You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@santuario.apache.org by Berin Lautenbach <be...@wingsofhermes.org> on 2004/07/01 13:49:57 UTC
Re: With comments
I've just run these two sample files against the CVS Java library. The
"with" file fails to verify, but the "without" is fine.
Dumping the post transformation bytes, I get the same output as the C++
library - in particular the PI stuff is output. With the Java 1.1
library, I don't get the PI output, only the nodes after it.
According to :
http://www.w3.org/TR/xml-c14n#Example-OutsideDoc
My read is that the PI *should* be there, so the C++ library and current
Java CVS behaviour is correct. Java library 1.1 and previous is
incorrect (again my read). This agrees with Raul's position, but I'm
happy to be contradicted if someone thinks I have it up wrong!
Cheers,
Berin
Milan Tomic wrote:
> Hi Raul,
>
>
>>The code in CVS of c14n in java and c++ looks good enouth. So
>>I need to know what version of the Java library are you
>>using
>
> I've checked with both 1.0 and 1.1 version of XSEC (checksig.exe app)
> and verification of signature fails. :<
>
> Thank you,
> Milan
>
>
>
>
RE: With comments
Posted by Milan Tomic <mi...@setcce.org>.
Thank you, guys. We took new version from CVS and it works OK.
Version 1.1 of Java libraries was not signing PIs. When I
removed PIs from Java XSEC 1.1 signed file and verified signature using
C++ libraries, signature was valid.
Best regards,
Milan
Re: With comments
Posted by ra...@r-bg.com.
> I've just run these two sample files against the CVS Java library. The
> "with" file fails to verify, but the "without" is fine.
>
> Dumping the post transformation bytes, I get the same output as the C++
> library - in particular the PI stuff is output. With the Java 1.1
> library, I don't get the PI output, only the nodes after it.
>
> According to :
>
> http://www.w3.org/TR/xml-c14n#Example-OutsideDoc
>
> My read is that the PI *should* be there, so the C++ library and current
> Java CVS behaviour is correct. Java library 1.1 and previous is
> incorrect (again my read). This agrees with Raul's position, but I'm
> happy to be contradicted if someone thinks I have it up wrong!
>
> Cheers,
> Berin
>
You are *complytly* right. I was getting mad because I was only testing in
CVS branch and it was correct.
Perhaps the issue is fixed in the patch where I rewrite the c14nsubtree
making it common between excl & incl.
Regards,