On Sun, Oct 08, 2017 at 02:41:03PM -0400, Scott Kostyshak wrote: > On Sun, Oct 08, 2017 at 12:48:36PM +0000, Enrico Forestieri wrote: > > > Thanks Scott. I'll have a look when I find some time. This kind of bugs > > can be very time consuming. > > Thanks, there is no rush and I don't consider this a blocker for 2.3.0.
Please, try the attached patch. This seems to be an issue related to argument insets only. However, I think the patch should be thoroughly tested to ensure that it introduces no regression. -- Enrico
diff --git a/src/output_latex.cpp b/src/output_latex.cpp index e0163ad2fa..02667aae9f 100644 --- a/src/output_latex.cpp +++ b/src/output_latex.cpp @@ -882,7 +882,7 @@ void TeXOnePar(Buffer const & buf, } // With CJK, the CJK tag has to be closed first (see below) if (runparams.encoding->package() != Encoding::CJK - && par_lang != openLanguageName(state) + && (par_lang != openLanguageName(state) || localswitch) && !par_lang.empty()) { string bc = use_polyglossia ? getPolyglossiaBegin(lang_begin_command, par_lang, par_language->polyglossiaOpts())