On Sat, Oct 10, 2015 at 08:23:03PM +0000, Guenter Milde wrote: > Dear Scott, dear LyX team, > > On 2015-10-10, Scott Kostyshak wrote: > > > 3. What features would you like to still implement for 2.2? Which of > > these features could not be put off for a 2.2.x release (e.g. requires a > > format change)? > > Unfortunately, my knowledge of C++ and the LyX code is too limited to > provide a patch, but: > > Please consider Ticket #9744: "allow parallel configuration of TeX and > non-TeX fonts" > > Try "View>PDF (XeTeX)" with the LyX manuals to see why this is important > (compiling fails, because the combination of XeTeX and TeX-fonts is poorly > supported by both, LaTeX and LyX). > > I propose a new config value "automatic" for "use non-TeX fonts". The > effect of useNonTeXFonts == "automatic" would be to use TeX fonts with > 8-bit engines and non-TeX fonts with Unicode-enabled engines (XeTeX, > LuaTeX).
I agree that this would be nice to fix for 2.2.0. Let's see if we can convince someone to work on it. Here is my attempt: In LyX 2.2.0 one of the changes that could be noticeable to users is that we now report missing characters (#9610). Without this fix, users might not be aware that their compiled PDF will not show some characters. Although some users might be annoyed by seeing errors they did not see before, (1) it is important to give these errors to them so they know about the problem and (2) there is a "Show Output Anyway" button so if they want to view their PDF as they did in 2.1.x, they still can. Fixing #9744 would have the following advantages: 1. It would allow fixing some of the missing glyphs errors. It is not clear how many users will run into this particular issue (see Günter's comment here: http://www.lyx.org/trac/ticket/9610#comment:18). 2. It would fix compilation of the manuals, as well as many of our tests. So many tests are failing now because of missing glyphs that it is difficult to figure out whether there are regressions caught by the tests or not. I am planning to temporarily revert the fix for #9610 when I run the export tests but this makes things more difficult. Scott
