[jboss-jira] [JBoss JIRA] (AS7-6139) ModelTypeValidator is overly lenient about numeric types
[
https://issues.jboss.org/browse/AS7-6139?page=com.atlassian.jira.plugin.s...
]
Jason Greene commented on AS7-6139:
-----------------------------------
you can call getType() on any ModelNode, and then do if/else with instanceof.
The rules I am thinking of that make the most sense are:
long|bigint|*bigdec|*double -> int (check range)
bigint|*bigdec|*double -> long (check range)
int|long|bigint|*bigdec -> double (allow)
int -> long (allow)
*bigdec -> bigint (allow)
string -> int|long|bigint|bigdec|double (verify parse)
* = we could theoretically verify no fractional values when going from double|bigdec to an
integral value. I think that we really should not.
ModelTypeValidator is overly lenient about numeric types
--------------------------------------------------------
Key: AS7-6139
URL: https://issues.jboss.org/browse/AS7-6139
Project: Application Server 7
Issue Type: Bug
Components: Domain Management
Affects Versions: 7.1.3.Final (EAP)
Reporter: Brian Stansberry
Assignee: Chao Wang
Priority: Minor
Attachments: AS7-6139.patch
ModelTypeValidator does some conversion checks to see if a passed in ModelNode meets the
requirements of the valid type. These checks are overly lenient when it comes to numeric
types as they rely only on the ModelNode.asXXX() methods not failing. But those methods
don't fail even if the conversion involves a narrowing.
For example, calling new ModelNode(Long.MAX_LONG).asInt() will not fail, but if the user
passed in such a value for use in an attribute of type INT the stored value would not be
what was expected.
Note this is minor as in many, many cases ModelTypeValidator subclasses like
IntRangeValidator are used, and those subclasses enforce ranges.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira