I have a LaTeX paper (multifile setup, using \input{}
statements) that , since a few days ago, keeps telling me at each compile (both with pdflatex
or latex
):
LaTeX Warning: Label(s) may have changed. Rerun to get cross-references right
Even recompiling, say, ten times does not fix it. All references in the paper seems fine, though (can't be 100% sure, would need to check each and every \ref
in the source with the PDF output).
It's probably because I misplaced a label or reference in some way. But how to find it, and fix the issue? What sort of misplacements result in this repeated message?
Edit: I've tracked it down to an issue with the acronym
package in combination with using \ac{}
in the abstract definition of an elsarticle (Elsevier's latex class, I am using the latest version, i.e. 1.20b from 2009-09-17) document. Below is a minimal copy/paste example where the issue appears:
\documentclass[final,5p, times,twocolumn]{elsarticle}
%\documentclass{article} --> no problem when using this one (and removing frontmatter)
\usepackage[nolist]{acronym}
\begin{document}
\begin{frontmatter}
\begin{abstract}
This is just a sentence with an acronym in it: \ac{ICT}.
\end{abstract}
\end{frontmatter}
\acrodef{ICT}{Information and Communication Technologies}
\end{document}
It is not clear to me if this is because of a bug in the acronym package, or elsarticle.
\ref{}
in things like margin notes. At every recompile the note jumped between two pages, always leaving changed references... So the answer is "yes", you must have done something. It is nearly impossible to tell you what you did wrong. My advice is to do the classic: Comment some inputs, does it go away? Can you find the one input for which the warning appears? or are those multiple inputs? once you found the input, comment parts of that file to pin point where the problem occurs..aux
file to a different name run latex again and compare the difference between the old and new aux files.\ac{}
(from the acronym package) in my abstract. Removing them makes it work (which is OK for me, as I did an\acresetall
after the abstract anyway). Usage of acronyms in the other parts of the paper doesn't give this problem, weird. @David: both.aux
fields are identical, also weird