EBSCOHost: Blocked a frame error

schultzbergin
Hello,

I hope this isn't the same issue as the one caused by proxy issues with EBSCOHost - the error is different, so I thought it warranted its own post at least.

I receive the following error when attempting to grab a file from EBSCOHost, and it prevents me from getting either the metadata or the document:

[JavaScript Error: "Blocked a frame with origin "http://0-web.b.ebscohost.com.maurice.bgsu.edu" from accessing a cross-origin frame." {file: "chrome-extension://ekhagklcjbdpajgpjgmbionohlpdbjgc/inject/http.js" line: 42}]

If I grab the Permalink from EBSCOHost, and paste that into the browser, then Zotero will save the metadata information, but will fail to provide me with the PDF document.

I am (currently) accessing EBSCOHost through my university's library, but off campus. I did have no problem, however, getting metadata and documents with zotero for Springer.

Any thoughts?

Thanks,
Marcus
  • adamsmith
    what's the full URL as you see it? (not the permalink one - the one where this actually fails)
  • schultzbergin
    My apologies for not mentioning that before. That URL is:

    http://pq9se9hp4e.search.serialssolutions.com/?ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info:sid/summon.serialssolutions.com&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Contractarianism%2C+Other%E2%80%90regarding+Attitudes%2C+and+the+Moral+Standing+of+Nonhuman+Animals&rft.jtitle=Journal+of+Applied+Philosophy&rft.au=COHEN%2C+ANDREW+I&rft.date=2007-05-01&rft.pub=Blackwell+Publishing+Ltd&rft.issn=0264-3758&rft.eissn=1468-5930&rft.volume=24&rft.issue=2&rft.spage=188&rft.epage=201&rft_id=info:doi/10.1111%2Fj.1468-5930.2007.00362.x&rft.externalDocID=JAPP362&paramdict=en-US
  • adamsmith
    yeah, I don't think we can currently get the serialssolution version of URLs to work reliably in Chrome, sorry.
    I think you should be able to access EBSCO in a way that it's not embedded in a serialssolution frame though. Try using
    http://0-search.ebscohost.com.maurice.bgsu.edu/login.aspx?authtype=ip,uid&profile=ehost
  • schultzbergin
    Yeah, that works in the same way as doing the permalink - Zotero will then save the metadata, but it will fail to grab the full PDF.
    That isn't the worst thing ever (I then download the PDF and then link it to the citation from within Zotero) but it does require extra steps.

    Should it be the case that accessing EBSCO in this way does let me grab the full text PDF? Is that perhaps a different issue I am having?

    Also, when you say you can't get zotero to work with serialssolution URLs in Chrome, does that mean they work fine in (say) Firefox?
  • aurimas
    edited March 18, 2014
    @adamsmith, is that because the EBSCO page is proxied through serialssolutions (in which case, would we even detect it as such?) [edit: I see that it's not]. Otherwise, this should (in theory) not be a cross-origin request. I don't think I have access through Serials Solutions

    @schultzbergin, to help us (or at least me) better understand how serials solutions works, could you submit a Debug ID (from Chrome, not Zotero Standalone) for an attempt to save an article through serials solutions?
  • adamsmith
    edited March 18, 2014
    You won't get PDFs from some sites (notable EBSCO, Proquest, and ScienceDirect) when using Chrome and a proxy connection. That will definitely work with Firefox (with Standalone closed)*.
    Serialssolution might work, too, but not sure (and no way to test for me).

    *note that if you install Zotero for Firefox it will, by default, use the same database as Zotero Standalone, so switching between the two is seamless.

    @aurimas - the way I think that works is that EBSCO (and other resources) run in a frame of the serialssolution page - which is why we detect them. But I don't have access either, so that's just based on my best understanding.
  • aurimas
    @aurimas - the way I think that works is that EBSCO (and other resources) run in a frame of the serialssolution page - which is why we detect them. But I don't have access either, so that's just based on my best understanding.
    That's what I think too. But then I'm not sure what frame Zotero is trying to access that is cross-origin.
  • schultzbergin
    Adamsmith, thanks again for the (especially quick!) assistance! I will checkout how things work in Firefox and report back.

