On Sat, 18 Sep 2004 10:19:56 +0100, Timwi timwi@gmx.net wrote:
Things I know I haven't done yet:
- HTML tags (currently everything between <something> and </something> is treated as an 'extension'; I need to limit that to <nowiki>, <pre>, <math>, <hiero>, <music> and <chem> for now)
- [http://url/ These sort of links]
- hrs (horizontal rules)
- the new -{ language variants }- syntax
If there is absolutely anything else missing, please let me know!
Have you worked out how to deal with "MagicWord" i18n yet? I ask because in the current not-a-parser, there is no definition of what a magic word looks like, just a class that can be asked "is this a one of those". Probably with a properly defined grammar, we would have to limit the style of magic-words, so that just likely tokens could be checked against the current list of magicwords. [I don't actually know if any translations make use of this feature, but it would be a shame to lose it.] I think all default magic words currently fall into one of: "#"<word> (as in "#REDIRECT") "__"<word>"__" (as in "__NOTOC__" et al) "<"<word>">" and matching "</"<word>">" (for extensions, and whitelisted HTML tags)
You may already have thought this through and come to your own conclusions, but this approach certainly seems more efficient than having to check *every* token against a run-time list.