On Tue, Apr 12, 2005 at 11:05:52PM +0300, Martin Vermeer wrote: > On Tue, Apr 12, 2005 at 10:46:35AM +0200, Alfredo Braunstein wrote: > > Martin Vermeer wrote: > > > > > BTW is the code added in this patch on the critical path for timing? I > > > remember seeing people discussing the lyxfont copy c'tor. > > > > Yes, it may well be that adding a couple of lyxfont constructions on char > > drawing will have some impact on speed. But don't worry about that for the > > moment, we can handle that latter. > > > > Unfortunately I'm almost completely clueless wrt. lyx font handling, so I > > cannot help you much. I can try to do some testing this weekend though. > > > > Regards, Alfredo > > > > It's in CVS. > > - Martin
I finally found also a way to fix the language "turd" behaviour in new lyxtexts in non-english documents. The patch has a bit of brute force over it and I am not sure that setCurrentFont was meant to be used in this way. But it works. I suspect that real_current_font, which is used to determine whether to display the foreign-language underline and the L-shaped cursor, is not maintained as strictly as it should. Anyway for now, for 1.4, this should be good enough (isn't current_font and real_current_font supposed to go?). - Martin
Index: text3.C
===================================================================
RCS file: /usr/local/lyx/cvsroot/lyx-devel/src/text3.C,v
retrieving revision 1.286
diff -u -p -r1.286 text3.C
--- text3.C 7 Mar 2005 11:03:45 -0000 1.286
+++ text3.C 13 Apr 2005 05:15:39 -0000
@@ -1124,6 +1124,9 @@ void LyXText::dispatch(LCursor & cur, Fu
cur.clearSelection();
LyXFont const old_font = real_current_font;
+
+ // Prevents language turds in new lyxtexts under non-english
+ setCurrentFont(cur);
string::const_iterator cit = cmd.argument.begin();
string::const_iterator end = cmd.argument.end();
pgpkbTVnK4j5b.pgp
Description: PGP signature
