On Thu, 2004-12-30 at 10:44 +1100, msevior@physics.unimelb.edu.au wrote:
> i Hub and Tomas,
>
> I've had this conversation before and have thought seriously about it. In
> fact there is code already in place for pasting XHTML that we could use
> for awml but I've always decided against it.
>
> The main reason is that there is a hellva lot of code in the RTF imp/exp
> that deals with the all the issues of pasting. Rewriting all that code to
> use awml format will be a lot of work.
I was more thinking about some sort of binary dump of the piecetable.
That would require adding object persistance to the piecetable objects.
On the otherside, I find this distinction of insertStrux / appendStrux
quite insane and bug prone. Why don't we rewrite insertStrux to work in
"transaction mode". Like when you start pasting, you do
beginTransaction(), do everything with insertStrux() that will be much
faster and then endTransaction()
>
> If you find RTF too restrictive for proper handling of bidi content, I
> suggest you extend it with your own '\*\abibidi' set of keywords.
MMmmmm. I'm not sure that this fix all the issue. Look at how long they
last.
Converting to RTF can be really consuming. Seriously. We don't all have
4GHz machine to run AbiWord.
> This has worked well for me and has enabled copy and pasting of lists and
> tables where what gets inserted depends on the context of the current
> document position.
This has brought largely unecessary bloat to RTF importer / exporter
that could have been avoided.
Hub
-- Crazy French - http://www.figuiere.net/hub/Received on Thu Dec 30 01:02:04 2004
This archive was generated by hypermail 2.1.8 : Thu Dec 30 2004 - 01:02:04 CET