Some utf-8 strings base64 encoded by php can not be decoded using iOS base64 library?

Question

Here is one piece of Chinese utf-8 text which is encoded by PHP on the server-side, but when I decode it with iOS, it returns null.

I also tried this online tool where text can be decoded well.

  NSData *decodedData = [[NSData alloc] initWithBase64EncodedString:content options:0];
  content = [[NSString alloc] initWithData:decodedData encoding:NSUTF8StringEncoding];

5aW96ZuF77yM5ZKx5p2l5LiA5L+X55qE77yM5pS56Ieq5Lic5Y2X6KW/5YyX6aOO44CCCuS4juS9oOebuOmAou+8jOWFqOaYr+acuue8mOW3p+WQiOOAguWPr+Wtpui1t+adpeWNtOW/g+aGlOaCtOOAggrmgLvmmK/ovpPkuoborqnlho3ljrvlrabvvIzlrabkuobo

Here is the test code for debug this issue with xcode:

    NSString * = @"5aW96ZuF77yM5ZKx5p2l5LiA5L+X55qE77yM5pS56Ieq5Lic5Y2X6KW/5YyX6aOO44CCCuS4juS9oOebuOmAou+8jOWFqOaYr+acuue8mOW3p+WQiOOAguWPr+Wtpui1t+adpeWNtOW/g+aGlOaCtOOAggrmgLvmmK/ovpPkuoborqnlho3ljrvlrabvvIzlrabkuobo";

//
    NSData *decodedData = [[NSData alloc] initWithBase64EncodedString: options:0];
    NSString *content = [[NSString alloc] initWithData:decodedData encoding:NSUTF8StringEncoding] ;

    NSLog(content);

Answer 1

Your revised question features a base64 string of:

5aW96ZuF77yM5ZKx5p2l5LiA5L+X55qE77yM5pS56Ieq5Lic5Y2X6KW/5YyX6aOO44CCCuS4juS9oOebuOmAou+8jOWFqOaYr+acuue8mOW3p+WQiOOAguWPr+Wtpui1t+adpeWNtOW/g+aGlOaCtOOAggrmgLvmmK/ovpPkuoborqnlho3ljrvlrabvvIzlrabkuobo

This string has a length that is a multiple of four bytes, so the lack of the =/== terminator at the end is not the problem. And, in fact, initWithBase64EncodedString decodes it successfully:

e5a5bde9 9b85efbc 8ce592b1 e69da5e4 b880e4bf 97e79a84 efbc8ce6 94b9e887 
aae4b89c e58d97e8 a5bfe58c 97e9a38e e380820a e4b88ee4 bda0e79b b8e980a2 
efbc8ce5 85a8e698 afe69cba e7bc98e5 b7a7e590 88e38082 e58fafe5 ada6e8b5 
b7e69da5 e58db4e5 bf83e686 94e682b4 e380820a e680bbe6 98afe8be 93e4ba86 
e8aea9e5 868de58e bbe5ada6 efbc8ce5 ada6e4ba 86e8

The issue here is that this appears to not be a valid UTF8 string. In fact, when I run it through the http://base64decode.net site you referenced in your original question, it is also unable to convert it to a UTF8 string (I notice that your screen snapshots are using a different converter web site). When I ran it through another converter, it converted what it could, but then complained about the character following 学了 (which is, coincidentally, the character at which your base64 converter web site stopped, too).

By the way, the UTF8 representation of 了 is e4 ba 86. And you'll see that near the end of the hex representation of your base 64 string, followed by one more byte, e8. The thing is, e8, by itself, is not a valid UTF8 character. It almost looks looks like you took a base64 encoded string and just grabbed the first 200 bytes, disregarding whether that resulted in cutting the UTF8 character off in the middle or not.

---

The original question featured a base64 string of:

5aW96ZuF77yM5ZKx5p2l5LiA5L+X55qE77yM5pS56Ieq5Lic5Y2X6KW/5YyX6aOO44CCCuS4juS9oOebuOmAou+8jOWFqOaYr+acuue8mOW3p+WQiOOAguWPr+Wtpui1t+adpeWNtOW/g+aGlOaCtOOAggrmgLvmmK/

That is not valid base64. It should be a multiple of four bytes in length, but that is only 163 characters, which is missing a character. Either your server isn't properly terminating the base64 string, or it got cut off for some reason.

For example, if I add a = to get it up to 164 characters, I get a valid base64 string:

5aW96ZuF77yM5ZKx5p2l5LiA5L+X55qE77yM5pS56Ieq5Lic5Y2X6KW/5YyX6aOO44CCCuS4juS9oOebuOmAou+8jOWFqOaYr+acuue8mOW3p+WQiOOAguWPr+Wtpui1t+adpeWNtOW/g+aGlOaCtOOAggrmgLvmmK/=

Adding the = would be the right solution if the server simply neglected to terminate the base64 string properly. Anyway, that can be base64-decoded to:

好雅，咱来一俗的，改自东南西北风。
与你相逢，全是机缘巧合。可学起来却心憔悴。
总是

Is that what you were expecting?

Perhaps you should take a look at your base64 routine on your server? Or if it's getting truncated, look at how you are receiving it and compare the server's original base64 string length to what you have here.

For information about adding = or == to the end of a base 64 encoded string, see the base64 wikipedia page.

Answer 2

@Rob is right.

Check this Check NSData won't accept valid base64 encoded string

But in case if your server is not returning valid JSON with "=" or "==" then you need to use external methods to perform base64decode. Those methods can decode even if base64string does not have "=" symbol at the end.