Re: Skipping parts of a serialized file

From:

"Tom Serface" <tom@nospam.camaswood.com>

Newsgroups:

microsoft.public.vc.mfc

Date:

Thu, 12 Mar 2009 21:04:16 -0700

Message-ID:

<E4BCE20A-0B67-4664-872C-6ABF2249B75E@microsoft.com>

If you can recognize the beginning of the useful data you could just skip
past most of it then read the rest until something you recognize, like not
one of those funny chars, shows up then start reading there. Sans any
useful information that's probably the best you could do.

Tom

"Scoots" <linkingfire@msn.com> wrote in message
news:5fdecb80-5f33-4ab5-9ff9-b6b8212ce8c2@v13g2000pro.googlegroups.com...

I'm sure this problem has been encountered before, but I haven't been
able to turn up anything in searches. It's fairly simple. I'm
decoding a file produced by another application (entirely legal
circumstances, the machine and application is provided for us, as well
as the source code.). The file produced has a minimum of about 3kb's
of runtime information from serializing a CDocument. However, they
don't serialize any of the information in that Document, it is written
out in a stable file format. In fact, I don't even know why they use
the serializing in the first place, as this is an "exported" file
format, and the only serializing is the CDocument itself. But they
do.

So I wind up with a bunch of information at the top of the file that
my application (which reads the files) completely doesn't care about.
Reading the rest of the information is really easy, and I've already
taken care of it. However, my trouble comes when it seems that this
initial runtime class information is NOT a fixed size. I haven't been
able to find any information at, say the top of the file, that
contains the size of this runtime information.

Is this size contained anywhere (if so, I haven't found it yet), or is
there an easy way to skip this runtime information so I can get to
what I need?

Thanks,
~Scoots