Ever since ec63800a414a72919dd96f874e8e9d85fd528afc we have non-sensical validation errors on CI .
listOrg.xml:216: element org: Relax-NG validity error : Invalid attribute id for element org
[9](https://github.com/readchina/ReadAct/runs/6424321663?check_suite_focus=true#step:10:9)
listOrg.xml:7: element org: Relax-NG validity error : Did not expect element org there
[10](https://github.com/readchina/ReadAct/runs/6424321663?check_suite_focus=true#step:10:10)
listOrg.xml:7: element org: Relax-NG validity error : Element listOrg has extra content: org
I can't reproduce these validation errors locally, the files in question validate using
xmllint --version
xmllint: using libxml version 20904
compiled with: Threads Tree Output Push Reader Patterns Writer SAXv1 FTP HTTP DTDValid HTML Legacy C14N Catalog XPath XPointer XInclude ICU ISO8859X Unicode Regexps Automata Expr Schemas Schematron Modules Debug Zlib
jing, saxon, and whatever else i can throw at it.
manually inspecting the files i see no reason why the 2.0 output should be invalid. Obviously listOrg can have direct org children, their xml:id are unique and valid NCNames, and the org at L7 has not been changed between v1.0 and v2.0.
This needs some further investigation, and experimenting. We might have a corrupt cache, an actual error in the new file hidden behind false log output, or we might need to switch validation method.
Ever since ec63800a414a72919dd96f874e8e9d85fd528afc we have non-sensical validation errors on CI .
I can't reproduce these validation errors locally, the files in question validate using
jing
,saxon
, and whatever else i can throw at it.manually inspecting the files i see no reason why the 2.0 output should be invalid. Obviously
listOrg
can have directorg
children, theirxml:id
are unique and validNCNames
, and the org at L7 has not been changed between v1.0 and v2.0.I've also inspected
4.3.0
and4.4.0
versions of the TEI schema files, and the rules in question are all as they should be.CI is using
libxml version 20910
so there seems to be a regression, leading to the false validation errors.Sadly this remains a showstopper for a new release.
this issue seems to experience similar problems https://gitlab.gnome.org/GNOME/libxml2/-/issues/223
This needs some further investigation, and experimenting. We might have a corrupt cache, an actual error in the new file hidden behind false log output, or we might need to switch validation method.