Examples and tests incorrectly use ByteBuffer.array()
Description
Release Notes
None
Activity
Show:
Alex Baranau November 11, 2014 at 7:37 AM
@Andreas Neumann I think it is great addition to Bytes - it can be used in many places. At same time - it may makes sense to have in StreamEvent too - as it will make it much more straightforward (better experience) for developer. IDE autocomplete helps here..
Andreas Neumann November 11, 2014 at 7:05 AM
@Alex Baranau yes, that is better. I added (in CDAP) a Bytes.toString(ByteBuffer bb) method that does exactly that. Should it be a convenience method of StreamEvent? Perhaps... but is that necessary with the utility in Bytes?
Alex Baranau November 11, 2014 at 1:03 AM
@Andreas Neumann should it be simply
String eventBody = Charsets.UTF_8.decode(event.getBody()).toString();?
Also, I believe it should be a convenience method of StreamEvent: https://issues.cask.co/browse/CDAP-92
Fixed
Pinned fields
Click on the next to a field label to start pinning.
Created November 6, 2014 at 1:52 AM
Updated March 27, 2015 at 7:31 PM
Resolved November 19, 2014 at 1:34 AM
For example, when reading a stream event, sample apps do:
String eventBody = new String(event.getBody().array());This is incorrect, because it ignores the offset and limit of the byte buffer within its underlying array. This also does not specify any encoding and is therefore susceptible to failure based on the system encoding. Better assure it is UTF8.
Instead, it should be
String eventBody = new String(event.getBody().array(), event.getBody().arrayOffset(), event.getBody().limit(), "UTF-8");We should add a utility method to common.Bytes and use that everywhere.