skip premature validation of required drop downs in metadata blocks

[jo-pol]

…mandatory

What this PR does / why we need it:

It skips a premature validation phase for a metadata block field with a mandatory drop down. This allows proper template change for a blank dataset and gives clearer error messages when saving.

Which issue(s) this PR closes:

Closes #10119

Special notes for your reviewer:

The HTML attribute required triggers a premature validation in metadata blocks. The db-table datasetfieldtype controls the desired validation phase.

Suggestions on how to test this:

• create a dataverse: https://demo.dataverse.org/ - add data - add dataverse
• select the dataverse - edit - general information
• make a field in a metadatablock required (e.g. type of article in the journal block)
• create a custom template and set a field(s) in the citation block (e.g. title).
• Create a new dataset, set none of the fields and change to the custom template, the title should be applied. Save the blank dataset and a clear error message appears for the type of article.

Does this PR introduce a user interface change? If mockups are available, please link/include them here:

no

Is there a release notes update needed for this change?:

no

Additional documentation:

[qqmyers]

Same change as in #10125

[scolapasta]

@jo-pol @qqmyers maybe this is a dumb question, but doesn't the fact that required checks to see if DatasetPage means that required would be false for templates? (at least that's how it reads it's supposed to work)

[qqmyers]

It's been a while, but I think this is a dataset page issue, when using a template.

[jo-pol]

@qqmyers @janvanmansum @scolapasta

I broke down the "suggestion to test" in smaller steps and therefore went through the procedure again as far as I remembered. Then I found more precise scenarios in #10119 and a relation with #2190

Though the PR is still open, the demo system now shows desired error indication:

[qqmyers]

Per a note from the author in the issue, it looks like this change has already been added in #10401. So this PR and the issue can be closed.