rgheck <[EMAIL PROTECTED]> writes: > There is a getLayout(BufferParams const &), but the one I was using is > getLayout(), which is only defined in InsetCollapsable. I could use > the other one, but there's a lot of weirdness here, now that I look at > it. I take it that Inset::getLayout(BufferParams const &) is supposed > to return the layout for that inset. But what it actually does is call > DocumentClass::insetLayout(docstring), which returns the layout with a > certain name. So, e.g., a call to > InsetCollapsable::getLayout(BufferParams const &) would return the > InsetLayout named "Collapsable"---or would, if it weren't overridden. > So this is just wrong, and we don't run into a problem because it's > only used with Collapsables.
But name() is virtual, isn't it? I think this is something we should keep and get working. > I propose we get rid of Inset::getLayout() altogether. It's not > needed. As the attached shows. The potential usefulness of getLayout for basic insets is to be able to define the fonts used by all insets (ref, cite, math) from the layout file. >>> lf.realize(buffer.params().getFont().fontInfo()); >>> >> >> I do not know why the family is singled out like that. >> >> > Should we wait and ask him? Probably. JMarc
