[
https://issues.apache.org/jira/browse/XMLBEANS-22?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12896790#action_12896790
]
Martin Heitz commented on XMLBEANS-22:
--------------------------------------
Because there is no support here I patched locally the method next of the class
Saver.PrettySaveCur to
boolean next ( )
{
int k;
if (_txt != null)
{
assert _txt.length() > 0;
assert !_cur.isText();
_txt = null;
_isTextCData = false;
k = _cur.kind();
}
else
{
int prevKind = k = _cur.kind();
if (!_cur.next())
return false;
_sb.delete( 0, _sb.length() );
assert _txt == null;
// place any text encountered in the buffer
if (_cur.isText())
{
// _cur.isTextCData() is expensive do it only if
useCDataBookmarks option is enabled
_isTextCData = _useCDataBookmarks && _cur.isTextCData();
CharUtil.getString( _sb, _cur.getChars(), _cur._offSrc,
_cur._cchSrc );
_cur.next();
// vvvvv fix: do not trim unconditionally vvvvv
// trim( _sb );
k = _cur.kind();
if (prevKind == ELEM && k == -ELEM) {
// System.out.println("not trimming '" + _sb + "'
-- prevKind==" + prevKind+ ", k=" + k);
} else {
// System.out.println("trimming '" + _sb + "' --
prevKind==" + prevKind + ", k=" + k);
trim( _sb );
// System.out.println("... to '" + _sb + "'");
}
// ^^^^^ fix: do not trim unconditionally ^^^^^
}
k = _cur.kind();
// Check for non leaf, _prettyIndent < 0 means that the save is
all on one line
if (_prettyIndent >= 0 &&
prevKind != COMMENT && prevKind != PROCINST && (prevKind
!= ELEM || k != -ELEM))
// if (prevKind != COMMENT && prevKind != PROCINST && (prevKind
!= ELEM || k != -ELEM))
{
if (_sb.length() > 0)
{
_sb.insert( 0, _newLine );
spaces( _sb, _newLine.length(), _prettyOffset +
_prettyIndent * _depth );
}
if (k != -ROOT)
{
if (prevKind != ROOT)
_sb.append( _newLine );
int d = k < 0 ? _depth - 1 : _depth;
spaces( _sb, _sb.length(), _prettyOffset +
_prettyIndent * d );
}
}
if (_sb.length() > 0)
{
_txt = _sb.toString();
k = TEXT;
}
}
if (k == ELEM)
_depth++;
else if (k == -ELEM)
_depth--;
return true;
}
Good luck, Mattin
P.S. Of course no guarantees that this does not have any side effects. I'm
using this code now for more than a month and have not not notified any problem.
> xmlText trims whitespace within elements if prettyprint is on
> -------------------------------------------------------------
>
> Key: XMLBEANS-22
> URL: https://issues.apache.org/jira/browse/XMLBEANS-22
> Project: XMLBeans
> Issue Type: Bug
> Affects Versions: Version 1
> Environment: Java 1.4.2_03, Win 2K
> Reporter: Aaron Dunlop
> Fix For: TBD
>
> Attachments: whitespace-patch
>
>
> When called with XmlOptions.SAVE_PRETTY_PRINT, xmlText trims whitespace
> within an element - e.g., ' a ' is converted to 'a'.
> Unless I'm completely missing something, the formatter shouldn't ever change
> the contents of an element.
> The following test snippet reveals the problem:
> XmlObject xdoc = XmlObject.Factory.parse("<test> </test>");
> Assert.assertEquals("<test> </test>", xdoc.xmlText());
> Assert.assertEquals("<test> </test>" + newLine, xdoc.toString());
> The current build from CVS returns <test/>[newline] instead of <test>
> <test>[newline].
> Complete testcase and patch to follow.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@xmlbeans.apache.org
For additional commands, e-mail: dev-h...@xmlbeans.apache.org
