Bug: Zotero saved proxy URL with entries

muelli

Apparently, Zotero stores the proxy, it used to access stuff on, say, ACM, with the entry. So the saved URL becomes, i.e. http://portal.acm.org.remote.library.dcu.ie/citation.cfm?id=1247402 as opposed to http://portal.acm.org/citation.cfm?id=1247402

I expected Zotero to save the real URL, because the proxy information is stored in Zotero anyway.

adamsmith

I just wanted to confirm this and put this on the screen again. This is also true for NBER papers such as
http://www.nber.org/papers/w15902
(the URL of which is saved as http://www.nber.org.turing.library.northwestern.edu/papers/w15902
for me)

ajlyon

I think this appears when translators try to pull tricks to get a good stable URL into the URL field; I have a potential fix for the ACM translator that might address this. Please take a look at http://github.com/ajlyon/zotero-bits/raw/master/ACM.js and see if it works better. I don't use a proxy for any access, so I can't really test this one.

If this doesn't work, it would be nice for Zotero to provide translators with a function that de-proxies a URL.

ajlyon

In the case of the NBER translator, it appears that your proxy is catching the URLs inside the RIS file and adding the proxy information to them. We could try to switch to MODS or BibTeX, which NBER also provides, but I don't know if that will help. This is certainly a case where we need to manually deproxify, if we can't get the proxy to let the bibliographic formats pass through untouched.