Hi folks,
this is the fourth time I'm sending this patch to this list: the first post
got ignored, the second time it got accepted (applied by Miguel in r36917 on
2004-12-01), then accidently reverted by Konstantin Triger in r44547 on
2005-05-16 (he managed to replace the whole file with one that was partially
out-of-date).
I reposted the adapted patch a second time about a month ago as I noted that
it isn't included in HEAD, which spawned concerns by Konstantin that the
patch doesn't address a deeper design flaw, which I disagreed with. The
complete code assumes rowCache is always an array instance and never null,
and my patch assures it really is never null. But nothing happened, no reply,
no applied patch.
So I'm sending this primitive patch again, since a project our company works
on relies on the correct behaviour of DataView. Please, could someone
re-apply it ? :-) Or at least this time tell me what's wrong with that patch
and what design issue to address to fix it once and for all.
Thnx a lot,
Marc Haisenko
--
Marc Haisenko
Systemspezialist
Webport IT-Services GmbH
mailto: [EMAIL PROTECTED]
Index: DataView.cs
===================================================================
--- DataView.cs (revision 47264)
+++ DataView.cs (working copy)
@@ -41,7 +41,7 @@
IExpression rowFilterExpr;
string sort = String.Empty;
protected DataViewRowState rowState;
- protected DataRowView[] rowCache = null;
+ protected DataRowView[] rowCache = new DataRowView[0];
// BeginInit() support
bool isInitPhase = false;
@@ -535,7 +535,7 @@
if (dataTable != null)
UnregisterEventHandlers ();
Index = null;
- rowCache = null;
+ rowCache = new DataRowView[0];
isOpen = false;
}
@@ -703,7 +703,7 @@
{
// TODO: what really happens?
Close ();
- rowCache = null;
+ rowCache = new DataRowView[0];
Open ();
OnListChanged (new ListChangedEventArgs (ListChangedType.Reset, -1 ));
}
_______________________________________________
Mono-devel-list mailing list
[email protected]
http://lists.ximian.com/mailman/listinfo/mono-devel-list
