[
https://issues.apache.org/jira/browse/JENA-72?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13049950#comment-13049950
]
Andy Seaborne commented on JENA-72:
-----------------------------------
I've parameterized the tests and added more tests.
If the conclusion is to use this patch, I'll apply it but you (Paolo) ended
with a question with a question on JENA-69 so I'm not sure if this is a
suggested approach or a proposed fix.
IMHO, a custom parser is fine.
Changing Tokenizer(Text|Bytes) to (optionally) expose whitespace is also fine
but probably (1) much more work (2) odd in that it is not generally used.
> TSVInput infinite loop and optional binding values not handled
> --------------------------------------------------------------
>
> Key: JENA-72
> URL: https://issues.apache.org/jira/browse/JENA-72
> Project: Jena
> Issue Type: Bug
> Components: ARQ
> Reporter: Laurent Pellegrino
> Labels: tsvinput
> Attachments: JENA-72_ARQ_r1136132.patch
>
>
> TSVInput that has been proposed in JENA-69 contains two bugs:
> 1) When you attempt to read an empty ResultSet (i.e. a ResultSet that contain
> no binding) the code will enter in an infinite loop. This issue is really
> simple to solve.
> 2) Optional binding values are not supported, especially with the following
> ResultSet:
> (resultset (?a ?b ?c)
> (row (?a 1) (?c 4))
> (row (?a 1) (?b 2) (?c 3)))
> it will be serialized as:
> a....b....c
> 1..........4
> 1....2....3
> When we call tokenizer.next(), we get the next token. The problem is that we
> are not aware that the tokenizer has consumed two separators (\t), and we
> cannot detect that the value 4 is associated to the variable c.
> Any idea to handle it?
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
