New Safari Extension Does Not Download From JSTOR

miles120

After upgrading OSX to version 10.11, I did the same for Zotero and the Safari connector. Now when I try and download articles from JSTOR very little happens:

1. The connector button changes to indicate it will download the target;
2. The dialog box appears in the lower right corner saying that it is downloading to the target folder;
3. Nothing appears in my database.

Are there any issues with El Capitan and Safari 9.0?

adamsmith

There's a well-known issue using Safari with JSTOR when going through a proxy. What's a URL exactly as you see it?

msitar

I just taught a workshop where a user who had just upgraded to 10.11 experienced the same issue with JSTOR. No proxy involved, on campus in the library, seeing the same behavior described above.

adamsmith

thanks. @miles -- if you're not using a proxy either, we'd be interested in troubleshooting that with you. Same for anyone else in that situation. The initial info would be a URL exactly as you see it as well as a debug ID from Safari for a failed import:
https://www.zotero.org/support/debug_output#zotero_connectors_chrome_and_safari

msitar

Here's a debug ID from Safari for a failed import: D827354815

The URL: http://www.jstor.org/action/doBasicSearch?Query=dogs&acc=on&wc=on&fc=off&group=none

adamsmith

thanks -- same problem on individual item pages or just for search results? And do you happen to know which items where selected on import for the above?

msitar

Both on individual item pages and search results.

The first item was selected: Prevalence of Neospora caninum Infection in Dogs from Beef-Cattle Farms, Dairy Farms, and from Urban Areas of Argentina
W. Basso, L. Venturini, M. C. Venturini, P. Moore, M. Rambeau, J. M. Unzaga, C. Campero, D. Bacigalupe, J. P. Dubey
The Journal of Parasitology, Vol. 87, No. 4 (Aug., 2001), pp. 906-907

The same thing happened with the individual item. Here's the debug ID for an attempted import from the individual item: D118725446

The user in my workshop was working with a completely different set of results, but with the same behavior.

adamsmith

@Dan -- could you check the debug? I'm suspecting this, too, is related to the forced https on exporting citations from JSTOR, but would be good to confirm that.

@msitar -- could you try from this URL (the same article, but using https)? https://www.jstor.org/stable/3285150

msitar

That did the trick. It successfully imported the citation and PDF.

adamsmith

OK, at least gives us a pretty solid idea about what the problem is. I think aurimas had some ideas on a fix; I don't.
I also do wonder why JSTOR jumps around with http/https on the page: search results don't support https at all, but citation data only goes through https. Don't really see the point, but there may be one.

miles120

adamsmith,

I am not using a proxy server. The debug ID of the log I just submitted is D1120203858. The URL is below. By the way, sometimes the connector button on my menu bar is not "aware" that the page can be downloaded. If I refresh the page it fixes the problem.

http://www.jstor.org/stable/41959627?Search=yes&resultItemClick=true&&searchUri=%2Faction%2FdoAdvancedSearch%3Fgroup%3Dnone%26amp%3Bq4%3D%26amp%3Bc5%3DAND%26amp%3Bed%3D%26amp%3Bisbn%3D%26amp%3Bq0%3Dliberal%2Bhumanism%26amp%3Bpt%3D%26amp%3Bacc%3Don%26amp%3Bar%3Don%26amp%3Bf0%3Dall%26amp%3Bf5%3Dall%26amp%3Bq1%3Dscience%2Bfiction%26amp%3Bc1%3DAND%26amp%3Bf6%3Dall%26amp%3Bf2%3Dall%26amp%3Bf1%3Dall%26amp%3Bc6%3DAND%26amp%3Bq5%3D%26amp%3Bc2%3DAND%26amp%3Bc3%3DAND%26amp%3Bf3%3Dall%26amp%3Bc4%3DAND%26amp%3Bq3%3D%26amp%3Bq6%3D%26amp%3Bla%3D%26amp%3Bq2%3D%26amp%3Bf4%3Dall%26amp%3Bsd%3D&seq=1#page_scan_tab_contents

miles120

I also tried to use HTTPS and the download worked. How should I approach this? Do I need to approach JSTOR or is there something you can do?

adamsmith

you can try asking JSTOR if they'd consider serving their citation data (RIS) without https (or forcing https on their article and search pages), but they're not strictly speaking doing something wrong--it's just unusual.

The main issue is that Safari doesn't allow Zotero to make requests from a http to a https connection. I'm not sure if we're going to be able to do something about this, but if so, it's not going to be fast.

