Huw D M Davies wrote:
TransGaming Add a Marlett replacement font
Cool! Is there any policy for using it? E.g. a lot of the arrows and stuff in the native comctl32 are drawn using this font, so would we now be able to do this with our version of the common controls?
Rob
Le jeu 16/09/2004 à 15:58, Robert Shearman a écrit :
Huw D M Davies wrote:
TransGaming Add a Marlett replacement font
Cool! Is there any policy for using it? E.g. a lot of the arrows and stuff in the native comctl32 are drawn using this font, so would we now be able to do this with our version of the common controls?
Far from everyone has fontforge installed, so unless there's a compiled version in cvs (a la configure) you might find most users don't have marlett.ttf.
Vincent
Far from everyone has fontforge installed, so unless there's a compiled version in cvs (a la configure) you might find most users don't have marlett.ttf.
Indeed, if large numbers of people aren't getting the fonts at all that's a pretty strong argument for having the binaries in CVS (or making fontforge a hard requirement of the build which would be a pain)
On Thu, Sep 16, 2004 at 04:08:53PM -0400, Vincent Béron wrote:
Le jeu 16/09/2004 à 15:58, Robert Shearman a écrit :
Is there any policy for using it? E.g. a lot of the arrows and stuff in the native comctl32 are drawn using this font, so would we now be able to do this with our version of the common controls?
Far from everyone has fontforge installed, so unless there's a compiled version in cvs (a la configure) you might find most users don't have marlett.ttf.
We can test for the presence of Marlett and fall back to the existing code if we don't find it.
Huw.
We can test for the presence of Marlett and fall back to the existing code if we don't find it.
That works but then we're maintaining two codepaths, one or other of which will hardly be tested at all. Why not just check in the font binaries? I know CVS isn't very good at it, but it does work and it'll probably save pain later.
Mike Hearn m.hearn@signal.qinetiq.com writes:
That works but then we're maintaining two codepaths, one or other of which will hardly be tested at all. Why not just check in the font binaries? I know CVS isn't very good at it, but it does work and it'll probably save pain later.
First because binaries are a pain with CVS, second and more important because there has to be a very good reason for putting any generated file in CVS; configure qualifies, but fonts don't. The worst case is that the window controls won't look exactly right, that's not a big deal, and it can be solved by having a binary font package somewhere for download.