XMLSerializer and NameTable Entries

microsoft.public.dotnet.framework

6 Aug 2006 3:35 PM

Fratt

I'm using the xmlSerializer that serializes/deserializes an xml document. I
ran the xsd.exe codegen tool against the schema. It appears that when I go
to deserialze my type for the first time that the xmlSerializer loads ALL
nested element types into its nametable, regardless of whether or not that
type is being used in the message. There are types that are not used and,
given the number of types I have, it's become a memory hog. Is there any way
to change the behavior of the xmlSerializer so that it doesn't preload every
possible type? I tried passing an xmlReader with my own NameTable, but I'm
not so sure that changed anything. I didn't preload it or anything. I just
passed my own in to share amongs many instances of xmlSerializer. Any
thoughts? I haven't found any good explanation of what the xmlSerializer is
doing behind the seens. Thanks.