Furthermore, it used to clobber first, then merge: https://github.com/apache/cordova-js/commit/99fadd12f74703a58e729664394570956d6ccbe4#lib/scripts/bootstrap.js
Ok, I'd like to change this but this requires everyone to grab a new cordova-js and re-test -- or I just tag iOS now with this fail and leave the fix for 2.3.0 final. What does everyone think? On Thu, Dec 13, 2012 at 2:20 PM, Shazron <[email protected]> wrote: > Found the culprit -- these two lines have to be reversed (which makes > sense -- clobber first, then merge, not the other way around): > > https://github.com/apache/cordova-js/blob/99fadd12f74703a58e729664394570956d6ccbe4/lib/scripts/bootstrap.js#L48-49 > > Unfortunately, this is an all-platform change to cordova-js :/ > > > On Thu, Dec 13, 2012 at 2:07 PM, Shazron <[email protected]> wrote: > >> Becky and I were debugging this. In lib/ios/platform.js, it seems >> FileReader is clobbered in the "clobbers" section but the iOS specific >> FileReader is not merged in the "merges" section. Not sure why yet. >> >> >> On Thu, Dec 13, 2012 at 1:22 PM, Becky Gibson <[email protected]>wrote: >> >>> So, it seems that the iOS specific version of readAsText is NOT being >>> called and the "standard" one is being used. The iOS version of >>> readAsText >>> performs the necessary decode. The iOS version IS in the cordova.js file >>> but I haven't debugged enough to see why it is not being used. >>> >>> -becky >>> >>> >>> On Thu, Dec 13, 2012 at 4:08 PM, Shazron <[email protected]> wrote: >>> >>> > Thanks Becky -- i'll take a peek as well. Filed: >>> > http://issues.cordova.io/2058 >>> > >>> > >>> > On Thu, Dec 13, 2012 at 12:18 PM, Becky Gibson <[email protected] >>> > >wrote: >>> > >>> > > So, I thought the encoding was necessary for internationalization >>> and any >>> > > special characters in the file data (for example a URL)? Do all of >>> the >>> > > other file tests still run with this change? We're there any >>> changes to >>> > > the file tests themselves? I try to take a look this afternoon. >>> > > >>> > > - Becky >>> > > >>> > > Sent from my iPad >>> > > >>> > > On Dec 13, 2012, at 2:15 PM, Marcel Kinard <[email protected]> >>> wrote: >>> > > >>> > > > Yup, I see that file-transfer.jitsu.com is returning a page >>> saying: >>> > > > >>> > > > An error has occurred: >>> > > > {"code":"ECONNREFUSED","errno":"ECONNREFUSED","syscall":"connect"} >>> > > > >>> > > > Something is obviously dead there. So yeah, ignore those test >>> failures >>> > as >>> > > > an infrastructure problem. >>> > > > >>> > > > Your File API failures look like the ones I was getting. If I >>> modify >>> > > > readAsText: from >>> > > > >>> > > > messageAsString:[pNStrBuff >>> > > > stringByAddingPercentEscapesUsingEncoding:NSUTF8StringEncoding] >>> > > > >>> > > > to >>> > > > >>> > > > messageAsString:[NSString stringWithString:pNStrBuff] >>> > > > then those failures disappear. So I'm still puzzled if it is really >>> > > > supposed to be URL encoding the result message. >>> > > > >>> > > > BTW, I noticed in your screenshot that your testsuite has 95 tests. >>> > Mine >>> > > > has 238. Are we using the same version of mobile-spec? >>> > > > >>> > > > -- Marcel >>> > > >>> > >>> >> >> >
