Save problem using proxy

When I access an article through my insitute's proxy service, Zotero is unable to save. The correct "Save to zotero" icon shows and uses the correct translator (I compared without the proxy), but the "Saving to.." popup hangs without doing anything, until I click anywhere - it then disappears. I have managed to save fine on this proxy in the past. Unfortunately the proxy is not a common one I can just give you the address of - its different each time, but uses the base https://ist.cea.fr/
As a workaround - can I disable the use of a translator, so I can at least save a snapshot to add to the item I saved without the proxy?
Debug ID: D1973746029
  • Can you submit a Debug ID from the Connector?
    As a workaround - can I disable the use of a translator, so I can at least save a snapshot to add to the item I saved without the proxy?
    You can save without a translator by right clicking the Zotero Connector button and choosing from the list under "Save to Zotero". The last two options don't use a translator and just save a snapshot.
  • Here's the Debug ID: D926058426

    > You can save without a translator by right clicking the Zotero Connector button and choosing from the list under "Save to Zotero". The last two options don't use a translator and just save a snapshot.
    Thanks, that worked.
  • edited September 6, 2023
    Saving via that proxy generally isn't going to work. Most of the time Zotero won't even be able to detect what site you're on in order to choose the right translator, since the original domain doesn't appear anywhere in the URL.

    The fact that you're getting the right translator here is just a function of the site using a publishing platform that's used across different domains, so Zotero detects it based on URL path rather than domain. But the translator still assumes that it's on a root domain without some random bit in the path, so here it's constructing a RIS download URL and getting https://ist.cea.fr/Citation/Download?resourceId=[…], which of course fails because it's missing the random proxy part.

    If the random part is actually consistent for each domain, you could (tediously) create separate proxy-server entries for each one, mapped to a single publisher site, and then Zotero could transform the URLs and know what site it's on, but if it's random each time, even that wouldn't be possible.

    Likely the best you can really do is rely on RIS/BibTeX download (the former sometimes labeled as "EndNote" or "Reference Manager") from sites, which in many cases the Zotero Connector can intercept and import directly. In cases where that doesn't work (because sites don't serve the files properly), you can save to disk and import manually via File → Import.
  • OK. Well, the workaround I used works - that is, create the item by saving from the non-proxied site, then adding in attachments by hand from the proxied site. Thanks again.
Sign In or Register to comment.