Importing a XAR with an attachment missing a <content> tag produces various issues

I tried to import a XAR XML file with an attachment lacking a <content> tag.

Like this (see also the attached minimal XAR archive to reproduce the issue):

  <attachment>
    <filename>hi</filename>
    <mimetype>text/plain</mimetype>
    <charset>UTF-8</charset>
    <author>XWiki.Admin</author>
    <date>1666270757000</date>
    <version>1.1</version>
    <comment/>
    <filesize>5</filesize>
  </attachment>

Expected behavior:

XWiki should handle the missing tag. It should produce an empty attachment, or ignore the attachment, or reject the page, or anything that does not lead to a broken state. A warning could be displayed.

Actual behavior:

The attachment appears in the Attachments page, but cannot be downloaded (XWiki says the attachment does not exist. In a CI system that upgrades XAR extensions automatically, the situation cannot be fixed by adding the missing <content> tag back (this does not happen when importing content from a XAR archive)

Errors can be seen in the logs:

{{2022-10-20 15:04:02,300 [XWiki Solr index thread] ERROR ttachmentSolrMetadataExtractor - Failed to retrieve the content of attachment [Attachment xwiki:Main.EmptyAttachment@hi]
com.xpn.xwiki.XWikiException: Error number 3002 in 3: The attachment [Attachment xwiki:Main.EmptyAttachment@hi] (file /home/raph/.xtool/instances/tmp/data/store/file/xwiki/f/4/bd82bfc89241883b9e066ec550727c/attachments/4/9/f68a5c8493ec2c0bf489821c21fc3b/f) could not be found in the filesystem attachment store.
at org.xwiki.store.legacy.store.internal.FilesystemAttachmentStore.loadAttachmentContent(FilesystemAttachmentStore.java:214)
at com.xpn.xwiki.doc.XWikiAttachment.loadAttachmentContent(XWikiAttachment.java:1107)
at com.xpn.xwiki.doc.XWikiAttachment.reloadAttachmentContent(XWikiAttachment.java:931)
at com.xpn.xwiki.doc.XWikiAttachment.getContentInputStream(XWikiAttachment.java:909)
at org.xwiki.search.solr.internal.metadata.AbstractSolrMetadataExtractor.getContentAsText(AbstractSolrMetadataExtractor.java:524)
at org.xwiki.search.solr.internal.metadata.AttachmentSolrMetadataExtractor.setLocaleAndContentFields(AttachmentSolrMetadataExtractor.java:114)
at org.xwiki.search.solr.internal.metadata.AttachmentSolrMetadataExtractor.setFieldsInternal(AttachmentSolrMetadataExtractor.java:93)
at org.xwiki.search.solr.internal.metadata.AbstractSolrMetadataExtractor.getSolrDocument(AbstractSolrMetadataExtractor.java:166)
at org.xwiki.search.solr.internal.DefaultSolrIndexer.getSolrDocument(DefaultSolrIndexer.java:510)
at org.xwiki.search.solr.internal.DefaultSolrIndexer.processBatch(DefaultSolrIndexer.java:419)
at org.xwiki.search.solr.internal.DefaultSolrIndexer.run(DefaultSolrIndexer.java:384)
at java.base/java.lang.Thread.run(Thread.java:833)
2022-10-20 15:04:02,303 [XWiki Solr index thread] ERROR .DocumentSolrMetadataExtractor - Failed to retrieve the content of attachment [Attachment xwiki:Main.EmptyAttachment@hi]}}