[SOLVED] Insert Bibliography fails -- Report ID 1882608647

DVBQQ

I've just completed a 100,000 word PhD thesis. I'm using Zotero 1.0.10, OpenOffice 3.1.1 with the extension Zotero-1.oxt, Firefox 3.5.2 and Vista Home Premium.

After weeks of final editing and checking of the text, this morning I clicked the button marked "Insert Bibliography" -- and my worst fear happened: an error page.

It's called [My Macros & Dialogs].Zotero. At the bottom it says "Ln 635, Col 60". One line is highlighted: mTagStart(nTagIndex) = Len(sResultText)

Can anyone help, please? I have to take my printed thesis to the binders next week.

adamsmith

OK - several suggestions:
First: do you get an error report in Zotero (gears menu, Report Error - send error and post ID here).

Second, what happens if you disassemble the dissertation - Do you get an error in all files or just in one?
What happens if you ctr.+A the entire text, and copy it into an empty document - still the same problem?
That's all I can think of atm.

DVBQQ

Hi Adam

Good to hear from you again, and thanks for coming back to me so quickly. (Sorry for my delay in replying -- I had to go out to buy five reams of printer paper...)

The Report Error, oddly, only shows the preview text of my post here -- preceded by "[JavaScript Error: "not well-formed" ".

Disassembling: if I do that, I'll only get part of the bibliography, relating to an individual chapter. That's why I waited until I'd put everything into one massive file (which had assorted OOo formatting problems, like losing some full stops immediately following citations).

I've just tried again, within the same file. There's an OOo error message: "Inadmissible value or data type. Overflow." Perhaps it's an OOo problem rather than a Zotero problem.

I'll have a go at copying the text and let you know how that goes.

David

DVBQQ

I copied all the text into a new file.

Same OOo error message: "Inadmissible value or data type. Overflow."
Same line highlighted in the file [My Macros & Dialogs].Zotero: Ln 635, Col 60, "mTagStart(nTagIndex) = Len(sResultText)"
I sent an error report to Zotero; the ID is 1882608647.

I'll contact the OOo forum as well, in case anyone there recognises the problem -- though they don't tend to reply anywhere near as quickly or usefully as people here.

adamsmith

I understand that disassembling won't solve the problem - that was meant as a suggestion for troubleshooting - to me this sounds a little like one corrupted citation is messing up your entire document. If you can isolate that citation you should be fine.

DVBQQ

Sorry I didn't get back last night -- on top of everything else my internet connection went down. I finally realised what you meant, stopped panicking and started analysing.

I created a bibliography from each of the original individual chapters. No problems -- that rules out dodgy data. I went back to a copy of the whole thesis, deleted the first half, no probs, deleted the second half, no probs, tried a copy of a copy of the whole thing: error. Deleted photos, diagrams, tables in case Zotero doesn't like non-textual stuff: error. So it seems the quality of the content isn't the problem. Could it be the quantity? I deleted the first third, no probs, deleted the last third, no probs, so it can handle c. 200 pages, so text length seems unlikely to be an issue. But that prompted a thought.

I went through the entire 288 pages deleting large blocks of text that didn't have any citations in them. Eventually I ended up with just 50 pages of text but with all the citations. Tried to create a biblio: error. Deleted 10 pages: error. Deleted another five pages: no probs.

Conclusion: OOo &/or Zotero can handle a long text file with no difficulties, but has a limit on the number of citations (or perhaps the number of entries in the biblio) it can handle when creating a bibliography. (And no, I haven't counted them -- yet.)

I went back to the whole thesis and experimented until I could successfully create a biblio from the largest number of chapters, 3-8, then did the same for 1-2. Then printed them both out and went through them with a red pen, deleting dupes and sorting out the inevitable confusions over 1986a, 1986b, 1986c etc. Then, not having any alphabetical file merging s/w that I know of (25 years ago I could have written a program in Unix shell language in 5 minutes to do it!) I put the two files side by side on screen and spent two hours laboriously copying from one to the other, one entry at a time, finishing well after midnight. But now I have a bibliography for the whole thesis: Joy.

I've gone into this much detail for two reasons: 1. to help anyone else who gets into the same scary situation with time running out; and 2. so that the programmers at both OOo and Zotero are aware of the problem and can maybe attempt to fix it in the next release. (If they haven't already; I'm aware that I'm not using the most recent version of Zotero, because it upgraded part way through my work and it didn't seem wise to change horses mid-stream.)

Thanks again for your help, and for calming me down when I was panicking.

David

adamsmith

thanks a lot! Much appreciated.
If you get a chance (and hopefully after ample celebration) - a rough estimate on the number of citations would be great to know.

Also, when the new plugin comes out (which might even be today) - could you give this one more try to see if it was fixed?

In any case - this really does sound like it's the number of citations, which is something that really shouldn't happen. If dev could take a look that'd be great.

DVBQQ

The celebration will have to wait until after the viva (I hope) in a few months' time -- but yes, once I've printed, bound and submitted the thesis I'll have time to look through the problem again, and report back.