Restrict where `<altIdent>` can go in tei_customization and p5odds

[sydb]

In response to Council’s decision of 2023-05-09 re: #2285, restrict where <altIdent> is allowed to only <elementSpec>, <attDef>, and <valItem> in p5odds and tei_customization:

p5odds Just added a <constraintSpec>

tei_customization Moved <altIdent> out of model.identSynonyms and then add it back into content models of <attDef> and <valItem>. While I was there also:

• Removed the <constraintSpec> that limited <altIdent> to an NCName (TEI P5 now does this)
• Updated the ODD file itself to be based on 4.6.0
• Increased the version number to 0.9.0b — I am guessing we forgot to increase it to 1.0.0 on actual release. Sigh.

I have run make test in the P5 directory, and all tests passed. What I have not done is actually test that p5odds and tei_customization do the right thing.

[HelenaSabel]

Hello @sydb!

The schema tei_odds.rng (generated with make exemplars) is not firing errors in a test file that contains <altIdent> as a child of <attDef> and as a child of <constraintSpec>.

In the case of tei_customization.rng, I tested a few ODDs and I always get the error conflicting ID-types for attribute "defaultPhase" of element "schema" from namespace "http://purl.oclc.org/dsdl/schematron" but this also happens with the version in dev.

[sydb]

Heh-heh. The test for p5odds is not in the RELAX NG, it is in the Schematron. However, I just tested and found a bug: I used ../tei:attDef where I should have said parent::tei:attDef. I plan to check in a fix in a few minutes.

In the meantime, the error you are getting from tei_customization is the result of having DTD compatibility mode turned on. If you are using jing on the commandline add a -i switch; if you are using oXygen uncheck the “Check ID/IDREF” box in the Options > Preferences > XML > XML Parser > RELAX NG pane. See https://journals.openedition.org/jtei/2573#annexe.

[sydb]

Fixed in 251161c.