Text after an empty nested blockquote is dropped (18.0.9+)

Marked version: 18.0.9 through 18.0.13 (18.0.8 is fine)

Describe the bug
Text on the line after an empty nested blockquote is dropped from the output, with default options.

To Reproduce

1. Marked Demo
2. CommonMark Demo

marked.parse('>>\ntext') in 18.0.8, matching the dingus:

<blockquote>
<blockquote>
</blockquote>
</blockquote>
<p>text</p>

In 18.0.9 to 18.0.13:

<blockquote>
<blockquote>
</blockquote>
</blockquote>

Same for >>>\ntext and > quote\n>>\ntext. With >>\ntext\n> more, the following > more blockquote disappears as well.

Expected behavior
The empty nested blockquote closes and text becomes a paragraph after it, as in CommonMark and 18.0.8.

This looks like it comes from #4030: the continuation branch in Tokenizer.blockquote() now appends every continuation line to raw, also when the re-tokenized nested blockquote did not consume them, so the lexer skips past text no token holds. Reverting that commit brings the paragraph back. I have a small fix with a spec test in test/specs/new/ and would be glad to open a PR if that is welcome.

Start by reading Tokenizer.blockquote() and the change referenced in #4030, then reproduce the issue with the nested blockquote examples in the report. Add or inspect the spec test in test/specs/new/ and verify that text after an empty nested blockquote is preserved while the existing blockquote cases still pass.