On Jun 24, 2014, at 12:13, René J.V. Bertin <rjvbertin@gmail.com> wrote:

On Jun 24, 2014, at 21:00, xquartz-dev-request@lists.macosforge.org wrote:

...

Send Xquartz-dev mailing list submissions to

...

I'm not sure why you think rendering quality is "considerably less good under XQuartz than it is on "stock" linux" ... do you have examples? We're using the sam freetype in XQuartz that Linux is using. Is there something I'm missing?

You're using the same freetype,

freetype-2.5.3 plus a few patches (inline at the bottom; I omitted the build system patches and just included the functional onces)

but built with the same options,

We just have --without-png. If there's an --enable-better that Linux distros are using, please let me know ;)

and using the same runtime configuration?

Obviously not the last one. Linux and darwin have different runtimes, but freetype shouldn't be caring about those differences. darwin Libc should behave similarly to glibc.

It's hard to give examples of something that's subjective to a large extent. Let's put it this way, I'm not the only one convinced that the result is largely worth the effort of installing a freetype with the patches.

Yeah. I'm not really a fonts guy. If someone points out "hey, enable this option in freetype as it makes it better", I'll do it (or at least try it out in a beta). --- Our freetype patches: --- a/modules.cfg.orig 2007-01-31 14:51:06.000000000 +0900 +++ b/modules.cfg 2007-03-31 18:59:22.000000000 +0900 @@ -124,7 +124,7 @@ # OpenType table validation. Needs ftotval.c below. # -# AUX_MODULES += otvalid +AUX_MODULES += otvalid # Auxiliary PostScript driver component to share common code. # diff -Naurp freetype-2.5.0.1.orig/src/base/ftrfork.c freetype-2.5.0.1/src/base/ftrfork.c --- freetype-2.5.0.1.orig/src/base/ftrfork.c 2013-06-05 14:56:56.000000000 -0700 +++ freetype-2.5.0.1/src/base/ftrfork.c 2013-06-20 22:26:21.000000000 -0700 @@ -504,29 +504,7 @@ /* Only meaningful on systems with hfs+ drivers (or Macs). */ - FT_Error error; - char* newpath = NULL; - FT_Memory memory; - FT_Long base_file_len = (FT_Long)ft_strlen( base_file_name ); - - FT_UNUSED( stream ); - - - memory = library->memory; - - if ( base_file_len + 6 > FT_INT_MAX ) - return FT_THROW( Array_Too_Large ); - - if ( FT_ALLOC( newpath, base_file_len + 6 ) ) - return error; - - FT_MEM_COPY( newpath, base_file_name, base_file_len ); - FT_MEM_COPY( newpath + base_file_len, "/rsrc", 6 ); - - *result_file_name = newpath; - *result_offset = 0; - - return FT_Err_Ok; + return raccess_guess_darwin_newvfs( library, stream, base_file_name, result_file_name, result_offset ); }