Connector PDF retrieval breakdown (Firefox)

edited February 22, 2018
Hello,

For me, Zotero 5.0.35.1 is not retreieving any PDFs using the Firefox connector. I have tried starting afresh, cleaning out the proxies for my institution. That's the extent of my tech abilities.

T&F, Sage, Wiley and Science Direct (edit: and JSTOR) have all failed on me this morning. For example:
http://onlinelibrary.wiley.com.ueaezproxy.uea.ac.uk:2048/doi/10.1111/anti.12298/abstract
http://journals.sagepub.com/doi/10.1177/1046878107300671
http://linkinghub.elsevier.com/retrieve/pii/S1474706508001666

I have not been able to find solutions in any other threads re. connector issues that work for me.

Any assitance would be grand.

Cheers,
  • I have the same (or similar) problem. In Firefox, using right-click on any pdf link, the right-click menu shows "Zotero Connector" menu item (with the red Z icon), and that has submenu item "Save to Zotero" with a little > symbol that denotes further submenu -- but there is no further submenu, and hence nothing to select to download the pdf.
  • @airquality No, different issue -- those functionalities simply don't exist for the connector currently. If you want to download a PDF in Firefox, open it in Firefox, then press the save to Zotero button. If you continue to have problems, please start a new discussion so we keep these issues separate.

    @a_rock_lobster could you try on an open access article please to help us see where the issue (which isn't a general problem) is?
    Try here
    http://journals.sagepub.com/doi/full/10.1177/1363461517725224
    If you don't get a PDF attachment could you open the PDF directly here:
    http://journals.sagepub.com/doi/pdf/10.1177/1363461517725224
    and click the save to Zotero button and see if that works?
  • @adamsmith No different I'm afraid. Zotero retrieves the metadata, but not PDF. If I follow the second link, the Zotero connector icon changes to a pdf icon; clicking this returns an error. Cheers,
  • Could we see a Debug ID from Zotero for the PDF save attempt?
  • edited February 23, 2018
    @dstillman

    From the connector (browser) side: D1838217631

    From Zotero app: D1856508610
  • (5)(+0000003): CookieSandbox: Managing cookies for journals.sagepub.com.ueaezproxy.uea.ac.uk:2048

    (5)(+0000000): CookieSandbox: Cleared cookies to be sent to journals.sagepub.com.ueaezproxy.uea.ac.uk:2048

    (5)(+0000064): CookieSandbox: Managing cookies for journals.sagepub.com.ueaezproxy.uea.ac.uk:2048

    (5)(+0000000): CookieSandbox: No Set-Cookie header received for journals.sagepub.com.ueaezproxy.uea.ac.uk:2048

    (5)(+0000001): CookieSandbox: Managing cookies for login.ueaezproxy.uea.ac.uk:2443

    (5)(+0000000): CookieSandbox: Cleared cookies to be sent to login.ueaezproxy.uea.ac.uk:2443

    (5)(+0000134): CookieSandbox: Managing cookies for login.ueaezproxy.uea.ac.uk:2443

    (5)(+0000001): CookieSandbox: No Set-Cookie header received for login.ueaezproxy.uea.ac.uk:2443
    (2)(+0000002): Downloaded PDF did not have MIME type 'application/pdf' in Attachments.importFromURL()
    It's getting a proxy login page instead of the PDF, so the normal cookie handling isn't working right. @adomasven might have an idea what's going on.
  • Downloaded PDF did not have MIME type 'application/pdf' in Attachments.importFromURL()

    If your browser does not continue automatically, click <<input type="submit" value="here">
    Cookie handling works at least partially, but Zotero downloads some sort of proxy tracking page with a JS redirect, instead of the PDF. When using your university proxy normally do you often see a blank page with the text above?
  • edited February 26, 2018
    When using your university proxy normally do you often see a blank page with the text above?
    @adomasven Sorry, I'm not sure what you mean.
  • When using your university proxy, do you often (sometimes?) see a page with text "If your browser does not continue automatically, click here", before being taken to a proxied page? Perhaps only once after login?
  • I don't recall ever having seen such a page.
  • Well, it seems that some sort of privacy system (Shibboleth ?) in your institution is intercepting Zotero's attempt to download the PDF. We're doing our best to show to such proxies that Zotero's request is genuine, and it usually works. You should try talking to your IT.
  • edited February 26, 2018
    Yes, my institution (the UEA) uses Shibboleth. They must have changed something recently, because I switched to Zotero a year ago and this has only recently started occuring. I am visiting another institution today and accessing T&F and Wiley through their eduroam seems to solve the problem. So I guess it is something to do with UEA (even though both insitutions use Shibboleth? is this a known issue?). Oh gawd, don't make me go to our IT help desk…

    Joking. Cheers for all your help,
  • edited February 26, 2018
    You can try the Zotero beta to see if that makes a difference. It probably won't, but it adds one additional bit of identifying information to PDF requests that could theoretically help.

    Basically, you're authenticated in the browser, and the connector forwards various info to Zotero so Zotero can make the request for the PDF itself. Usually this works, but in your case the proxy/authentication system is denying the request from Zotero as unauthenticated. So the technical question for IT would basically be, what other than IP address, cookies, referrer header, and user agent header is used to determine whether a request is valid? This is something that could be tested using the browser devtools and curl — find the PDF request (for a regular in-browser load) in the Network tab, copy the complete request for curl, and then (assuming it still results in the PDF in curl) figure out what has to be removed from the request for it to stop working.
  • (But given that this is happening across sites, there isn't that much left here that would make sense, so it may very well be the thing we fixed in the beta.)
Sign In or Register to comment.