Content is corrupted in case of utf8 multibyte characters

[Biktop]

I found that in version 5.1.0 downloaded cvs parsed incorrectly in case if source data contains utf8 multibyte characters (whereas 5.0.2 didn't have this issue). I noticed that PapaParse tried to download second chunk even if there no any extra bytes:

First chunk:

< Content-Length: 184027
< Content-Range: bytes 0-184026/184027

Second chunk (not necessary)

> Range: bytes=183546-5426425
< Content-Length: 481
< Content-Range: bytes 183546-184026/184027

After some digging I think that problem here
this._start += xhr.responseText.length;

Because Content-Length can be greater then text length (because some characters can be 2 or more bytes).

I think it should be something like this:
this._start += +xhr.getResponseHeader('Content-Length');

[pokoli]

I think this should be fixed by #745 could you please test it?

[Biktop]

I've tested it and it seems that it works!

[pokoli]

Ok, so I've merged #745 which fixes this

[Biktop]

Are going to publish this fix soon?

[pokoli]

I've just published 5.1.1 which contains this fix!