> On Aug 13, 2015, at 2:31 PM, Steve Thompson <scth...@gmail.com> wrote: > > Hal, > I think I did understand. You want your collaborator to be able to make > arbitrary changes to the .tex file, then have tex2lyx figure out how to > create a new .lyx file. But your collaborator could have done anything. That > makes the “modified” file no different from any other completely arbitrary > new input file. > Steve > > > > Steve - > > I do not think you understood me. Suppose a create a LyX file and use it to > generate a tex file. I send it to my coauthor who does not use LyX. He > makes modifications to the LaTex file without changing the front material, > and sends me his Tex file. I should be able to use tex2lyx to make a revised > version of my LyX file. But this requires many corrections by hand. > > Hal
Steve - First of all, tex2lyx already comes close to making a good LyX file. The main problem is that there always seem to be a handful or errors that must be fixed before the file will run. For some reason the program cannot handle options like \begin{thm}[Main Lemma], but also other options. It also seems that comments in the Latex file can sometimes produce errors. A huge annoyance is that theorems, etc. are displayed in ERT. I understand that for arbitrary LaTex files some of these issues maybe very complicated or impossible, but what I am asking for does not involve arbitrary files. Let’s take an easy case. Suppose my coauthor only makes changes between \begin{document} and \end{document} of the LaTex file, and does not use any new constructs. Shouldn’t it be possible, perhaps with a helper file that records settings and any other special issues, to make a good LyX version? Or even easier, shouldn’t I be able to export a Latex file and then without changing it, import it and get the same LyX file back. Yes, I know there is no reason to do this, but it is a benchmark. Hal