Item title contains unicode character and result in bibliography export issue

edited September 18, 2019
I have an article (https://doi.org/10.1039/C7LC00644F ) with the character, μ, in its title. When I copy the item as Better BibLaTex by going to Edit >> Copy as Better BibLaTex, the resulting copied text no longer contains the μ (%u03BC), character and it gets replaced to the character Μ (%u039C).

I think this is NOT an Better BibTeX issue because when I right click on item and do Create Bibliography from item >> Chicago Manual of Stytle 17th edition (note) >> Copy to Clipboard , the resulting copied text replaced the original μ character to Μ (%u039C).

This issue is causing a problem in LaTex compiling. I get the following error below because the exported Better BibLaTex item contains the character Μ instead of μ.

Package inputenc Error: Unicode character Μ (U+39C)(inputenc) not set up for use with LaTeX.

When I manually replace the Μ character in my .bib file of LaTex project to μ character, I once again get the following error:

Package inputenc Error: Unicode character μ (U+3BC)(inputenc) not set up for use with LaTeX.


I found out that if I use the character µ (%B5) , I don't get any issues. However, if I used the default character μ (%u03BC), I get issues. The default character is the character originally in the article title.

I am unsure what's the core problem is and how to fix it.

My system:
Zotero v5.0.74
Better BibTex v5.1.139
Windows 10 x64 v1903 Build 18362.356
  • I can't help with the LaTeX part — I assume that's just a configuration problem — but for the bibliography output, citeproc-js is capitalizing the word as per the style, and the uppercase letter for μ is Μ (not to be confused with the regular ASCII M):

    > 'μ'.toUpperCase()
    'Μ'


    @fbennett, any ideas here? Is title-casing in citeproc-js currently restricted to any specific character ranges, or will it just do anything toUpperCase() supports?
  • If you open an issue on the BBT tracker I can look at the bibtex export.
  • edited January 6, 2020
    I'd be curious to know this as well. BBT's titlecaser builds on the citeproc titlecaser, but has a facility to exclude parts from titlecasing, so I can adjust it to not titlecase μ. But in general I try to have BBT create bib(la)tex that would render to the same output that citeproc creates -- so if Zotero title-cases μ to Μ so would (does) BBT.

    Unless citeproc would be clearly at fault, but between me knowing more about bibtex, or Frank knowing more about CSL, there's absolutely no contest (in this you might picture Frank as Usain Bolt, and me as, well, me, doing a 200 meter dash), so I consider this to be extremely unlikely.
  • @emilianoeheyns Thanks for the update. I just want to confirm you guys don't need additional logs or troubleshooting steps on my end? It is my understanding you guys are waiting on Frank to provide input on how to handle situations like these?
  • I don't need further logs. I'm indeed waiting for Frank's feedback, but frank is always crazy busy, so we have to allow for this to take some time.
Sign In or Register to comment.