Martin Vermeer wrote:
> The patch that I proposed restricts going "Wide" to insets having a
> minimum total text height corresponding to more than two rows ("Tall").
> This appears to work well in my testing. It is still possible to
> construct pathological cases, but ordinary typing even inside a nested
> inset will not normally make the inner inset extend on top of buttons or
> outer text.
Yes, it works very well for me, too. I still could provoke wide insets, but
in most cases this does not happen anymore. Nice!
> I attach this patch, to which I have now added the inset methods that I
> discussed about with Georg, and the names you proposed (or at least one
> of them).
For the other one I think it would be better to use
virtual InsetText * asInsetText() { return 0; }
insetad of isTextInsetBased. That is symmetric to
virtual MathInset * asMathInset() { return 0; }
, and you don't need so many casts (if you want to know whether an inset is
derived from InsetText you normally want to do something with it)
> And ah yes, I used Angus's script to generate the raw changelog entry,
> requiring only modest manual post-editing. I think we're on to something
> here if we are going to switch from changelog to svn log.
Yes, it is very nice. I use it already.
Georg
