Hi,

I just release a new preview of my planned VeraPDF plugin (on
https://github.com/ZUGFeRD/ZUV/releases/tag/v0.4.1) and unfortunately in
the past months

a) the need for ph-schematron has been confirmed and
b) I could not solve the deployment issues.

It turned out I do need ph-schematron, which in turn references saxon,
for for performance issues: we're talking about 11 seconds (ZUGFeRD 1)
respectively 4 seconds run time per ZUGFeRD file with when checking with
XSLT2 (i.e. ph-schematron and saxon2) while it takes 11 *minutes* for a
check without, i.e. with java's XSLT1. And so far I could not even
produce a XSLT1 for the ZF2 (more precisely UN/CEFACT SCRDM CII 16B EN
16931) schematron.

Despite something like java -jar ZUV-0.4.1-SNAPSHOT.jar -f <any ZUGFeRD
PDF file like from
http://www.mustangproject.org/MustangGnuaccountingBeispielRE-20170509_505.pdf>
works, if one tries to use the (shaded) ZUV-0.4.1-SNAPSHOT.jar as
VeraPDF plugin there will be exceptions like

Feb 12, 2018 10:36:31 AM org.verapdf.processor.ProcessorImpl extractFeatures
WARNUNG: Exception caught when extracting features of item
java.lang.NoClassDefFoundError: org/junit/rules/ExternalResource

Does anybody have an idea how this could be solved?
I e.g. tried a shaded jar with relocations but that did not work because
saxon had to be be refered by it's original name which AFAIK is taken by
VeraPDF internal usage.

My only remaining straw would be turning the thing around and attempting
to embed VeraPDF in the ZUGFeRD validator instead of embedding the
ZUGFeRD validator in VeraPDF as a plugin.

thanks a lot and kind regards,
Jochen

Mustangproject.org
Chief ZUGFeRD amatuer
_______________________________________________
Users mailing list
Users@lists.verapdf.org
http://lists.verapdf.org/listinfo/users