Error in toc extent calculation

[mojavelinux]

If a TOC entry does not have an ID, that entry an all of its descendants will be skipped when computing the extent of the TOC. However, by the time the TOC is inked in the main document, this entry will have a pdf-anchor attribute, which will cause it to be included in the TOC. As a result, the TOC extent calculation will be incorrect and it will overrun the first page of content in the body.

This is a regression introduced by f8fa537.

There is still a larger issue here, which is that entries that were not included in the extent calculation end up being included when the TOC is added to the primary document. This issue only deals with the fact that the children were excluded in the TOC extent calculation.

[mojavelinux]

This issue can only be observed when the sectids attribute is turned off for parts of a document, which is not a typical usage of AsciiDoc.