[
https://issues.apache.org/jira/browse/LUCENENET-226?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
George Aroush resolved LUCENENET-226.
-------------------------------------
Resolution: Fixed
Thanks for the patch Mike!
I committed it without your changes to TestFieldCacheSanityChecker.cs. Your
comment in the issue about the change you made to this file made me debug this
further and I found that we have a different formatting for both double and
float like so:
double: [java] "1.7976931348623157E308" vs. [.NET]
"1.79769313486231570000E+308"
float: [Java] "3.4028235E38" vs. [.NET] "3.402823E+38"
The double format for .NET is with your change of using "E20" as the formatter;
without it (original code) we get "1.79769313486232E+308".
Here is what worries me about this. If this format is only an in memory format
(it is not), than that's OK. However, if this is going to the index (it is),
than we have an issue, where the index will not be compatible. And it looks to
me that this may have been an issue since the early days of Lucene.Net.
Comments? Anyone else, please pitch in.
> Tests in Lucene.Net.Util.TestFieldCacheSanityChecker
> ----------------------------------------------------
>
> Key: LUCENENET-226
> URL: https://issues.apache.org/jira/browse/LUCENENET-226
> Project: Lucene.Net
> Issue Type: Bug
> Reporter: Michael Garski
> Attachments: TestFieldSanityChecker.patch
>
>
> There are 3 tests failing at this time:
> Lucene.Net.Util.TestFieldCacheSanityChecker.TestInsanity1
> Lucene.Net.Util.TestFieldCacheSanityChecker.TestInsanity2
> Lucene.Net.Util.TestFieldCacheSanityChecker.TestSanity
> I'll be working with those and updating this issue with progress or issues.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
