Regarding wrapping the IStream - send me an email
Wrapping COM interfaces:
http://www.amazon.com/exec/obidos/tg...glance&s=books
Imaging documentation (or the lack of thereof) - I agree completely
"plinth" <(E-Mail Removed)> wrote in message
news:88CD4C0A-341C-4AC1-8D36-(E-Mail Removed)...
> Thanks - here's what I found:
> 1. The crash was happening in Seek because the out long newPosition was
> being sent in as null (which means "I don't care about the new
> position") - I
> changed it to IntPtr and made the appropriate code changes.
> The same crash happens later now. I've got breakpoints set at the entry
> in
> every method in the interface implementation and none are getting hit.
> This
> leads me to believe that the problem might be in the definition of the
> interface itself.
> IStream in the header files is defined as inheriting from
> ISequentialStream.
> I made the assumption that the vtables would just be sequential so I just
> injected the member declarations into the IStream. I'm assuming that this
> is
> a mistake.
>
> Now, here's a vent for my frustration: is there anywhere where the process
> of wrapping of COM interfaces is documented properly? What I've stumbled
> upon is basically "add the COM object to the reference in your project and
> it
> will just work", which unfortunately is not an option in CF.
>
> Furthermore, I'd like to see the IImagingFactory, IImage, and other
> interfaces actually documented as to their intended use. The existing
> documentation is very much in the category of "what", but this section of
> code needs both a "why" and "how" level of documentation. I went through
> the
> sample code in the CE 5.0 SDK and found one example which only does the
> bare
> minimum, and thanks, I already figured that much out.
>
> It'd be nice to see something clear and coherent along the lines of,
> "here's
> how you write your own codec". "here's how you can expect your codec to
> be
> called by an IImageSink" "here's how you write your own sink" "Here's
> the
> overall architecture and relationship of the objects in this space and the
> intent"
>