(Android) Fix UTF-8 related crashes

[mjmasn]

See: https://developer.android.com/reference/java/nio/charset/Charset.html#defaultCharset()

Android note: The Android platform default is always UTF-8.

Before this change, the following Exception would be thrown when streaming a file containing Arabic characters:

W System.err: java.nio.charset.MalformedInputException: Input length = 1
W System.err: 	at java.nio.charset.CoderResult.throwException(CoderResult.java:281)
W System.err: 	at java.nio.charset.CharsetEncoder.encode(CharsetEncoder.java:835)
W System.err: 	at com.RNFetchBlob.RNFetchBlobFS.readStream(RNFetchBlobFS.java:332)
W System.err: 	at com.RNFetchBlob.RNFetchBlob$11.run(RNFetchBlob.java:307)
W System.err: 	at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1167)
W System.err: 	at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:641)
W System.err: 	at java.lang.Thread.run(Thread.java:764)

Fixes: #297

Gave it a quick manual test a json file containing Arabic using:

RNFetchBlob.fetch('GET', 'http://localhost:8123/test.json')
  .then(res => console.log('res', res.json()))
  .catch(err => console.log('err', err));

and the json was logged correctly.

For the file stream change, again I tested this manually with a of react-native-local-mongodb I'm working on which loads data from files. Occasionally with a large enough file the app can hit OOM so I'm introducing streaming to reduce the likelihood of this. After this change the app now loads again.

[ronzilla]

FWIW, this works for any multi-byte UTF-8 on Android--Chinese, Japanese, Polish, symbols, etc. (we were seeing the same bug, this PR branch fixed the problem)

[erennyuksell]

is this PR tested well ? Should i merge it into https://github.com/Frekansapp/rn-fetch-blob here?

[Traviskn]

@mjmasn and @ronzilla Thank you for posting how you tested to confirm the fix works, apologies for the delayed review

[Traviskn]

This seems to have caused a regression, more info described in #442

[Traviskn]

I've reverted the merge of this PR and published a new version. If a new PR can be opened that fixes UTF-8 related crashes without breaking base 64 responses then I'd be glad to take another look!

[mjmasn]

That's unfortunate, I suspect specifying the encoding explicitly when calling fetch should work but either way it's still a regression so makes sense to have reverted this.

I don't have time to look at this for the foreseeable so if someone else wants to take it on, feel free!