67105 2011-08-28 22:46:20 -0700 REGRESSION: iframe load event resets the Japanese Input Method 2013-04-12 16:44:48 -0700 1 1 1 Unclassified WebKit Frames 528+ (Nightly build) Mac (Intel) OS X 10.7 RESOLVED DUPLICATE 59121 P2 Normal --- 0 naofumi webkit-unassigned ap oldest_to_newest 458073 0 105461 naofumi 2011-08-28 22:46:20 -0700 Created attachment 105461 Minimal HTML to reproduce the regression I have noticed the issue on all Safari builds (including the final released versions) on Mac OS X 10.7 (Lion). I have added an attachment that isolates the issue. To reproduce: 1. Unzip and access index.html from the browser. 2. Turn on the Japanese Input Method. 3. Enter Japanese text in the text area field. 4. Notice that whenever the iframe is refreshed, the Japanese text is reset. Japanese text is refreshed only when the frame content is loaded. Frames that don't load correctly (i.e. wrong URL) do not clear Japanese text. This was not an issue with the previous release of Safari, but since I only have Lion testing environment, I can't pin down exactly when the regression happened. I have tested all Webkit Nightlies that run on Lion, down to r90316, and have confirmed the regression. Real-world issues This is a problem when you have an iframe on a page that either refreshes or takes time to load. When the iframe is fully loaded, it will reset all text that has been entered in the Japanese Input Method. Since many webpages load advertisements as iframes, and since these are often slow or refreshed, the issue might be rather common. 458104 1 ap 2011-08-29 00:36:32 -0700 *** This bug has been marked as a duplicate of bug 59121 *** 458106 2 ap 2011-08-29 00:38:17 -0700 > Since many webpages load advertisements as iframes, and since these are often slow or refreshed, the issue might be rather common. Are you aware of any Web sites that are affected? That would help prioritize fixing this. 458113 3 naofumi 2011-08-29 00:50:42 -0700 I did notice some similar behavior on the Japanese Amazon website (http://amazon.co.jp). However, the Ad format (Flash, iframe /w JS, AJAX) changes depending on what ads are being featured. Also since this is a timing issue, it is difficult to replicate on external websites. Hence I'm not sure if what I saw on Amazon is the exact same issue, or something else. Simple answer, "I'm not sure". (In reply to comment #2) > > Since many webpages load advertisements as iframes, and since these are often slow or refreshed, the issue might be rather common. > > Are you aware of any Web sites that are