[Vim-latex-devel] Re: [Vim-latex-cvs] vimfiles/plugin imaps.vim,1.7,1.8

[Luc H. <her...@fr...>]

* On Sun, Dec 08, 2002 at 06:54:47PM -0800, Srinath Avadhanula <sr...@fa...> wrote:
> Therefore we could conceivably have ftplugin/c.vim containing lines
> like:
> let b:Imap_PlaceHolderStart = '?'
> let b:Imap_PlaceHolderEnd = '?'
> call IMAP('for', "for (??; ??; ??) {\n??\n}??", 'c')

Indeed, but you loose flexibility in that case. The day the new language
C% arise, it would not be able to take advantage of the C mapping for
'for' as it defines a new operator : '??' 
;-)


> The problem you mention is when some new language has legitimate
> characters such as '<+' and Imap_PutTextWithMovement() tries to replace
> those with the placeholder characters. 

That's my point: defining a highly evolutive and language/configuration
independent plugin.

> That problem can be solved by using a new function
> Tex_PutTextWithMovement() instead of Imap_PutTextWithMovement() which
> first replaces things like « (or '<+' or '!mark!') with the user's
> choice and then calls Imap_PutTextWithMovement(). This makes

That's an option.

> > It is not really problematic !  I will use '!mark!' for instance
> > (instead of '¡mark!' currently used) to express the markers (aka
> > placeholders)
> 
> Well, on further thinking, it does make sense to use longer things
> like '!mark!' or something similar and then replace with the user's
> placeholder choice in `Imap_PutTextWithMovement'()... 

> But this will mean again that the scripts themselves become really
> long and unreadable...

Not always. Check again the mappings ']em' and ']ee' I proposed you.
The mecanics that permits such writings is very complex [1], but once
written, the end-coders that write and maintain TeX (/HTML/...)
code-snippets have a very easy job left.

However, I must admit that my mappings for C&C++ constructs and common
brackets are quite complex. [I think that today I should be able to
simplify them as my VimL skills have improved.]

> Well, there seems to be fair arguments for and against the using of
> things like !mark!. But I would think that a solution where imaps is
> completeley independent of the choice of placeholder characters is best.

['!mark!' is not a placeholder -- actually, this is a mapping with my
settings ; it is replaced by maparg('!mark!', i')]

And I definitively agree, functions like IMAPS() should be independent
of any placeholder characters. Somehow, '<+' and '+>' (as well as '«' &
'»' today) look like placeholders characters, and more important: they
look like things that some languages may use -- like for instance '«'
and '»' in LaTeX documents written in French.

> And I think each language choosing its own placeholders is most
> elegant and robust...
 
100% agree.


[1] but it does not handle insert-mode mappings only ... command-,
insert- and normal-mode mappings _and_ menus are handled. So, I think it
is a fair price.
-- 
Luc Hermitte