    @aurimas - the information I provided before was from the zotero connector, not standalone, but here is also what chrome reports via developer tools:

    SecurityError: Blocked a frame with origin "http://0-web.a.ebscohost.com.maurice.bgsu.edu" from accessing a cross-origin frame. at chrome-extension://ekhagklcjbdpajgpjgmbionohlpdbjgc/inject/http.js:42

    Hopefully that is what you are looking for... let me know if not
  • adamsmith
    edited March 18, 2014
    aurimas is asking for a debug ID for the process of trying to import the item. See instructions here:
    https://www.zotero.org/support/debug_output#zotero_connectors_chrome_and_safari

    @aurimas - I see now, ebsco is the origin frame, this is odd indeed. Or maybe that message is just for the PDF and not what actually causes import to fail?
  • aurimas
    Or maybe that message is just for the PDF and not what actually causes import to fail?
    PDF download would be happening (and failing) via Standalone though. But this could be caused by us trying to get to the PDF or just the metadata. Debug ID would tell us more.
  • schultzbergin
    Hi again,

    Sorry for the mistake earlier and thanks for the instructions on how to do it right.

    I followed those instructions and the debug ID is D1748700667

    Just to clarify:
    If I access ebsco normally and attempt to import, nothing happens (as far as the page is concerned - i.e., no popup indicating it is grabbing metadata or PDF).
    If I then open the same page with the permalink, Zotero will now import the metadata but not the PDF.
    The debug I did was the former case (the regular ebsco page rather than the permalink)


    Thanks again for the assistance.
  • aurimas
    @Dan, could you post the output to a gist?
  • adamsmith
    yes, that was right. We know what's happening with the permalink, that's expected.

    @Dan - could you give us the relevant parts of the debug?
  • dstillman
    https://gist.githubusercontent.com/dstillman/3673fd49287e8e4aa117/raw/96c3e779f5e19aba8fbb67729f6b6186759b35b3/gistfile1.txt
  • aurimas
    edited March 18, 2014
    So that would be breaking here: https://github.com/zotero/zotero-connectors/blob/master/src/common/inject/http.js#L42 Seems that accessing location.protocol of the parent frame (serialssolutions.com in this case) is not a safe operation. I'm not entirely sure why we're looking at the parent frame though. Perhaps Simon could comment (and we should probably move this to github)

    Edit: ticket created https://github.com/zotero/zotero-connectors/issues/20
  • aurimas
    Looks like this might be fixed in the next version of Zotero Chrome connector. @schultzbergin, if you have a way, could you verify that this works correctly in Firefox?
  • schultzbergin
    In Firefox, if I try to import from the normal link (not permalink) then it will import the metadata, but will fail on importing the PDF (so it does the same thing Chrome does when using the Permalink with Chrome).

    In either case, the full text PDF fails to download.
  • adamsmith
    with Firefox and the permalink you don't get the PDF? Do you have Zotero Standalone closed when you try this?
  • aurimas
    edited March 18, 2014
    Is that with Zotero Standalone open? I forgot to mention that you should try with standalone closed. If it is with standalone closed, then could you submit a debug ID from Firefox with Standalone closed for an attempt to import article?
  • schultzbergin
    My apologies, I did have standalone open.

    When I try to import via Firefox with standalone closed, everything works fine (using the regular link)
  • adamsmith
    cool, so we have everything figured out now.
    I hope Simon pushes out the new connector soon, there are a couple of other things I really want that have been implemented in the connectors.
  • schultzbergin
    I was wondering if there was any update on this front?

    The Chrome Connector is still version 4.0.8.2 which, according to the Chrome extensions page was released in June 2013.

    And I still have issues with any serialssolutions database. I have a basic workaround I now use (download the pdf, import pdf to zotero, have zotero grab metadata) but I would prefer to be able to do it all directly.

    Thanks!
    -Marcus
  • adamsmith
    no, as you say, the new connector version hasn't been released. I can't say why not, but last week Simon updated the submodules, which usually precedes a release. No promises, though.

This is an old discussion that has not been active in a long time. Before commenting here, you should strongly consider starting a new discussion instead. If you think the content of this discussion is still relevant, you can link to it from your new discussion.

Sign In or Register to comment.