sep is undefined

edited January 26, 2017
Hello, could you help me, please?
Zotero refuses to update the old references in the document or insert a new one flashing a small window with "sep is undefined" message.
Note: I am using several forms of accents and have no idea which item is causing the problem...
«1
  • Haven't seen that in ages -- could you start by confirming your Zotero version?
  • Hello thanks for your reply.
    It is 4.0.29.15
  • Is there anything I can do on my end to help with your research?
  • Thanks for your advice. I did points 1-7. Zotero gave me this error ID: 987256886

    Now I am deleting citations one by one to find the problematic citation.
  • note the exact instructions there -- You'll be much faster if you split the document and narrow the error down than if you actually go one- by-one.
  • Hello again, the message "sep is undefined" stopped when I got the the citation 27 out of 200. Here it is:
    ‘Hashiw Hogabardzut‘yan Oriordakan Yerjankahishatak Arghut‘eants‘ H. Dprots‘in (The Report of the Armenian School for Girls Named after Late Arghut‘eants‘)’, 152–54.

    Could you advise me what to do? All my transliterated citations carry this kind of accents, and unfortunately, I cannot eliminate them. Is there a way to keep them and not to experience problems with zotero?
  • What happens if you use that same citation in a new document?
  • Nothing extraordinary happens. It inserts the reference normally.
    I made another copy of the same document and deleted this problematic citation, but the "sep is undefined" window appeared again. I am assuming it is not this particular citation but a combination of my transliterated citations that zotero does not like.

    Do you think there are other solutions?
  • @Dan%Stillman -- could you check the error report if there is anything else?
    @fbennett -- I assume this is citeproc-js; how do you want to troubleshoot this?
  • I assume this was just a bug in the citeproc-js version (1.1.136) in 4.0.29.15, which is the latest version available for macOS. It's likely fixed in the 5.0 beta, which has a newer citeproc-js version.
  • @hayarpi.papikyan if you don't want to use the beta, you can try installing this add-on: https://github.com/Juris-M/propachi-vanilla/releases/latest which will just update the citation processor, leaving Zotero as it is.
  • Hi, thanks for the suggestion, I tried the plugin. It is not helping. This is getting discouraging. Moreover, you are saying 5.0 beta "is likely" to fix...
  • @adamsmith Could still be failing in the latest (?) - I haven't touched that part of the code in a long, long time. Is there a test fixture that reproduces the bug?
  • I am in the middle of my dissertation and have spent three days trying to fix this problem. Do you think 5.0 will help me and will not create extra issues with these accents and the transferring of my library?
  • No, I don't think 5.0 is likely to help if the propachi add-on doesn't.
    The most helpful thing you could do is to try to create a minimal case of a citation/document that triggers this error. It's very hard to fix an error that we can't replicate on our side.
  • Thank you for your reply. Please, correct me if I got your message wrong. Are you suggesting to create documents with fewer citations?
  • See step 8 at https://www.zotero.org/support/word_processor_plugin_troubleshooting#debugging_broken_documents. Assuming there is only one problematic citation, that will help you identify it.
  • Hello, I did what you suggested about seven hours ago (see in the discussion above). After finding seemingly problematic citation, I made another copy of the whole document and deleted this problematic citation, but the "sep is undefined" window appeared again.

    I tried with older documents, where I did not experience any problems. It opens the same window... (!)
  • Meanwhile I tried Beta 5.0. It does not help either.
  • If the processor plugin doesn't fix it, 5.0 won't either.

    Hang on for a day. I have just finished several days' work on text-case transforms in the processor code.

    By coincidence, the method I used there can also be applied to quote/apostrophe formatting. If that is done, this problem will go away.

    I just need to put in time on it.
  • Thank you @fbennett ! Of course, I will wait. It is like life-and-death situation for me, as I have to submit my dissertation this June and there is not time to do the references and bibliography manually.
  • Thanks for understanding. I'm very sympathetic to the pressures you're under (I'm an academic supervisor myself).

    I'm about halfway to a solution with the new coding pattern, and still very confident that it will be more reliable than what I wrote originally. I hope to get it up and running in the processor tomorrow.
  • Thank you @fbennett . Your and @adamsmith 's support and responsiveness is much appreciated.

    Just a little note. Now I have Zotero 5.0 on my computer (it's mac os version 10.12.3). Yesterday after numerous experiments when I tried 5.0 beta, it did not allow me to go back to the 4.0.29.15 that I was using.
  • Yeah. Zotero 5.0 is a one-way upgrade. There are ways to downgrade, but the processor fix I'm working on can be installed in Zotero 5.0 also, so you can stay on that version.

    More news fairly soon!
  • I did what you suggested about seven hours ago (see in the discussion above)
    Oh, sorry, missed that.
    the message "sep is undefined" stopped when I got the the citation 27 out of 200. Here it is:
    ‘Hashiw Hogabardzut‘yan Oriordakan Yerjankahishatak Arghut‘eants‘ H. Dprots‘in (The Report of the Armenian School for Girls Named after Late Arghut‘eants‘)’, 152–54.
    Have you already tried making a copy of your original file, and deleting everything in that copy except the citation above (and maybe the Zotero bibliography)? If that still gives the error, that might make for a interesting minimal example, even if citing the same item in a new document doesn't give the error.
  • @hayarpi.papikyan,

    Making good progress on the solution. I have the new in-text markup + quotes-parsing code running. It needs some small refinements to get it to pass all of our tests, but it's looking good, and should be ready for use in another day or so.

    Thanks for your patience.
  • Thank you very much @bennett for starting my morning with good news!!!
  • Hello @Rintze, sorry for a late reply. I have done what you recommended already several times. As I mentioned earlier in the discussion, the problem is not in this particular reference only, but the accumulation of my transliterated citations (I use French, transliterated Armenian and Russian). In theory there is nothing wrong with these citations as I use the Library of Congress guide for romanising Armenian and Cyrillic alphabets. As for zotero, I suppose it does not like this large collection of the accents, because when I delete the half of the document it starts updating and inserting the citations normally.
  • edited January 31, 2017
    @hayarpi.papikyan,

    Still poking a long with the update. It's taking a little more time than predicted, because a portion of my initial code needed a rewrite (specifically for quotation marks and apostrophes, the semantics of which are quite tricky for the machine, although not so much for people). The parsing works are now all in place, and there is just a small amount of work to do. I'm pretty confident that the revised processor will be ready for testing tomorrow. I'll be eager to see whether it solves the problems with your document.
Sign In or Register to comment.