Am Mittwoch, 1. März 2017 um 11:35:45, schrieb Guenter Milde <mi...@users.sf.net> > On 2017-03-01, Scott Kostyshak wrote: > > On Tue, Feb 28, 2017 at 06:24:51PM -0500, Scott Kostyshak wrote: > >> On Tue, Feb 28, 2017 at 02:07:38PM -0500, Scott Kostyshak wrote: > > >> > Note that this affects lib/examples/es/linguistics.lyx > >> > > >> > Can anyone else reproduce? Is the problem obvious? > >> > > >> > I can reproduce the issue on 2.2.0 and on master. I cannot reproduce on > >> > 2.1.0. I can do a git bisect if someone thinks the time would be worth > >> > it. > > >> Perhaps this is http://www.lyx.org/trac/ticket/9871 ? > >> I inverted the affected test. > > >> But that issue is not labeled as a regression. Why does the MWE work > >> with 2.1.0? > > > The commit that changed the behavior was 36b9645b. > > > Günter, can you confirm that the export just worked by chance (perhaps > > even bad output?) before, so that 36b9645b did not really break the > > export? I have a vague memory that this is the case, but I want to > > double-check. > > 36b9645b is part of the fix to #9740 (XeTeX with TeX fonts). > (BTW: how do you get the content/description of a change-set given this > strange number only?)
# git show 36b9645b > Possible regressions due to the restrictions of XeTeX with TeX-fonts are > expected (and should be eventually handled) but don't outweigh the > benefits of the change. No reason to restore 2.1 behaviour. > > See also the similar discussion in https://www.lyx.org/trac/ticket/10199 > > Could you try whether export works with pdftex if the input encoding is set > to ASCII? Does not work here. Even lualatex cannot compile if encoding is set to ascii. Setting to utf8 works with pdflatex and lualatex but not with xelatex. > Günter Kornel
signature.asc
Description: This is a digitally signed message part.