Is it true about document chapter "5.1. Chunk-Oriented Processing"?

[soracane]

Hi.
I am not good at english, so I use google translator.
I say sorry about my strange english in advance.

I read document about "5.1. Chunk-Oriented Processing",
and I tried to write log about timing of calling ItemReader, ItemProcessor, ItemWriter.
(I made each custom classes, just only do System.out.println("xxx()"))

 read()
 read()
 read()
 process()
 process()
 process()
 write() size=3

But document says that each classes are called reader -> processor -> reader -> processor -> writer.
I think the document should be improved, if there are mistakes.(document and images)

To write Chapter "5.1. Chunk-Oriented Processing" accurately is important for Skip or Retry process.(I think)

I think below image is accurately image. Or do I mistake?
commit.JPG


Thank you.

[mminella]

Technically you are correct. This is due to the buffering of input so that it can be retried on a failure. If buffering is turned off, however, then the documentation is correct.

[soracane]

Technically you are correct. This is due to the buffering of input so that it can be retried on a failure. If buffering is turned off, however, then the documentation is correct.

Thank you for your reply.
Maybe you mentioned this(javadoc ChunkOrientedTasklet).
Is it correct?

Then I see what you mentioned.
And I know it is easy for Spring Batch beginners to understand normal process of the chunk tasklet.
...But if someone uses retry or skip process, they make a mistake.

I found thread below.
http://forum.springsource.org/showth...e-is-exception

Maybe a lot of people misunderstand skip , retry process.
I think that Chapter of "5.1.5. Configuring Skip Logic" had better to be added the information about something like this image.
SkipBehavior.JPG


But I understood why the document is written like that.
Thank you.