[xmlsec] problems with dsig2 and dsig3 examples
aleksey at aleksey.com
Thu Aug 29 14:20:47 PDT 2002
Well, the xmlsec reported that it could not find a node with Id.
This is one of a tricky questions about XML and XPath that
the ID attribute means nothing w/o DTD or schema )
I spent about a day when I first run into this :)
Devin Heitmueller wrote:
>Thank goodness! I thought I was going insane.
>Since I had never seen the examples work, I didn't have any reference
>point. I spent the day tinkering with my environment (and even built a
>new workstation), trying to figure out why my setup didn't work. I had
>just broken out gdb and started to step through the code when I got your
>mail. I'm quite relieved to find out that it wasn't me.
>Added the three lines, and now the example appears to work.
>On Thu, 2002-08-29 at 16:53, Aleksey Sanin wrote:
>>Yes, you are absolutelly right. There is a problem with the example test
>>because 'id' attribute is useless w/o DTD or schema
>>(see http://www.aleksey.com/pipermail/xmlsec/2002/000057.html for long
>>And I forgot to change examples :( To fix this you simply need to add 3
>>to the test.xml and test.tmpl files as follows (<?xml...?> is already in
>><?xml version="1.0" encoding="UTF-8"?>
>><!DOCTYPE test [
>><!ATTLIST Info Id ID #IMPLIED>
>>Devin Heitmueller wrote:
>>>Did some more testing:
>>>1. Installed a vanilla install of Redhat 7.2
>>>2. Compiled libxml2-2.4.24 and libxslt-1.0.20. Installed into /usr/lib
>>>3. Compiled xmlsec 0.0.9. Ran 'make check'. Only one test failed in
>>>the digital signature section. The same test failed on my Mandrake
>>>system. [is this normal?]
>>>4. ran 'make' in the docs/examples/dsig2 directory.
>>>5. ran './dsig2 dsakey.pem test.tmpl', and got the same errors as seen
>>>on the Mandrake system.
>>>So, at least I have narrowed it down to not being specific to Mandrake.
>>>It should be noted that I'm still using OpenSSL 0.9.6 (which ships
>>>standard for both Mandrake and RedHat). I guess I could try upgrading
>>>to 0.9.7, but the problem doesn't look like it's related to crypto at
>>>all, but rather the XML processing.
More information about the xmlsec