9uLmNtYWtlCisrKyBiL1NvdXJjZS9jbWFrZS9PcHRpb25zQ29tbW9uLmNtYWtlCkBA IC05Myw2ICs5MywxNCBAQCBpZiAoREVCVUdfRklTU0lPTikKICAgICBzZXQoQ01BS0VfU0hBUkVE X0xJTktFUl9GTEFHUyAiJHtDTUFLRV9FWEVfTElOS0VSX0ZMQUdTfSAtV2wsLS1nZGItaW5kZXgi KQogZW5kaWYgKCkKIAorIyBFbmFibGUgZXJyb3IgbG9ncyBhbmQgZGVidWcgbG9ncyBvbiBkZXZl bG9wZXIgYnVpbGRzLiBFcnJvciBhbmQgZmF0YWwgbG9ncyBoYXBwZW4gYWx3YXlzLgorIyBEZWJ1 ZyBsb2dzIG9ubHkgaGFwcGVuIGlmIHRoZSBlbnYgdmFyIFdFQktJVF9ERUJVRyBpcyBzZXQgdG8g dGhlIGxvZ2dpbmcgY2hhbm5lbCBvciB0byAiYWxsIi4KK2lmIChERVZFTE9QRVJfTU9ERSkKKyAg ICBTRVRfQU5EX0VYUE9TRV9UT19CVUlMRChMT0dfRElTQUJMRUQgRkFMU0UpCisgICAgU0VUX0FO RF9FWFBPU0VfVE9fQlVJTEQoRVJST1JfRElTQUJMRUQgRkFMU0UpCisgICAgU0VUX0FORF9FWFBP U0VfVE9fQlVJTEQoRkFUQUxfRElTQUJMRUQgRkFMU0UpCitlbmRpZiAoKQorCiBpZiAoVVNFX0FS TV9MTFZNX0RJU0FTU0VNQkxFUikKICAgICBmaW5kX3BhY2thZ2UoTExWTSBSRVFVSVJFRCkKICAg ICBTRVRfQU5EX0VYUE9TRV9UT19CVUlMRChIQVZFX0xMVk0gVFJVRSkKZGlmZiAtLWdpdCBhL0No YW5nZUxvZyBiL0NoYW5nZUxvZwppbmRleCBhNjBhNzgxNjMxY2RjMjAzMmI3YWNiYjEyMzQ0NDAx YmIwMmZlNDE4Li42MWZhZWViOTA3YTM3NzM3OTFlNTYxNGJkODM5YTA1NjU5NWUxYTE5IDEwMDY0 NAotLS0gYS9DaGFuZ2VMb2cKKysrIGIvQ2hhbmdlTG9nCkBAIC0xLDMgKzEsMjEgQEAKKzIwMTgt MDEtMTcgIENhcmxvcyBBbGJlcnRvIExvcGV6IFBlcmV6ICA8Y2xvcGV6QGlnYWxpYS5jb20+CisK KyAgICAgICAgW0NNYWtlXSBFTkFCTEUoREVWRUxPUEVSX01PREUpIHNob3VsZCBlbmFibGUgbG9n Z2luZyBzdXBwb3J0IGZvciBSZWxlYXNlIGJ1aWxkcy4KKyAgICAgICAgaHR0cHM6Ly9idWdzLndl YmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTE3NzQ1OQorCisgICAgICAgIFJldmlld2VkIGJ5IE5P Qk9EWSAoT09QUyEpLgorCisgICAgICAgIEN1cnJlbnRseSBsb2dnaW5nIHN1cHBvcnQgaXMgZW5h YmxlZCB1bmNvbmRpdGlvbmFsbHkgb24gRGVidWcgYnVpbGRzLgorICAgICAgICBUaGlzIHBhdGNo IGVuYWJsZXMgaXQgYWxzbyBvbiAocmVsZWFzZSkgZGV2ZWxvcGVyIGJ1aWxkcyB3aXRoIENNYWtl ICgtRERFVkVMT1BFUl9NT0RFPU9OKQorCisgICAgICAgIExvZ3MgZG9uZSB2aWEgTE9HKCkgYXJl IHByaW50IG9ubHkgaWYgdGhlIGVudmlyb25tZW50IFdFQktJVF9ERUJVRyBpcyBzZXQgdG8gdGhl IGxvZ2dpbmcgY2hhbm5lbCBvciB0byAiYWxsIgorICAgICAgICBMb2dzIGRvbmUgdmlhIExPR19F UlJPUigpIGFyZSBwcmludCBhbHdheXMuCisgICAgICAgIExvZ3MgZG9uZSB2aWEgRkFUQUwoKSBh cmUgcHJpbnQgYWx3YXlzIGFuZCB0aGV5IGFsc28gY3Jhc2ggdGhlIHByb2Nlc3MuIFRoaXMgaXMg b2ssIGFzIHRoaXMgb25seSBkZWZhdWx0cyB0bworICAgICAgICBUcnVlIG9uIGRldmVsb3BlciBi dWlsZHMgd2hlcmUgYSBGQVRBTCBzaG91bGQgbm90IGJlIGlnbm9yZWQgKGV2ZW4gb24gUmVsZWFz ZSkuCisgICAgICAgIEluIGFueSBjYXNlIEkgZ3JlcHBlZCB0aGUgc291cmNlIGNvZGUsIGFuZCBk b24ndCBzZWVtcyB0byBiZSBjdXJyZW50bHkgYW55IHVzYWdlIG9mIEZBVEFMKCkuCisKKyAgICAg ICAgKiBTb3VyY2UvY21ha2UvT3B0aW9uc0NvbW1vbi5jbWFrZToKKwogMjAxOC0wMS0xNyAgQ2Fy bG9zIEdhcmNpYSBDYW1wb3MgIDxjZ2FyY2lhQGlnYWxpYS5jb20+CiAKICAgICAgICAgVW5yZXZp ZXdlZC4gVXBkYXRlIE9wdGlvbnNHVEsuY21ha2UgYW5kIE5FV1MgZm9yIDIuMTkuNiByZWxlYXNl Lgo=