JSTOR inconsistencies

slg-lc
We're experiencing an inconsistency at our institution that I can't figure out so far, related to JSTOR. Each machine has the same OS (WinXP) and the same Firefox (4.0.1) and the current Zotero.

On some of the machines, when "Automatically attach associated PDFs" is checked that's exactly what happens for JSTOR articles. On other machines, for the exact same articles even, the PDF is not downloaded and attached--and this is after the user manually opens a PDF and accepts the JSTOR terms of service for the session. There is no error message, of course--just no PDF. The metadata is fine.

On the machines that won't download JSTOR articles, it is still possible to get PDF attachments from other databases, such as Project MUSE, so there seems to be something specific going on with JSTOR. Any ideas? Could it have something to do with preferences set by the users elsewhere in Firefox?

Thanks!

slg
  • ajlyon
    edited July 28, 2011
    Can you provide debug output (http://www.zotero.org/support/debug_output) for an attempt to save from the computer that is having trouble? Post the debug ID here.
  • slg-lc
    From a computer that is not grabbing the JSTOR PDFs:

    The Debug ID is D1758098412.
  • ajlyon
    Dan: Can you look in slg-lc's debug output and see if the PDF is being discarded for a MIME type mismatch?
  • Simon
    Yup.
    (3)(+0000004): Importing attachment from URL

    (3)(+0000000): HTTP HEAD http://www.jstor.org/stable/pdfplus/20355384.pdf?acceptTC=true
    [...]
    (2)(+0000253): Downloaded PDF did not have MIME type 'application/pdf' in Attachments.importFromURL()
  • slg-lc
    Same JSTOR article from a computer that did bring down the PDF along with the metadata:

    The Debug ID is D1212934957.

    Thanks!
  • Simon
    The URL of the PDF in the second report is the same as above. What happens if you go to that URL on the computer that won't download the PDF?
  • slg-lc
    Hi, Simon, it opened as expected. The only aspect of JSTOR that doesn't seem to be working on the first computer is that Zotero doesn't automatically bring down the PDF along with the metadata, even when the appropriate box is checked in preferences.
  • ajlyon
    I know it's not supposed to matter for Firefox 4, but can you make sure that you have Firefox set to allow third-party cookies in the Firefox preferences?
  • slg-lc
    edited July 28, 2011
    Ok, get this. None of the computers had allow third-party cookies enabled, and I tried that on the one that wouldn't auto download PDFs and that still didn't work.

    But this made me realize that the one where auto download does work has had Firefox/Zotero installed for a long time, going back to some point in the past where JSTOR functionality required the explicit addition of a cookie "allow" exception. The exception was still present.

    So I went over to a machine where the auto download was not working, added the explicit "allow" cookies from www.jstor.org, and voila!

    Is this some weird problem due to the way Firefox is locked down at our institution? The way JSTOR is served?

    Thanks for your help!

    slg
  • ajlyon
    You would have to allow some cookies for it to work, so perhaps you had all cookies disabled? Still, the important thing is that it's working now. Happy reading.
Sign In or Register to comment.