Christian Ridderström wrote:
> Speaking of that, I used my script to see how the lfuns have changed
> between the recent releases. The lfuns for Lyx 1.3.1 - 1.3.3 are the
> same, but the n:o lfuns has dropped from 241 to 238 in the latest
> CVS. However, the n:o lfuns in
>
> 1.3.3 is... 241
> HEAD is... 238
> both 1.3.3 and HEAD is... 206
Which indicates two things:
1. Some lfuns are becoming more powerful. You can do more from
outside lyx 1.4.x than you could from outside lyx 1.3.x with fewer
lfuns.
2. We make no effort to translate old-style lfuns to new style lfuns.
I have been thinking for a while of an lfun2lfun script. As an
example:
citation-insert -> dialog-show-new-inset citation
But perhaps this is only part of a wider scripting capability. It is
possible to define the properties of an inset _exactly_ from outside.
We use a snippet of the lyx format to pass the information. For
example
inset-apply citation LatexCommand \citeN[see]{leeming2003}
an aside: I should strip out the part handled by the buffer (as
oppossed to the inset), "LatexCommand" here...
The point I'd like to make is that this lfun syntax is powerful but
cumbersome. I can imagine an lfun2lfun script that generated such
commands from simpler input by the user.
That way, LyX will have to handle fewer lfuns and the user will still
have his syntactic sugar.
--
Angus
