Tom Hindle wrote:
When checking param index type (which is 1 as SizeParamIndex == 1)
(about line 6778) the Array itself is being checked (byte[] pv) as
m->sig has had and extra parameter inserted at the beginning of its
parameter list. (which I assume is for communicating the HRESULT return
value back to the C++ as this is a COM method call.)
It's the converted "this" arg of the managed sig at the front
of the unmanaged sig.
Could someone tell me if this is the correct place to adjust the
param_num or should
spec->data.array_data.param_num be expected to contain the adjusted
value?
There are apparently 2 places in cominterop.c where
spec->data.array_data.param_num has to be increased by one.
The attached patch is fixing one place, but there seems
to be another one in cominterop_get_native_wrapper_adjusted.
Please try it out as I don't have a test case.
Robert
Index: cominterop.c
===================================================================
--- cominterop.c (revision 135665)
+++ cominterop.c (working copy)
@@ -1917,6 +1917,12 @@
mspecs[mspec_index] = g_new0
(MonoMarshalSpec, 1);
mspecs[mspec_index]->native =
MONO_NATIVE_VARIANTBOOL;
}
+ } else {
+ /* increase SizeParamIndex because
we've added a param */
+ if
(sig_adjusted->params[param_index]->type == MONO_TYPE_ARRAY ||
+
sig_adjusted->params[param_index]->type == MONO_TYPE_SZARRAY)
+ if
(mspecs[mspec_index]->data.array_data.param_num != -1)
+
mspecs[mspec_index]->data.array_data.param_num++;
}
}
_______________________________________________
Mono-devel-list mailing list
[email protected]
http://lists.ximian.com/mailman/listinfo/mono-devel-list
