On Mon, May 31, 2004 at 01:18:01AM +0300, [EMAIL PROTECTED] wrote: > Oh, that makes sense. > > Of course if someone really has patience, understanding, and maybe a > functional-programming-bent, I'd guess a style could be designed that > overrides \newcommand and other macro definition facilities so that they > create lists of available macros, instead of actually defining them, and > run this in the background to create a list of available macros which > could then be used for the completion.
Not just that, it also requires a more-than-healthy addiction to masochism. > Again, its probably true that a very approximate and simplistic solution > would probably provide 90% of the benefit, but I wouldn't claim that > that would be simple to do, either. Examples of probably reasonable > approximations: > - only pick up some of the definition types > - analyze styles in batch, and present lists based on those > - ignore completely the cancellation/redefinition of macros, and simply > show anything that was ever defined (autocompletion doesn't have to > guarantee results that are correct at all levels) This might be do-able. However, unless a volunteer steps up I don't see it coming into existence - even if quite a bit of the necessary infrastructure is already in place thanks to the math preview facility, > I can even almost imagine a regular expression based hack providing > significant benefits... ... but certainly not with regular expressions if you aim for "90+%" "reliability". One on the main problems of the TeX language is that it is too flexible to be predictable, and the outstanding major problem is that even standard .sty files use too much of the fancy language gimmicks to be easily parsable for any program different from TeX itself. Andre' PS: Please don't top-post-bottom-quote on this list.
