Open GoogleCodeExporter opened 8 years ago
[deleted comment]
the problem is inside svenson JSONTokenizer here is a link to an issue in that
project and a patch file that fixes some of the problems. The main problem is
large
Long values like Long.MAX_VALUE get serialized incorrectly, in the case of
Long.MAX_VALUE, 103 is being added to the value and then svenson is adding .0
to the
end which causes svenson to convert it to a Double instead of a Long. Here is
the
link to the patch.
http://code.google.com/p/svenson/issues/detail?id=14
Original comment by jarrod.r...@gmail.com
on 15 Apr 2010 at 10:35
Actually after more patching and testing, I found that Long.MAX_VALUE is getting
serialized to the wrong value, it is getting 103 added to it some how when it
gets
written to CouchDB, that causes the overflow inside of svenson. I send in
9223372036854775807 Long.MAX_VALUE in Java
and 9223372036854776000 is what gets stored in the document.
Original comment by jarrod.r...@gmail.com
on 19 Apr 2010 at 11:45
I submitted this bug to the couchdb team.
https://issues.apache.org/jira/browse/COUCHDB-749
Original comment by jarrod.r...@gmail.com
on 23 Apr 2010 at 5:27
Original issue reported on code.google.com by
jarrod.r...@gmail.com
on 14 Apr 2010 at 11:23