Postprocessing doesn't show errors during processing of bibliography

[asmaier]

When processing the bibliography latexmlpost only shows the number of errors encountered, but doesn't give any hint, what went wrong:

latexmlpost --verbose --verbose --verbose --dest=equations.html equations.xml 

(MakeBibliography equations.html processing... [Converting bibliography equations/test.bib ...

Conversion complete: 1 error.
Status:conversion:2 
... converted!] [using bibliographies test.bib] [154 bibentries, 9 cited] [DBStatus: 8195/0 objects] 13.45 sec)

With a big *.bib file, it can be quite painful to track down these errors. It would be nice, if latexmlpost would at least give the line number, where it encountered problems parsing or processing the *.bib file.

[brucemiller]

Do you happen to have a small sample case handy to test?

[asmaier]

Here is a test latex and test bib file:

test.tex:

\documentclass[a4paper,12pt,twoside,openright]{book}
\usepackage[english]{babel}     
\usepackage[utf8]{inputenc} 
\usepackage{natbib}
\begin{document}
\tableofcontents
\chapter{Introduction}
blabla
\nocite{*}
\bibliographystyle{apalike}
\bibliography{test}

\end{document}

test.bib:

@ARTICLE{Bryan1997,
  author = {Bryan, Greg L. and Norman, Michael L.},
  title = {{A Hybrid AMR Application for Cosmology and Astrophysics}},
  year = {1997},
  eprint = {astro-ph/9710187},
  pdf = {Bryan1997.pdf},
  slaccitation = {%%CITATION = ASTRO-PH 9710187;%%},
}

@ARTICLE{Ensslin2006,
  author = {Enßlin, T.~A. and Vogt, C.},
  title = {{Magnetic turbulence in cool cores of galaxy clusters}},
  journal = {A\&A},
  year = {2006},
  volume = {453},
  pages = {447-458},
  month = jul,
  adsnote = {Provided by the SAO/NASA Astrophysics Data System},
  adsurl = {http://adsabs.harvard.edu/abs/2006A%26A...453..447E},
  doi = {10.1051/0004-6361:20053518},
  eprint = {arXiv:astro-ph/0505517},
  keywords = {galaxies: cluster: general, cooling flows, magnetic
     fields, turbulence, X-rays: galaxies: clusters, intergalactic medium},
}

@ARTICLE{Kiessling2003,
  author = {Kiessling, M.K.-H.},
  title = {{The ''Jeans swindle'' - A true story-mathematically speaking}},
  journal = {Advances in Applied Mathematics},
  year = {2003},
  volume = {31},
  pages = {132-149(18)},
  month = july,
  doi = {doi:10.1016/S0196-8858(02)00556-0 },
  pdf = {Kiessling2003.pdf},
  url = {http://www.ingentaconnect.com/content/els/01968858/2003/00000031/00000001/art00556},
}

@ARTICLE{Veynante2002,
  author = {Veynante, D. and Vervisch, L.},
  title = {{Turbulent combustion modeling}},
  journal = {Progress in Energy and Combustion Science},
  year = {2002},
  volume = {28},
  pages = {193-266(74)},
  month = March,
  doi = {doi:10.1016/S0360-1285(01)00017-X},
  pdf = {Veynante2002.pdf},
  url = {http://www.ingentaconnect.com/content/els/03601285/2002/00000028/00000003/art00017},
}

@INPROCEEDINGS{Norman1999,
  author = {Norman, M.~L. and Bryan G.~L.},
  title = {{Cosmological Adaptive Mesh Refinement$^{CD}$}},
  booktitle = {ASSL Vol. 240: Numerical Astrophysics},
  year = {1999},
  pages = {19-+},
  adsnote = {Provided by the NASA Astrophysics Data System},
  pdf = {Norman1999.pdf},
  url = {http://adsabs.harvard.edu/cgi-bin/nph-bib_query?bibcode=1999numa.conf...19N&db_ key=AST},
}

@ARTICLE{Shyy1997,
  author = {Shyy, W. and Krishnamurty, V.S.},
  title = {{Compressibility effects in modeling complex turbulent flows}},
  journal = {Progress in Aerospace Sciences},
  year = {1997},
  volume = {33},
  pages = {587-645(59)},
  abstract = {... In the present review, the
     compressibility effect is investigated in the context of engineering models
     needed for complex flow computations, particularly the k-&unknown;
     model. ...},
  doi = {doi:10.1016/S0376-0421(97)00005-5},
  pdf = {Shyy1997.pdf},
  url = {http://www.ingentaconnect.com/content/els/03760421/1997/00000033/00000009/art00005},
}

[brucemiller]

I guess you figured out how to process the bibliography separately, so that you do see the error messages, but I'm wondering why it's so silent here. And it appears to be intentional (the verbosity is set to -5 when transforming the bibliography within MakeBibliography).

@dginev do you have any recollection why we did that?

[dginev]

Great chance for a git blame dance from the interface, the commit that introduced that was here: https://github.com/brucemiller/LaTeXML/commit/eb417ebbb5a60a4caebce0cf5fd3d8a3bfd098e5

And looking at it the memory is resurfacing - I believe I had a case where the bibtex conversion through this call ended up printing out errors and it overall looks worse than the silent pass - as the conversion of the bibtex is done by a separate+dedicated LaTeXML object, and the errors are not merged back into the main conversion error report.

So if I remember this right, I hid the entirety of the bibtex conversion details as a way to maintain some log sanity, as at that point reintegrating the two "status reports" was not obvious.

This comment suggests that I did not consider it fully done either: https://github.com/brucemiller/LaTeXML/commit/eb417ebbb5a60a4caebce0cf5fd3d8a3bfd098e5#diff-8b047080e526df5d43e4a6de69b833c6R159

[dginev]

@asmaier thanks again for the report, the error-reporting is now improved in the latest master branch of latexml.