On Mon, Oct 20, 2003 at 10:06:58PM +0000, Angus Leeming wrote:
> I'm a bit pissed off with the way we handle fonts and strings.
Good starting point.
> It seems to me that we are duplicating code in the inset metrics and
> draw functions. Only in the very inner block of code do things
> differ, one ascertaining the width of the string in this particular
> font and the other calling the painter with this info.
>
> As an example, I append the code from render_graphic.C to print out a
> diagnostic message.
>
> It seems to me that this mess could be avoided if we had some
> container that could contain both the string and the font information.
> Indeed, the STL stream classes do something very similar.
>
> I can imagine code like:
>
> void RenderGraphic::metrics(MetricsInfo & mi, Dimension & dim) const
> {
> if (!image_ready)
> lyx::rendered_stringstream rss;
> create_message(rss);
> dim.wid = std::max(50, rss.width() + 15);
> }
> }
>
> RenderGraphic::draw(PainterInfo & pi, int x, int y) const
> {
> if (!image_ready)
> lyx::rendered_stringstream rss;
> create_message(rss);
> pi.pain.text(x + InsetOld::TEXT_TO_INSET_OFFSET + 6, y - 4,
> rss);
> }
> }
>
> where the magical helper function contained code like:
>
> void RenderGraphic::create_message(lyx::rendered_stringstream & rss)
> {
> rss << setFont(mi.base.font)
> << setFamily(LyXFont::SANS_FAMILY);
>
> string const justname = OnlyFilename(params_.filename);
> if (!justname.empty()) {
> rss << setSize(LyXFont::SIZE_FOOTNOTE)
> << justname;
> }
>
> string const msg = statusMessage(loader_.status());
> if (!msg.empty()) {
> rss << setSize(LyXFont::SIZE_TINY)
> << msg;
> }
> }
>
> Is this just a wild flight of fancy, or does anybody else think that
> it's a sensible proposal?
I think this could be handled on a case-by-case base by using some
helper functions in the insets that really need it.
Most insets are simple enough not to need such a thing, so any generic
solution is like to complicate matters.
However, we certainly _have_ a font problem. I'll ask a question in the
next mail ;-} IMNSHO the optimal solution would be "font insets" but
we'd need Asger's three-box model for insets implemented to go this
route...
> ps, current code follows...
That's indeed a case that could make use a local helper function.
But I don't think this very helper would be useful for, say,
MathSqrtInset.
Andre'
