robmaris <[EMAIL PROTECTED]> wrote:
> switch (type) {
> - // byte array types
> - case FieldDefs.TYPE_BYTE: //(keesj:) byte in mmbase stands for byte
> array
> - //I'm not sure a byte array is numerical
> - if (!(value instanceof byte[])) {
> - ok = false;
> - }
> - break;
> + case FieldDefs.TYPE_BYTE: //(keesj:) byte in mmbase stands for byte array
> + //I'm not shure a byte array is numerical
Now, the value of a byte-array field must be Number again. This seems very
odd. Why was this part of the change rolled back?
> - case FieldDefs.TYPE_NODE:
> - if (value instanceof MMObjectNode) { // core node as a value
> - value = new Long(((MMObjectNode)value).getNumber());
> - break;
> - }
> - if (value instanceof Node) { // bridge node as a value
> - value = new Long(((Node)value).getNumber());
> - break;
> - }
I do think it makes some sense that the value of a 'NODE' typed field must be
(or at least can be) a Node.
> // String types.
> - case FieldDefs.TYPE_XML:
> - // XXX TODO: This code does not take into account the use of DOM
> objects
Also the TODO was removed, but is the same issue I think.
I think the idea is that only database-types are taken into consideration,
so 'TYPE_NODE' and 'TYPE_XML' do not really exist? Should we add this kind
of code on a higher level? (bridge's query?).
Michiel
--
Michiel Meeuwissen |
Mediapark C101 Hilversum |
+31 (0)35 6772979 | I hate Internet
nl_NL eo_XX en_US |
mihxil' |
[] () |
