[ http://issues.apache.org/jira/browse/IBATISNET-110?page=all ]
Gilles Bayon reassigned IBATISNET-110:
--------------------------------------
Assign To: Gilles Bayon
> Whitespace is not maintained properly when CDATA tags are used
> --------------------------------------------------------------
>
> Key: IBATISNET-110
> URL: http://issues.apache.org/jira/browse/IBATISNET-110
> Project: iBatis for .NET
> Type: Bug
> Reporter: Ron Grabowski
> Assignee: Gilles Bayon
> Fix For: DataMapper 1.3
>
> The ParseDynamicTags method in DomSqlMapBuilder.cs calls Trim() when building
> the sql statement. The following statement:
> <statement id="GetMany"><![CDATA[SELECT * FROM User ]]></statement>
> <statement id="GetOne" extends="GetMany"> WHERE Id=3</statement</statement>
> is incorrectly constructed as:
> SELECT * FROM UserWHERE Id=3
> when it should be:
> SELECT * FROM User WHERE Id=3
> The same error occurs when two or more CDATA nodes and/or a Text node is used
> inside the same statement. This statement:
> <statement id="GetMany">
> <![CDATA[SELECT * ]]>
> <![CDATA[FROM ]]>
> User
> </statement>
> is incorrectly rendered as:
> SELECT *FROMUser
> This is the offending code:
> string data = child.InnerText.Replace('\n', ' ').Replace('\r', '
> ').Replace('\t', ' ').Trim(); //??
> One solution is to trim the beginning of the first node and trim the end of
> the last node:
> string data = child.InnerText.Replace('\n', ' ').Replace('\r', '
> ').Replace('\t', ' ');
> if (i == 0)
> {
> // trim whitespace from the beginning of the string
> data = Regex.Replace(data, @"^\s*(.*)$", "$1");
> }
> else if (i == children.Count - 1)
> {
> // trim whitespace from the end of the string
> data = Regex.Replace(data, @"^(.*?)\s*$", "$1");
> }
> Yes, I know there is a TrimStart and TrimEnd method but I think there are
> more whitespace characters than just " ", "\n", "\t", "\r"...especially if
> the xml file is encoded with a funky content-type. There may be a more
> effecient regular expression to remove whitespace.
> Another solution is to remove the call to Trim after Replace.
> I think that if the user is using a CDATA node, they probably don't want
> their sql messed with. Perhaps if we encounter a CDATA node we shouldn't try
> to make it all fit on one line:
> trimCdataNodes="false" ???
> Thank you to Chad Humphries for encountering this bug.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see:
http://www.atlassian.com/software/jira