affected? That would help prioritize fixing this. 875035 4 naofumi 2013-04-12 16:23:01 -0700 (In reply to comment #2) > > Since many webpages load advertisements as iframes, and since these are often slow or refreshed, the issue might be rather common. > > Are you aware of any Web sites that are affected? That would help prioritize fixing this. This issue is now rather severe. It happens when updating your status on Facebook, each time they load up new ads. 875045 5 ap 2013-04-12 16:44:48 -0700 Yes, thanks for the update! 105461 2011-08-28 22:46:20 -0700 2011-08-28 22:46:20 -0700 Minimal HTML to reproduce the regression iframe-meddling-with-japanese-input.zip application/zip 3212 naofumi UEsDBAoAAAAAADd0HT8AAAAAAAAAAAAAAAAkABAAaWZyYW1lLW1lZGRsaW5nLXdpdGgtamFwYW5l c2UtaW5wdXQvVVgMAJEkW066JFtO9QEUAFBLAwQUAAgACAA4dB0/AAAAAAAAAAAAAAAALQAQAGlm cmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0Ly5EU19TdG9yZVVYDACvJFtOuyRbTvUB FADtmD9PhDAYh39FBqJL4+TY0UWN36C54GZcjLtnOP+eBzkgceSj+5b+Tkm4G5wO9X1IeUrf8lIY KAWAmbXFJWABZIg2obKFjGVEMrDpc3zgHM9o8I7l9lyTI4z9EC9YocBiOP6qebhu62Yt8WPcoZbo WvYX0nOOUpxLS4036V2ikuPdeW6Y51Tij5JlLu0LnMkzKqVeSC0cF7It+wxP0vIqkUrKqr9K6B0i FVrJ+tuesaIoijItTFR2tN9hKIoyQcL7wdGe7qIN4wmdDs6xtKM93UUb9kvolM5oSzva0100X1qG iw/DK28WL8bSjvY/umVF+TccRNkw/1/tXv8rivKHMWl+m8/wtSAYEeZaJ+V+cwI4m2P8EZDEn4Un +I472tNdtH4IKMq++ARQSwcIKIP81ikBAAAEGAAAUEsDBAoAAAAAAEN0HT8AAAAAAAAAAAAAAAAJ ABAAX19NQUNPU1gvVVgMAM4kW07OJFtO9QEUAFBLAwQKAAAAAABDdB0/AAAAAAAAAAAAAAAALQAQ AF9fTUFDT1NYL2lmcmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0L1VYDADOJFtOziRb TvUBFABQSwMEFAAIAAgAOHQdPwAAAAAAAAAAAAAAADgAEABfX01BQ09TWC9pZnJhbWUtbWVkZGxp bmctd2l0aC1qYXBhbmVzZS1pbnB1dC8uXy5EU19TdG9yZVVYDACvJFtOuyRbTvUBFABjYBVjZ2Bi YPBNTFbwD1aIUIACkBgDJxAbAbECEIP4QQwgDkwFDgAAUEsHCPLF+CEoAAAAUgAAAFBLAwQUAAgA CABMcR0/AAAAAAAAAAAAAAAAMAAQAGlmcmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0 L2ktZnJhbWUuaHRtbFVYDADOJFtOQB9bTvUBFABtj8FOwzAMhs/0KUzuram2A2JpDrRDIA2YUBDi GFaXRGqbkXiUvT3tChIHJEuWf/v/bMvz6rHUr9s1WO5a2D5fb+5KECniy6JErHQFt/p+A8vsIgcd TB8dO9+bFnH9IBIAEJZ5f4U4DEM2LDIf3lE/4URbYut9pKzmWqgkkZOmxkSmVsmZ7IgNTOaUPg7u sxCl75l6TvVxTwJ2c1UIpi8+8VawsyZE4uLATXop/oUEagJF+8efr06T7Lgl5W6C6ei3KXFWE4nz VfLN10c1viVtDq4uhFDaughjsCVwzcnN1jD8LKKxyZHaJhsZ+USaERLnd78BUEsHCNV1LDz0AAAA YQEAAFBLAwQUAAgACABMcR0/AAAAAAAAAAAAAAAAOwAQAF9fTUFDT1NYL2lmcmFtZS1tZWRkbGlu Zy13aXRoLWphcGFuZXNlLWlucHV0Ly5faS1mcmFtZS5odG1sVVgMAM4kW05AH1tO9QEUAGNgFWNn YGJg8E1MVvAPVohQgAKQGAMnEBsBcQcQg/i7GIgCjiEhQVAmSMcMIFZCU8KIEBdNzs/Vy01MLsrP TSxJLdZLTixKLWGo5gK5Ijk/pzQ3T8FWwdjCGiyQk5mXCuSaWXPVAgBQSwcIR92hw2cAAAC6AAAA UEsDBBQACAAIABp0HT8AAAAAAAAAAAAAAAAuABAAaWZyYW1lLW1lZGRsaW5nLXdpdGgtamFwYW5l c2UtaW5wdXQvaW5kZXguaHRtbFVYDADOJFtOgyRbTvUBFACVU9Fu0zAUfaZfcclLN4nWzVZ1XUkq QTdpQxubWNDgCbn1TWPh2sW+WShP+1d+hOtkSEVITLwksX3O8Tn33mQvz24Wxefbc6hoY+D249ur ywUkAyHujxdCnBVncFFcX8F4OEqh8NIGTdpZaYQ4f5/0ACCpiLYzIZqmGTbHQ+fXovggotpYGOcC DhWpZN7rZXFvzi+Uat57kW2QJETyAL/V+iFPFs4SWhoUuy0msOpWeUL4nVq917CqpA9IeU3lYJpE EdJkcF5gIHAl6NLLDcIGlTLarqHRVME7uZUWA8Kl3daUiY7Sy0RnJFs6tZtzkqxK/0+I8ZG2jU+A c7brIZoFbckBVQhLNK7p9qRHCXWIYvHkTy24RqqcGrZCb0KLeLLgsfQYKgygKaAp4QAf0O8gIBdI Hb4C60ivkCmSL7E7MEhsJLRS7WYlHzCi2A1aKDU3T/9AxS7be/YNwIG2Sq8k8fFyx3JQW4WeS4CH MQG74fJHm5nY7ofPYsY2omXTXc/iMgHvmpAn09hPwx/jUdIxADqi+A2d76k+ZdcqT7rPL2nCPVBU 5f3xZNqHCvW6oryfjkZ9CH7FuEELHMZBSTqtjtq19qhtLce6x+VXTWAj3+xgWWujAhzcWLiTpfQa ruXq5g4+wRWP+SE3+agVcOYpqNFzf5qm08kMLmTggqy5JoGxmeCjPcxoMjl5FnOcTmY/Hx//iZqe TtOjGSxcbZTtE49TnFHbGvwLenLyHDQTMUsmurHPRPdX/gJQSwcIPf7qHzoCAAAIBAAAUEsDBBQA CAAIABp0HT8AAAAAAAAAAAAAAAA5ABAAX19NQUNPU1gvaWZyYW1lLW1lZGRsaW5nLXdpdGgtamFw YW5lc2UtaW5wdXQvLl9pbmRleC5odG1sVVgMAM4kW06DJFtO9QEUAGNgFWNnYGJg8E1MVvAPVohQ gAKQGAMnEBsBcQcQg/i7GIgCjiEhQVAmSMcMIFZCU8KIEBdNzs/Vy01MLsrPTSxJLdZLTixKLWGo 5gK5Ijk/pzQ3T8FWwcAazM/JzEsF8owsrLlqAVBLBwidZ93RZgAAALoAAABQSwECFQMKAAAAAAA3 dB0/AAAAAAAAAAAAAAAAJAAMAAAAAAAAAABA7UEAAAAAaWZyYW1lLW1lZGRsaW5nLXdpdGgtamFw YW5lc2UtaW5wdXQvVVgIAJEkW066JFtOUEsBAhUDFAAIAAgAOHQdPyiD/NYpAQAABBgAAC0ADAAA AAAAAAAAQKSBUgAAAGlmcmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0Ly5EU19TdG9y ZVVYCACvJFtOuyRbTlBLAQIVAwoAAAAAAEN0HT8AAAAAAAAAAAAAAAAJAAwAAAAAAAAAAED9QeYB AABfX01BQ09TWC9VWAgAziRbTs4kW05QSwECFQMKAAAAAABDdB0/AAAAAAAAAAAAAAAALQAMAAAA AAAAAABA/UEdAgAAX19NQUNPU1gvaWZyYW1lLW1lZGRsaW5nLXdpdGgtamFwYW5lc2UtaW5wdXQv VVgIAM4kW07OJFtOUEsBAhUDFAAIAAgAOHQdP/LF+CEoAAAAUgAAADgADAAAAAAAAAAAQKSBeAIA AF9fTUFDT1NYL2lmcmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0Ly5fLkRTX1N0b3Jl VVgIAK8kW067JFtOUEsBAhUDFAAIAAgATHEdP9V1LDz0AAAAYQEAADAADAAAAAAAAAAAQKSBFgMA AGlmcmFtZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0L2ktZnJhbWUuaHRtbFVYCADOJFtO QB9bTlBLAQIVAxQACAAIAExxHT9H3aHDZwAAALoAAAA7AAwAAAAAAAAAAECkgXgEAABfX01BQ09T WC9pZnJhbWUtbWVkZGxpbmctd2l0aC1qYXBhbmVzZS1pbnB1dC8uX2ktZnJhbWUuaHRtbFVYCADO JFtOQB9bTlBLAQIVAxQACAAIABp0HT89/uofOgIAAAgEAAAuAAwAAAAAAAAAAECkgVgFAABpZnJh bWUtbWVkZGxpbmctd2l0aC1qYXBhbmVzZS1pbnB1dC9pbmRleC5odG1sVVgIAM4kW06DJFtOUEsB AhUDFAAIAAgAGnQdP51n3dFmAAAAugAAADkADAAAAAAAAAAAQKSB/gcAAF9fTUFDT1NYL2lmcmFt ZS1tZWRkbGluZy13aXRoLWphcGFuZXNlLWlucHV0Ly5faW5kZXguaHRtbFVYCADOJFtOgyRbTlBL BQYAAAAACQAJAJsDAADbCAAAAAA=