On Wed, 2005-08-10 at 11:06 +0000, J. Ali Harlow wrote:
> On 10/08/05 04:03:29, Owen Taylor wrote:
>
> > So, if people want to test out what is in CVS now I'd appreciate it,
> > especially for those deviating from the central stream of
> > Linux+recent X.org.
>
> The changes to configure.in seem fine, but the changes to gdkfont-
> win32.c that were checked in at the same time seem to need a little
> more tweaking. I've commited a fix-up for gdkfont-win32.c itself, but
> that still leaves the following which I don't know how to fix:
Always a problem writing code you can't compile at the moment. Thanks
for fixing up.
> Creating library file: .libs/im-ime.dll.a
> .libs/gtkimcontextime.o(.text+0x1ac7): In function
> `gtk_im_context_ime_set_preedit_font':
> /home/avrc/external/gtk/build/mingw32/modules/input/
> gtkimcontextime.c:882: undefined reference to
> `_pango_win32_font_logfont'
> collect2: ld returned 1 exit status
Hmm, that usage of PangoWin32 should still be OK with the cairo changes;
(PangoCairoWin32Font inherits from PangoWin32Font).
while it wouldn't be that hard to get GTK+ building on Windows with
using only FreeType fonts (just change a few checks in Pango's
configure) I don't think anybody would do it.
You could either add pangowin32 back to PANGO_PACCKAGES for win32, or
you could use a separate set of CFLAGS/LDFLAGS for gtkimcontextime.
The first is probably simpler - go back to:
if test "x$gdktarget" = "xwin32"; then
PANGO_PACKAGES="pangowin32 pangocairo"
else
PANGO_PACKAGES="pango pangocairo"
fi
Since pangocairo pulls in pangowin32 on Windows, there is no added
inefficiency, really.
Regards,
Owen
Attachment:
signature.asc
Description: This is a digitally signed message part