Re: /srv/bzr/emacs/trunk r101338:*lisp/emacs-lisp/syntax.el(syntax-ppss): More sanity check to catchtch

[E Sabof <esabof@gmail.com>]

Dmitry Gutov <dgutov@yandex.ru> writes:

> Until you define what such a function would do, it's hard for things to
> go wrong.

I should probably get re-acquainted with mmm. It might contain useful functionality that I haven't considered.

From personal experience of dealing with multi-lingual files, I'd say I'd like fontification, language-sensitive navigation and indentation. I suppose some keys might change as well, but I'not convinced that the benefits outweigh the additional complexity.

It might be possible to achieve consistent navigation by making functions such as beginning-of-defun consult a region property before consulting the buffer-local beginning-of-defun-function.

I might also like error-checking, but it's conceivable to move this responsibility to flymake.

>
> But let's just look at things in motion. What would make a major mode
> call this function? It would have to know about a set of template
> delimiters, right? And somehow translate them into `set-region-mode'
> calls?
>
> Are BEGINNING and END buffer positions? What happens if the user inserts
> some text in the buffer? Will all regions below suddenly become invalid?
>
> Okay, let's make them markers. What happens when the user removes a
> region? Or just one delimiter? Or adds a new delimiter, or region?

This requires a parsing loop, which I'd guess is how mmm handles it.

Evgeni