Helge Hafting wrote:

> Well, but without changing the name, the format will only be extended,
> not changed.  So no need to convert old files then, only conversion
> to ERT when opening a new file with an older lyx.

You are right, but in our terminology this is a format change, too. We have
serveral of these with an empty convert() function in lyx2lyx.

> I was thinking of isValidGlueLength, but I see this too is independent
> of vspace.  I misunderstood as it was implemented in the vspace.C file.

This is indeed not very nice. Should be cleanep up one day.

>> The really interesting thing is to extend LyXLength to handle user
>> defined lengths. How would that be done? Having a central place where the
>> user could define his own lengths, or simply allowing an arbitray unit
>> for user defined lengths?
> 
> A good question indeed.  I was planning on allowing anything starting
> with a backslash, so the user can use things like \medskipamount and
> all the other lengths latex (and various packages) defines for us.

This would be good for a start. The only problem: How do we represent such a
length on screen?

> But of course this makes latex compile errors possible too. If that
> isn't good enough, I'll drop this part for now.  It can always be
> added later. Document settings could certainly have a page for user
> defined lengths, and .layout files could be extended to provide
> "allowed" lengths provided by latex or the packages the .layout
> brings in. But that is material for a later patch.  It is something
> I'm not missing right now.  User defined lengths don't seem that
> necessary when we can write stuff like -2.45pt.

I disagree. Example: I have a complicated table in our lecture notes. When I
want to include it in a beamer presentation I have to change several (~20)
lengths. They cannot be described with the help of text% etc. If LyX would
accept user defined lengths for column widths and in minipages I would only
need to redefine one length variable, because all these lengths can be
described as a multiple of it.

Of course this has not much to do with the hspace changes, it is a more
general thing.

> Well, I am not so sure how useful it is to change a "hfill" into
> something else, be it a fixed-length space or a different fill.
> One can always just delete and insert the other kind.  Changing
> from dotfill to rulefill is something I can imagine doing to
> fine-tune the look though.  Or changing a quad to enskip.  But a fill
> doesn't seem so likely to be replaced with a non-fill.

Indeed.

> Anyway, it can be done later.  Actually getting a hspace dialog
> is a step forward on its own.

Yes. So maybe you start with just that?


Georg

Reply via email to