CLDR-15990 Generate .xsd and .md from .dtd

[srl295]

CLDR-15990

• Updates GenerateDtd, which is normally run automatically as part of BRS, to do two things:

1. Generate a .xsd file for each .dtd with XML Schema for the DTD files. These are checked into the repository and intended to be released with CLDR.
2. Generate a .md file in docs/ldml/dtd2md/ with a shell that can be copied into the specification for each element. This directory is .gitignored and the resulting files are not added to the repository.

• Updates XMLValidator to add a --schema= option allowing it to be used to validate against a schema.

• Note that DTD @-annotations such as @MATCH are not supported by the converter at present, so such comments are removed from the .xsd

• CLDR-15990 generate .xsd from .dtd #2350 is the original PR

---

Example of the generated .md content:

### <a name="Element_languages" href="#Element_languages">Element: languages</a>

> <small>
>
> Parents: [localeDisplayNames](#Element_localeDisplayNames)
>
> Children: [alias](#Element_alias), [language](#Element_language), [_special_](tr35.md#special)
>
> Occurrence: ?, ?
> </small>

_Attribute:_ `draft`

_Attribute:_ `references`

_Attribute:_ `standard`

_Attribute:_ `validSubLocales`

* * *

• This PR completes the ticket.

ALLOW_MANY_COMMITS=true

[jira-pull-request-webhook]

Hooray! The files in the branch are the same across the force-push. 😃

~ Your Friendly Jira-GitHub PR Checker Bot

[pedberg-icu]

Out of curiosity, what does it mean to have the two successive @METADATA/@DEPRECATED annotations here?

[srl295]

These are now removed.

[pedberg-icu]

Still confused about how these annotations are being used. Is there a mapping from the sequence of annotations to items inside the previous block?

[srl295]

The annotations aren't used, I've now removed them.

[pedberg-icu]

It all looks reasonable, but I did not check the xsd files in detail, I assume they are generated. I have a a general question about what the annotations like @deprecated mean in these xsd files and how they are used, but that is not blocking.

[macchiati]

I looks like the xsd is tossing in all the annotations way after what they apply to. Also, the description for the PR needs much more information about what is being done and why.

[srl295]

This conversion doesn’t handle annotations yet. So it’s not suitable for replacing CLDR validation. However it is useful for clients checking data that’s already from CLDR. And I can keep iterating it. I’d need to mutate the xsd based on dtd data.

[srl295]

I looks like the xsd is tossing in all the annotations way after what they apply to.

They are just comments. It (the off the shelf dtd to xsd converter) is probably grouping them to the item they follow. Perhaps a step would be to just remove the annotation comments.

Also, the description for the PR needs much more information about what is being done and why.

Makes sense. Thanks.

[srl295]

ping… can i get a review on this? it now removes the annotation comments.