pdfs no longer automatically downloaded - proxy problem?

Sometime this summer my Zotero lost the ability to automatically download pdfs from most of the sites I use. It happened across a host of translators (a list is below). From the address bar I can click the icon and generate a new item and there aren't necessarily any error messages.

Additionally, I can no longer use the right-button - 'Zotero>Save Link As A Zotero Item' to download the pdf. Again, I get no error, it just never downloads the pdf.

Occasionally I've seen the pdf show up in the Zotero window and then disappear. Looking at the Debug output (D2048123116) it looks to me like the pdf was downloaded and then deleted because it didn't have the right MIME type:

.....
(2)(+0000880): Downloaded PDF did not have MIME type 'application/pdf' in Attachments.importFromURL()

(3)(+0000000): Deleting item 1674

.....
Another strange symptom is that after I've downloaded the pdf and dragged it into the item in Zotero the 'Rename File from Parent Metadata' works but doesn't update within the Firefox screen. (Looking in the zotero storage folder for the item shows the new name.)

I've tried reinstalling Zotero - I get the same problem even before I copy back my Zotero data. I've tried running Firefox without other extensions. Neither helped. I haven't tried reinstalling Firefox.

To access pdfs I'm typically going though a proxy server - this worked before but may be the On the arxiv.org site (arXiv.org translator) I can access pdfs without the proxy. When I've found a free article on the aps.org site (APS translator) Zotero downloads the pdf - if I go through the proxy it doesn't work. Going through the proxy gives me access to the pdf, but Zotero fails to access it.

A URL that doesn't work:

http://prl.aps.org.proxy.lib.umich.edu/abstract/PRL/v105/i12/e120501


Translators that don't work with proxy -
ACS Publications, APS, Nature, DOI, Highwire,

Translators that do work (without proxy)
arXiv, APS
  • A lot of this may have to do with the fact that many for-pay sites have started to "hide" the pdfs behind pseudo URLs, which breaks the pdf download for Zotero.
    Zotero looks at the file it's being pointed to and, as you rightly conclude, sees that it's not a pdf and deletes it.

    DOI never had pdf download in the first place.

    That doesn't explain the APS issue, though - the article you link to works for me with and without proxy. I can't come up with a good reason why your proxy would break the pdf link here. hmmm
  • Accept Third-Party Cookies !

    I tracked down the source of the proxy / pdf download issue. The behavior described above was the result unchecking the 'Accept Third-party Cookies' option in Firefox.

    I'm running Firefox 3.6.10 and Zotero 2.0.8 and I didn't see any mention of having to accept 3rd party cookies on the support web pages for this combination. There are a couple of pages on university library sites that suggest this for general problems.

    Perhaps someone else could try checking this with other proxy servers to see if it's a general problem or specific to the umich proxy.
Sign In or Register to comment.