stevedlawrence commented on a change in pull request #158: Daffodil 1080
sequences and separators - preliminary review
URL: https://github.com/apache/incubator-daffodil/pull/158#discussion_r242158874
##########
File path:
daffodil-runtime1-unparser/src/main/scala/org/apache/daffodil/processors/unparsers/SeparatedSequenceUnparsers.scala
##########
@@ -392,7 +431,74 @@ class OrderedSeparatedSequenceUnparser(
Assert.invariantFailed("No event for unparsing.")
}
}
- //
+
+ case null => { // potentiallyTrailingGroupUnparser:
PotentiallyTrailingGroupSequenceChildUnparser => {
+ //
+ // We have to unparse the group, for side-effects potentially.
+ // however, we don't know whether to output the separator.
+ //
+ // The problem is that whether we put down this separator depends on
whether
+ // the children of this group are in fact all absent and so there's
nothing to output here, AND
+ // there's no subsequent siblings (i.e, this model group is truly
trailing empty.)
+ //
+ // This really is a case where we can't stream.
+ //
+ // Simpler DFDL implementations could implement this behavior by
some sort of backtracking on the output stream.
+ // They could use similar techniques for dealing with the storing of
prefix-lengths. They don't implement
Review comment:
should the word "implement' be removed here?
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