So I guess the best I have if you want a solution quickly is to use Chrome or Firefox or add the s manually every time.

@aurimas -- would it be worth trying to disable Safari compatibility in the translator so that it runs in Standalone?

miles120

adamsmith,

Thanks for the information. I will ask JSTOR but keep my expectations low. I have Chrome, but am not a big fan. Will try it and see how it goes.

adamsmith

actually miles, msitar or anyone else interested in this, if you want, you could help me test a theory:
Find the translator folder in the Zotero Data directory:
https://www.zotero.org/support/zotero_data
Download the file JSTOR.js from
https://gist.github.com/adam3smith/ecbd3a055c17124eee72/raw/c4b9d3307705a1c00d3893a9e537ccad2c75e8b6/JSTOR.js
and move it to the translator folder, replacing the existing file of the same name.

Restart Zotero, restart Safari, and try again with JSTOR. If you hover over the Save to Zotero icon, it should now say something like
Save to Zotero(JSTOR via Zotero Standalone).

Does that work?

(there's no risk in doing this, it's easy to revert).

miles120

This made no difference. Things are still the same:

1. When I hover over the Save to Zotero button, it does present the text box;
2. When I click the dialog box appears at the bottom of the screen, indicating the target;
3. There is no indication that the citation or PDF is downloading;
4. Nothing is transferred to my database.

adamsmith

too bad. Just to be sure, you do see the changed text (including Zotero Standalone) under 1., right?

miles120

It says:

Save to Zotero (JSTOR)

adamsmith

hmm - that shouldn't be the case. It should mention Standalone when using the version I posted. I don't want to take up too much of your time on this, but if you could check that there's only one JSTOR.js file in your translator folder and it has " "browserSupport": "gcib", in the header (the current one has "gcsib"), try restarting both Zotero and Safari one more time and check again.

If that doesn't change anything, we'll try to find someone with a Mac to look at this more closely (@zuphilip -- you can get to one pretty easily, right?)

miles120

There is only the one Javascript file. It does contain the header line with gcib. After restarting everything I get the same text box.

zuphilip

I can try to look at JSTOR and its behaviour with the Zotero translators with a Mac. I will try to do it today. Is there anything special I should look for?

adamsmith

yeah -- first try if you can replicate the issue that JSTOR doesn't work on Safari when on a regular http connection.

If that's the case, it'd be great to try with safari compatibility disabled. That's what I asked miles to try and it sounds like he did everything right, but the connector should show a different text on hover, so it'd be good to try this again. My understanding is that this should just pass the entire process on to Standalone, which I think could work.

zuphilip

Yes, I can replicate that JSTOR is not working on Mac with Safari. The changes the JSTOR.js in the data directory of the standalone will not make any difference. Actually, I even tried to put the empty string as browser report or changed its label, and the result is always the same and the new name is not reflected. Am I missing something here?

zuphilip

( 3 debug Reports: https://gist.github.com/zuphilip/919cdd781d2ae3f077bc )

DWL-SDCA

Please note that testing of JSTOR may not be particularly useful for a few days. Any problems encountered may be on their end and unrelated to anything Zotero.

Here is the latest from JSTOR (14 OCT):

We know this is an especially busy time of year for many of you. We are very sorry for the inconvenience JSTOR access issues may have caused for the past few days. Please know that we are continuing to make progress on restoring full access to JSTOR. At this point, most users should be able to search, browse, and access archival journal articles and primary sources on www.jstor.org. Access to books as well as journal content published in the most recent few years may be intermittent.

We will continue to provide updates as we have them through email, social media channels, and on this webpage:

http://about.jstor.org/jstor-help-support/jstor-updates

I have had various and intermittent problems for 2 1/2 days now -- time-outs, 404 errors, permissions problems -- whether on campus through a proxy or at home via my JPass.

DWL-SDCA

Update: JSTOR is now working

Thank you for your patience during this week’s site-wide access problems. We have identified the root cause and solved the issue. We strive to provide the very best service and availability of JSTOR that we can, and we will continue to do so in the coming weeks as we complete the migration to our new platform. We encourage you to check our platform updates page for information about this work, and to please let us know if any further issues arise.

We appreciate your patience and support. Please feel free to contact us with any questions or comments that you have. Your feedback is welcome.
Sincerely,
Brian Larsen
Associate Director, User Support

adamsmith

(That is JSTOR itself--unfortunately I'd expect the Safari related Zotero/JSTOR problem to remain)