Aparently, we're not supposed to callback to the parent
for lParam when in LVS_OWNERDATA. Very strange, but it
kindda makes some sense. Anyhow, it fixes Ender's IDA
app.
ChangeLog
Do not callback to the parent for lParam when in LVS_OWNERDATA.
Docs update.
--- dlls/comctl32/listview.c.Wm Tue Oct 22 19:09:46 2002
+++ dlls/comctl32/listview.c Tue Oct 22 21:05:45 2002
@@ -59,7 +59,12 @@
* -- we should keep an ordered array of coordinates in iconic mode
* this would allow to frame items (iterator_frameditems),
* and find nearest item (LVFI_NEARESTXY) a lot more efficiently
- *
+ *
+ * Flags
+ * -- LVIF_COLUMNS
+ * -- LVIF_GROUPID
+ * -- LVIF_NORECOMPUTE
+ *
* States
* -- LVIS_ACTIVATING (not currently supported by comctl32.dll version 6.0)
* -- LVIS_CUT
@@ -4757,12 +4762,12 @@
{
dispInfo.item.state = 0;
- /* if we need to callback, do it now */
- if ((lpLVItem->mask & ~LVIF_STATE) || infoPtr->uCallbackMask)
+ /* apprently, we should not callback for lParam in LVS_OWNERDATA */
+ if ((lpLVItem->mask & ~(LVIF_STATE | LVIF_PARAM)) || infoPtr->uCallbackMask)
{
/* NOTE: copy only fields which we _know_ are initialized, some apps
* depend on the uninitialized fields being 0 */
- dispInfo.item.mask = lpLVItem->mask;
+ dispInfo.item.mask = lpLVItem->mask & ~LVIF_PARAM;
dispInfo.item.iItem = lpLVItem->iItem;
dispInfo.item.iSubItem = lpLVItem->iSubItem;
if (lpLVItem->mask & LVIF_TEXT)
@@ -4777,7 +4782,10 @@
*lpLVItem = dispInfo.item;
TRACE(" getdispinfo(1):lpLVItem=%s\n", debuglvitem_t(lpLVItem, isW));
}
-
+
+ /* make sure lParam is zeroed out */
+ if (lpLVItem->mask & LVIF_PARAM) lpLVItem->lParam = 0;
+
/* we store only a little state, so if we're not asked, we're done */
if (!(lpLVItem->mask & LVIF_STATE) || lpLVItem->iSubItem) return TRUE;
--
Dimi.
