journals.cambridge.org attaches HTML not PDF
The translator for journals.cambridge.org erroneously attaches a HTML page, not the PDF file. I guess the journals.cambridge.org site tries to load the PDF in a separate page, but if that separate page is fetchable for Zotero the PDF should be too.
What I'm puzzled by is that I thought Zotero was supposed to check if a pdf is actually a pdf before attaching it?
It should start saving PDFs (and working on search results and tables of contents) again. If this works for you, please post here so that I can submit this change to be pushed to all users.