On Fri, Dec 02, 2005 at 03:42:50PM +0100, Jean-Marc Lasgouttes wrote:
> >>>>> "Martin" == Martin Vermeer <[EMAIL PROTECTED]> writes:
> 
> Martin> From within an inset, I can confirm that indeed the default
> Martin> font settings are returned when you change layout to Section.
> Martin> When in the main text, ALL_INHERIT is returned as should.
> 
> Martin> Looking at the code in Paragraph::getFontSettings, these
> Martin> default settings seem to come from the font list structure.
> Martin> But how did they get there in the first place? And why only
> Martin> inside an inset?
> 
> Ha! Been there, done that :)
> 
> The result is the following one-liner patch to text2.C (the patch to
> text.C does nothing except remove code that is indimidating but does
> nothing).
> 
> The problem is that the cache LyXText::current_font, which is (if I
> understand correctly) the uninstantiated default font of the LyXText,
> was not initialized, and thus received the default ALL_SANE (roman,
> medium...) setting, instead of a good ALL_INHERIT.
> 
> The situation now is much more credible, but touching such core code
> make me nervous. Could you test that (without your other patch)?
> 
> I think this current_font/real_current_font should eventually go away.
> Also, it would be nice if defaultfont_ was either removed of
> initalized to something somewhere :)
>

Do you know a way to similarly initialize current language? That would
fix a bug properly that I sort-of fixed a long time ago in an ugly way.
(And the fix doesn't work; just make your doc language different from
the default, and delete all chars in your doc, and start writing again).

The patch works fine BTW in what little testing I did.

- Martin

Attachment: pgpm8KebZHKhP.pgp
Description: PGP signature

Reply via email to