[Base64] Codereview request for 8006315 8006530 AND 8006295

[Ulf Zibis]

Am 04.02.2013 17:05, schrieb Chris Hegarty:
> On 04/02/2013 14:20, Alan Bateman wrote:
>> On 01/02/2013 16:55, Xueming Shen wrote:
>>> Hi,
>>>
>>> For 8006295 it appears the more reasonable approach is to continue
>>> throw IOE
>>> but document this behavior clearly at the wrap(is) API. So the change
>>> is joined
>>> with these two.
>>>
>>> 8006295: Base64.Decoder.wrap(java.io.InputStream) returns InputStream
>>> which throws unspecified IOException on attempt to decode invalid
>>> Base64 byte stream
>>>
>>> http://cr.openjdk.java.net/~sherman/8006295_8006315_8006530/webrev/
>>>
>>> Thanks!
>>> -Sherman
>> In Decoder's javadoc then I would suggest "as if followed by padding
>> characters(s)" rather than "as if there are padding charcacter(s)
>> followed".
>>
>> In wrap's javadoc it currently reads:
>>
>> "If the input byte stream is not in valid Base64 encoding scheme then
>> IOException will be thrown when reading any invalid Base64 byte(s) from
>> the returned input stream."

I most like this, but adding some {@code ...} tags would help, at least for IOE. Maybe one of the 
terms "Base64" would be enough, like:
If the input byte stream is not of valid Base64 encoding scheme then {@code IOException} will be 
thrown by it's {@code read} methods when encountering any undecodable byte(s)

>>
>> I guess this okay but an alternative would be:
>>
>> "The {@code InputStream} methods that reads from the stream throw {@code
>> IOException} if bytes are read that cannot be decoded".

Better: ... that read ...  (no 's', typo ?)
Also: ... byte(s) ...

>
> OR
>
> "The {@code read} methods of the returned {@code InputStream} will throw {@code IOException} if it 
> encounters bytes that cannot be decoded".

My 2 ct:
... encounters any Base64-invalid byte(s)

-Ulf