On Sunday 02 February 2003 11:49 am, Lars Gullik Bjønnes wrote: > Angus Leeming <[EMAIL PROTECTED]> writes: > | On Sunday 02 February 2003 1:41 am, Lars Gullik Bjønnes wrote: > | > Angus Leeming <[EMAIL PROTECTED]> writes: > | > | I've reopened the bug and put the patch and test case there but still > | > | think that this should be applied to 1.3. > | > > | > Then test test test... you still have some days before 1.3.0 go out > | > the door. > | > > | > We should create some testsuite kind of thing for this stuff. > | > | I'd quite like to learn enough perl to say I don't like it. Seems to me > | that a reasonable first step is to upgrade reLyX so that it outputs stuff > | in the current format. Thereafter, adding support for missing stuff like > | minipages has at least a chance of being coherent. > | > | Roll on André's TeX parser. > > Perhaps Boost.Spirit might be a way to go? > > Spirit is a parser framework written in C++.
Interesting. I envisage the following step-by-step framework for our handling of LaTeX: 1. Do not add any new functionality to reLyX but get it to output stuff in the current lyxformat not lyxformat 215. 2. Add support to reLyX for stuff that LyX can now handle natively. Things like minipages. At this point reLyX is fully working again. 3. reLyX has already been modified to pass math unchanged because André's math parser can make sense of this subset of LaTeX. In other words, reLyX handling of math has been 'retired'. 4. As André's insets take over the world, then more and more LaTeX can be passed through reLyX unchanged because his parser will be able to deal with it. Gradually reLyX shrinks away to nothing as the 'math parser' evolves into a 'LaTeX parser' 5. If we want to re-design the math parser using Boost::Spirit then that fits into this framework too. Angus
