stevedlawrence commented on a change in pull request #517:
URL: https://github.com/apache/daffodil/pull/517#discussion_r601708496
##########
File path:
daffodil-test/src/test/resources/org/apache/daffodil/section14/sequence_groups/SequenceGroupNestedArray.tdml
##########
@@ -219,6 +219,7 @@ notice blank lines are skipped
<ex:record>
<ex:item>notice lines of all commas are skipped</ex:item>
</ex:record>
+ <ex:record />
Review comment:
I managed to get the ibm DFDL crosstester working, and it has the same
behavior as Daffodil--there is an empty ``<ex:record />`` element when a line
is all commas. Likewise, empty lines that contain no commas do not create a
record element at all. So I think we're are consistent with IBM DFDL and are
interpreting the spec the same. The other ``test_csv_hang_*`` tests pass as
well.
Unfortunately, ``test_csv_nohang_1`` does not pass in IBM DFDL. It fails
with a parse error with message
> Unexpected data found at offset '94' after parsing completed. Data:
'0x2c...'
The 0x2c byte at offset 94 is the first comma on the line
```
,preceded by an empty field,
```
So it seems IBM does not like a line starting with a zero-length field in
the nohang test that has ``minOccurs="1"``. It feels like Daffodil has the
correct behavior, but this area of the spec is complicated, so I can't say for
sure.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
