Package: python3-lxml Version: 4.6.3-1 Severity: important Dear Maintainer,
The Relax NG validator does not correctly detect the errors in the XML file. The simple set of files reconstructing the problem is attached in archive "bug_demo.tar.gz". The agwb.rng defines the XML schema. The example1.xml contains the conforming xml file. The example1a.xml file has an error - field "B2" in creg "X2" does not contain required "width" attribute. The example1b.xml has another error - the attribute "name" in block SYS1 is written as "nafme". Validating the erroneous files, however, gives incorrect results. The tests are performed both with Python lxml module and with xmllint utility. The errors are similar, so probably the error is located in the lxml library. $ ./test0.sh example1a.xml [...] example1a.xml:11: element field: Relax-NG validity error : Element field failed to validate attributes example1a.xml:10: element field: Relax-NG validity error : Element creg has extra content: field Relax-NG validity error : Extra element creg in interleave example1a.xml:9: element creg: Relax-NG validity error : Element block failed to validate content Relax-NG validity error : Extra element block in interleave example1a.xml:3: element block: Relax-NG validity error : Element sysdef failed to validate content example1a.xml fails to validate $ ./test0.py example1a.xml False example1a.xml:11:0:ERROR:RELAXNGV:RELAXNG_ERR_ATTRVALID: Element field failed to validate attributes example1a.xml:10:0:ERROR:RELAXNGV:RELAXNG_ERR_EXTRACONTENT: Element creg has extra content: field <string>:0:0:ERROR:RELAXNGV:RELAXNG_ERR_INTEREXTRA: Extra element creg in interleave example1a.xml:9:0:ERROR:RELAXNGV:RELAXNG_ERR_CONTENTVALID: Element block failed to validate content <string>:0:0:ERROR:RELAXNGV:RELAXNG_ERR_INTEREXTRA: Extra element block in interleave example1a.xml:3:0:ERROR:RELAXNGV:RELAXNG_ERR_CONTENTVALID: Element sysdef failed to validate content The error message lacks specifity. There should be a detailed information about the lacking field. In case of example1b.xml the situation is even worse: $ ./test0.sh example1b.xml example1b.xml:3: element block: Relax-NG validity error : Expecting an element sreg, got nothing Relax-NG validity error : Extra element block in interleave example1b.xml:3: element block: Relax-NG validity error : Element sysdef failed to validate content example1b.xml fails to validate ./test0.py example1b.xml False example1b.xml:3:0:ERROR:RELAXNGV:RELAXNG_ERR_NOELEM: Expecting an element sreg, got nothing <string>:0:0:ERROR:RELAXNGV:RELAXNG_ERR_INTEREXTRA: Extra element block in interleave example1b.xml:3:0:ERROR:RELAXNGV:RELAXNG_ERR_CONTENTVALID: Element sysdef failed to validate content There is absolutely no reason to expect "sreg" element in a "sysdef". It is against the defined schema. However, if I convert the schema to the DTD: $ trang -I rng -O dtd agwb.rng agwb.dtd then the further validations wit DTD give the correct results for both erroneous files: $ ./test1.sh example1a.xml [...] example1a.xml:11: element field: validity error : Element field does not carry attribute width Document example1a.xml does not validate against agwb.dtd $ ./test1.py example1a.xml False example1a.xml:11:0:ERROR:VALID:DTD_MISSING_ATTRIBUTE: Element field does not carry attribute width $ ./test1.sh example1b.xml [...] example1b.xml:3: element block: validity error : Element block does not carry attribute name example1b.xml:3: element block: validity error : No declaration for attribute nafme of element block Document example1b.xml does not validate against agwb.dtd $ ./test1.py example1b.xml False example1b.xml:3:0:ERROR:VALID:DTD_MISSING_ATTRIBUTE: Element block does not carry attribute name example1b.xml:3:0:ERROR:VALID:DTD_UNKNOWN_ATTRIBUTE: No declaration for attribute nafme of element block The above reports precisely describe errors in the validated files. Therefore the problem is rather not in the Relax NG scheme definition but in the validator itself. -- System Information: Debian Release: bullseye/sid APT prefers testing APT policy: (500, 'testing'), (500, 'stable') Architecture: amd64 (x86_64) Kernel: Linux 5.10.0-5-amd64 (SMP w/12 CPU threads) Locale: LANG=pl_PL.UTF-8, LC_CTYPE=pl_PL.UTF-8 (charmap=UTF-8), LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash Init: systemd (via /run/systemd/system) LSM: AppArmor: enabled Versions of packages python3-lxml depends on: ii libc6 2.31-10 ii libxml2 2.9.10+dfsg-6.3+b1 ii libxslt1.1 1.1.34-4 ii python3 3.9.2-2 Versions of packages python3-lxml recommends: ii python3-bs4 4.9.3-1 ii python3-html5lib 1.1-3 Versions of packages python3-lxml suggests: pn python-lxml-doc <none> pn python3-lxml-dbg <none> -- no debconf information
