crc problems or fake files?

Hi, I have a strange problem, could you help me, gurus? ;)

I have two .zip archives. Both are protected with password and contain
one .doc file.
I've found passwords for both archives and they were decompressed with
WinRAR without any problems.
But decompressed files look like a garbage in MS Word viewer.

I think that files was just fake. But is it possible that I've found
wrong passwords which produces right CRC after decompression?

On Mar 28, 4:21=A0am, scriptg...@gmail.com wrote:
> But is it possible that I've found
> wrong passwords which produces right CRC after decompression?

No.  Especially since you have two instances.

Not only would the CRC have to match, so would the decryption check
that occurs before decompression even starts, and furthermore the
random data that would result from a faux password would have to both
have no deflate format errors (which is quite unlikely if the files
are longer than a few hundred bytes), and the self-terminating deflate
format would have to accidentally terminate at exactly the right
compressed byte and in order to produce no errors within the zip
format, would need to accidentally generate exactly the right number
of decompressed bytes.

I haven't calculated the probability of all of those accidents
occurring, but I'd guess somewhere around 10^-50 (+/- 20 in the
exponent).  Since you had it happen twice, around 10^-100.

So, no.  You have produced exactly the files that were originally
compressed.  Now you have to figure out what they are.

Mark

On Mar 28, 11:22 am, Mark Adler <mad...@alumni.caltech.edu> wrote:
> On Mar 28, 4:21 am, scriptg...@gmail.com wrote:
> 
>
> No.  Especially since you have two instances.
>
> Not only would the CRC have to match, so would the decryption check
> that occurs before decompression even starts, and furthermore the
> random data that would result from a faux password would have to both
> have no deflate format errors (which is quite unlikely if the files
> are longer than a few hundred bytes), and the self-terminating deflate
> format would have to accidentally terminate at exactly the right
> compressed byte and in order to produce no errors within the zip
> format, would need to accidentally generate exactly the right number
> of decompressed bytes.
>
> I haven't calculated the probability of all of those accidents
> occurring, but I'd guess somewhere around 10^-50 (+/- 20 in the
> exponent).  Since you had it happen twice, around 10^-100.
>
> So, no.  You have produced exactly the files that were originally
> compressed.  Now you have to figure out what they are.
>
> Mark

You look REALLY bored.

Thanks a lot, Mark!
I think you're right. I am sure that two files with same size can have
the same CRC (it is easy to prove), but i didn't thought about
possible errors in deflate format.
Somebody just compressing garbage files, compressing them with
password and trying to sell it. I don't think these files contain some
kind of structured information, I've analyzed the content of that
files.