Hi Joseph Thanks for taking your time to look at this. I can see now that what I need to fix but still I would say that LuaTeX breaks TeX's backward compatibility.
On Sun, May 15, 2016 at 6:02 PM, Joseph Wright < [email protected]> wrote: > On 15/05/2016 08:37, Vafa Khalighi wrote: > > Hi > > > > I have been a lollipop user (http://ctan.org/pkg/lollipop) since 15 > years > > ago and its developer from April 2014. For the past 15 years, I have only > > been using Knuth's TeX engine since that gives me all I want. If you try > to > > build the lollipop format for Knuth's TeX by: > > > > tex -ini -jobname=lollipop -progname=lollipop lollipop.ini > > > > > > and process the following minimal example with lollipop format: > > > > > > \Start > > > > Let's say $x$ is a variable > > > > \Stop > > > > You see no problem in the output. However, if you build the format by > > using luatex v0.95.0 by > > > > > > luatex -ini -jobname=lollipop -progname=lollipop lollipop.ini > > > > > > and process the same file, then the x in the inline math mode > > disappears. Should not luatex at least behave like Knuth's TeX engine? > > Does anyone know what is wrong here? > > > > The content of my lollipop.ini is > > > > \scrollmode > > \input lollipop > > \dump > > \endinput > > > > > > Vafa > > > > There've been some changes in the internals of handling \Umathcode, > which at least in LaTeX terms requires some updates to parts of the > codebase. What I notice here is that the \Umathcode for 'x' is certainly > wrong, and doing > > % Set up extra primitives > \ifx\directlua\undefined\else > \directlua{tex.enableprimitives("",tex.extraprimitives())}% > \fi > % Direct PDF output for convenience > \ifx\pdfouput\undefined > \outputmode=1 % > \else > \pdfoutput=1 % > \fi > > \Start > > \Umathcodenum`x=31457400 % > > %\showtokens\expandafter{\the\Umathcode`\x }% > Let's say $x$ is a variable > > \Stop > > works. > > Joseph >
