From: Karl K. <ko...@or...> - 2002-02-07 19:41:07
|
Hi , One of the new options added is -G cjk, which adds right-to-left, to-to bottom text. This leads to some wishes for the wishlist: a) scrollback should work left-to-right, not bottom-to-top. b) arabic from bottom-to-top looks strange, especially as it is=20 not joined. One could think of doing bidi the following way: + right-to-left: rotate 90deg counterclockwise + left-to-right:=A0maybe rotate 90deg clockwise In attempting b), I have a question: In ml_font.c, the font-name is created with: =20 sprintf( fontname , "-*-%s-%s-%s-%s--%d-*-%s-*-%s" , family , weight , slant , width , fontsize , spacing , encoding) = ; This specifies pointsize to be '*', and leaves out xres and yres values. Is there a reason to not do a full specification, that is: sprintf( fontname , "-*-%s-%s-%s-%s--%d-*-*-*-%s-*-%s" , family , weight , slant , width , fontsize , spacing , encoding) = ; ( the three stars being pointsize, xres,yres ). Only then, rotation seems to work in a reliable way: 90deg becomes sprintf( fontname , "-*-%s-%s-%s-%s--[0 %d ~%d 0]-*-*-*-%s-*-%s" , =09 family , weight , slant , width , fontsize ,=20 =09 fontsize , spacing , encoding) ; Of course, one has to alloca more memory and _not_ rotate asian glyphs .. = =20 As a side-note, those X11R5-servers need an own solution: best they die out ( but there still are some here ).=20 =09Karl |