Description of Problem: I'm attempting to use Norm Walsh's diffmk script on Red Hat Linux 8.0. This script takes two XML files, calculates the differences between them, and outputs a third file with some special markup of the changes. Running the diffmk script on Red Hat Linux 8.0 fails if I'm using — from DocBook. Version-Release number of selected component (if applicable): perl-XML-Parser-2.31-12 How Reproducible: Always Steps to Reproduce: 1. Download and install Algorithm::Diff from http://www.cpan.org/modules/by-category/06_Data_Type_Utilities/Algorithm/ 2.Grab my slightly modified diffmk tarball from http://www.ibiblio.org/gleblanc/diffmk.tar.bz2 (it just has some slightly modified test?.xml files to work with the DTDs normally found on RHL 8.0 and with — added) 3. change to diffmk/work and run '../diffmk --doctype docbook test1.xml test2.xml changes.xml' Actual Results: An error message stating: undefined entity at line 5, column 13, byte 287 at /usr/lib/perl5/vendor_perl/5.8.0/i386-linux-thread-multi/XML/Parser.pm line 185 Expected Results: A changes.xml file. Additional Information: I can't reproduce this bug on Red Hat Linux 7.2, which has a slightly older perl-XML-Parser. I tried to rebuild the older version on Red Hat Linux 8.0, but didn't have any luck.
This happens on other entities defined by the DTD, such as #.
Hmm, I tried recompiling the perl-XML-Parser 2.30. I ended up using the 2.31-12 spec file, with 'make test' removed, and the versions changed in order to make it output an RPM. Unfortunately, this doesn't appear to have fixed the problem, in fact, it doesn't even appear to change the error message.
Closing due to lack of activity and likely fixed in modern perl. If this is still an issue in RHEL4, FC3 or FC4 please open a new bug with new test information and exact versions of packages.
The only lack of activity that I see here is that nobody from Red Hat so much as acknowledged that they'd even read this. It comes with full steps to reproduce the bug, links to all relevant bits and pieces, what more could you want? Closing this as "CURRENTRELEASE" is just plain irresponsible.