[
https://issues.apache.org/jira/browse/DIRSERVER-1296?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Lorenz Breu updated DIRSERVER-1296:
-----------------------------------
Attachment: patch_SearchIT.txt
hope this worked...
should be a diff between my working copy and the repo version of SearchIT.java
in core-integ.
I added an AT "integerAttribute" with "integerOrderingMatch", added the OC
"extensibleObject" to the test oranizationalUnits and gave them new
integerAttributes matching their number. In addition I added 3 new such test
OUs... The old tests still pass.
Now the integerComaprisonTests test for the right issue. They fail without the
"fix" in ComparableComparator and pass if it is there, so my initial fix does
work...
However, I do realize it is a hack. The problem is, that the
"integerOrderingMatch" MatchingRule is assigned a ComparableComparator in the
ComparatorRegistry. I do not know, where this happens, but if somebody does,
then it would be much nicer to change that to some NumericalComparator there,
than trying to parse the strings as I have done.
> integer attribute types are not compared correctly
> --------------------------------------------------
>
> Key: DIRSERVER-1296
> URL: https://issues.apache.org/jira/browse/DIRSERVER-1296
> Project: Directory ApacheDS
> Issue Type: Bug
> Components: core
> Affects Versions: 1.5.4
> Reporter: Lorenz Breu
> Fix For: 1.5.5
>
> Attachments: ComparableComparator.java, patch_SearchIT.txt,
> SearchIT.java
>
>
> When searching for entries that have attributes with the INTEGER syntax, the
> values are compared lexicographically, not numerically. This happens even if
> the ordering and equality types are explicitly set to their integer versions
> when injecting the attribute types into ADS.
> Example:
> dn: cn = foo, dc = example
> cn: foo
> integerAttribute: 435
> now a search using "(integerAttribute<=500)" will correctly return the
> entry....
> but a search using "(integerAttribute<=44)" will ALSO return the entry, which
> it clearly should not.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
