Michael Beckerle created DAFFODIL-1950:
------------------------------------------
Summary: Documentation/Example of how to use JAPI in "no extra
data allowed" mode
Key: DAFFODIL-1950
URL: https://issues.apache.org/jira/browse/DAFFODIL-1950
Project: Daffodil
Issue Type: New Feature
Components: Documentation
Affects Versions: 2.1.0
Reporter: Michael Beckerle
Fix For: 2.2.0
This code may exist in our JAPI unit tests already.
A common usage pattern for our user community is to parse, inspect, sanitize,
and reassemble (unparse) data to provide guaranteed spec-compliant clean data
to applications.
Left over data past the normal end marker or length of a file is a common thing
people want to eliminate.
However, whether left over data is a problem or not depends on the application
need. For some applications, e.g, message streaming, the "left over" data is
really just the first part of the next message, and must be where we pick up to
keep parsing after finishing a first message.
Since both (and perhaps other) modes of operation are reasonable and expected,
we need examples of how to implement these patterns, and they need to be
implementable from our Java API.
There may already be unit tests illustrating this in the JAPI unit test set.
Our TDML runner certaily does that, but it is written in Scala. If so then
this issue is really just requesting a Wiki documentation page highlighting
this code and the usage pattern of interest.
This issue may be subsumed by the Message streaming API issue, which is adding
a new mode of operation where left-over data is expected, and doesn't indicate
any problem with the data stream. (DAFFODIL-1565, DAFFODIL-1065)
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
