Ref data successfully grabbed, but PDFs and snapshots consistently fail

benramage
EDIT: Here's the Report ID: 1268870514

I'm having a problem saving PDFs and snapshots from supported websites with the Chrome extension. I've tried many different sites that I have full institutional access to, and I am consistently able to get the reference data without any trouble, but in the pop-up in the lower right, an "X" appears next to "Full Text PDF" and the PDF never arrives in Zotero standalone. No "X" appears next to "Snapshot", but that doesn't arrive in the standalone program either. This error is reported in the Chrome extension every time:

[JavaScript Error: "Cannot read property 'id' of undefined" {file: "chrome-extension://ekhagklcjbdpajgpjgmbionohlpdbjgc/background.js" line: 75}]

And yes, I have confirmed that the extension is reporting that the standalone program is currently available and that both "Automatically..." check boxes are checked. Thanks so much for the help...
«1
  • aurimas
    Please provide a debug id for a save attempt http://www.zotero.org/support/debug_output
  • adamsmith
    provide a couple of example URLs exactly as you see them
  • benramage
    Here you go, aurimas: The Debug ID is D894261792. Thanks.

    adamsmith, here are two examples:
    http://www.sciencedirect.com/science/article/pii/S0378112713005112
    http://onlinelibrary.wiley.com/doi/10.1111/1365-2664.12149/abstract
    I definitely have full access to both
  • adamsmith
    @Dan - anything helpful in the debug?

    @benramage - and just to be 100% sure - those are the URLs as you see them in Chrome, right? You didn't remove any proxy or so?
  • benramage
    yes, that's right adamsmith. I do have a proxy server set up (which is necessary to have access to the full articles), but those are the URLs copied directly from Chrome
  • adamsmith
    That might still be relevant, though ideally it shouldn't be. What type of proxy is it?
  • benramage
    http://proxy.lib.berkeley.edu:7777/proxy.pac
    "Automatically detect proxy settings" checked
    and
    "Use automatic configuration script" checked
  • adamsmith
    Dan will have to weigh in but the pac proxies used on the UC campuses have caused problems before - I believe Standalone doesn't pick them up correctly.
  • benramage
    You might be right adamsmith. I just disabled my proxy server and I was able to successfully import a ref and PDF from an open-access journal. I intended to then turn my proxy server back on and try the same open-access article again, but now Chrome is refusing to recognize my proxy settings (despite them working fine in IE), but that's another (extremely frustrating) story... ugh, it's just been one of those days. In any case, unfortunately, I am currently unable to test anything with the proxy server functioning in Chrome.
  • adamsmith
    also, could you confirm that you have Zotero version 4.0.11 (those things are in the error report and debug, but only Simon and Dan can actually access those)
  • benramage
    UPDATE: The proxy server seems to be working (at least partially) again. I can still properly retrieve the open-access PDF, but I have the same old problem with the paywall content. So... it seems that the proxy server is not blocking PDF downloads in general, but the Chrome extension is failing to use it to access PDFs for which it is required. ???
  • benramage
    Yup, newest version. Just installed it today.
  • adamsmith
    right I'd expect that: for whatever reason the request for the PDF doesn't appear to go through the proxied connection, so Zotero gets a paywall instead of the PDF and throws it out bc it has the wrong mimeType.
    I don't have the expertise to troubleshoot this further, though, but I'm pretty sure the proxy is going to turn out to be the reason. Dan or Simon will have to take it from here.
  • aurimas
    Make sure that 'network.proxy.type' is set to '5' in Zotero Standalone under Preferences -> Advanced -> General -> Open about:config
  • benramage
    It is. I just checked.
  • benramage
    Hi again, aurimas. Any other thoughts about how to fix this problem? Thanks.
  • aurimas
    Unfortunately nothing that you can do atm. We're waiting for Dan to take a look at the Report ID you submitted, which should give us a lot more information. Only the official Zotero developers have access to those reports. I think Dan just missed this thread yesterday.
  • benramage
    Anyway I can draw this thread to Dan's attention? (or are you saying he is indeed aware of it, but just hasn't gotten to it yet?) Thanks!
  • aurimas
    Yeah, idk what's going on. Usually Dan is very fast to respond to these and he has responded to a few (though very few) threads yesterday... Perhaps he's on vacation. I'll try to contact him directly.
  • dstillman
    PDFs download via Standalone, so if there's a problem downloading PDFs we'd need a Debug ID from Standalone in addition to the connector.
  • benramage
    Here you go, Dan. D2041971435
    Thanks.
  • dstillman
    proxy2.lib.berkeley.edu:443 uses an invalid security certificate.
    Berkeley's proxy server is misconfigured, and they'll need to fix it. You can point them to this post (https://forums.zotero.org/discussion/32231/#Item_23) for the details.

    === Technical details ===

    If you take a look at https://www.ssllabs.com/ssltest/analyze.html?d=proxy2.lib.berkeley.edu&hideResults=on you can see that the Berkeley proxy server's HTTPS support has lots of problems. They get an F because they're using very weak encryption and because of some other issues, but the relevant issue for Zotero is the "Chain issues: Incomplete" and "Extra download" lines.

    The server isn't serving the necessary intermediate (Internet2-provided) SSL certificate, and Firefox — and therefore Zotero Standalone — won't download it on its own. Other browsers might do so, though I don't know for sure. (I think Chrome does.)

    If you can load https://proxy2.lib.berkeley.edu in Firefox, it's likely because you previously loaded another site of your university's (or another university's) that included the intermediate certificate, which Firefox would then cache and use even on sites that didn't serve it properly. But generally speaking, sites should always serve their intermediate certificates, and are broken if they don't. If you create a new Firefox profile, you'll get an sec_error_unknown_issuer error trying to load that URL, which is essentially the situation Zotero Standalone is in.

    === End of technical details ===

    So the proper fix here is for the Berkeley IT folks to fix the server. In the meantime, I have an awkward workaround for you. Go to the Zotero sync preferences and set it to use WebDAV for file syncing. For the URL, set it to "HTTPS" and "spaces.internet2.edu". Use "test" and "test" for the username and password. Click on Verify Server. (You might need to click it twice due to a bug.) Once you get the message that https://spaces.internet2.edu isn't a valid WebDAV URL, clear those fields and change the setting back to syncing via Zotero (or whatever you were using before). Since spaces.internet2.edu uses (and properly provides) the same intermediate certificate, the certificate will be downloaded into Standalone, and after that Standalone should be able to access the Berkeley proxy server.
  • benramage
    Thanks so much for the assistance, Dan, but unfortunately it's still not working. The behavior has changed (it now successfully downloads the snapshot despite getting stuck on the PDF, which it didn't do before), but the PDFs still fail to come through.
  • dstillman
    Can you restart Standalone and provide another Debug ID?
  • benramage
    Here you go: D875241385.
  • dstillman
    <HTML>
    <HEAD>
    <TITLE>Proxy: Cookies Required</TITLE>
    </HEAD>
    <BODY>
    <H1>UC Berkeley Library Proxy</H1>
    <p>The proxy was unable t...
    OK, that's a new, unrelated issue that aurimas or possibly Simon will need to look into. The good news is that you can now at least connect to the Berkeley proxy server, so that's a start.
  • aurimas
    @Dan, is the PDF host different from the web page?
  • benramage
    Thanks so much for your input. I'll wait for a response from the others.
  • adamsmith
    @ben - could you try this on the
    http://onlinelibrary.wiley.com/doi/10.1111/1365-2664.12149/abstract
    page?
  • benramage
    same problem
Sign In or Register to comment.