Re: [Devel] 2.1.5rc1 and bytecode interpreter

[James H. Cloos Jr.]

>>>>> "David" == Turner David <address@hidden> writes:

JimC> Am I missing something obvious?

David> in previous releases, the line in "ftoption.h"
David> that read #undef TT_CONFIG_OPTION_BYTECODE_INTERPRETER was
David> simply commented off, and thus never read by the pre-processor.

David> In the current sources, this line will simply disable any
David> command-line settings you use when compiling the sources.

Blind, I say, BLIND I am.  :-/

I looked at that part of the diff from 2.1.4 to 2.1.5rc1 several
times w/o noticing that.  (OTOH, it wasn't a simple - one line
and + on the next, and those long runs of UPPERCASE do tend to
all look the same after a while.  But I should have picked up
on that anyway.)

David> I didn't envision that someone would force bytecode interpreter
David> compilation through additional CFLAGS. The fix is simply to
David> comment out the offending line as before.

I beleive the author of the ebuild got the idea from either debian or
one of the rpm dists, so I presume it is now the standard way everyone
is turning it on.  Saves having to keep track of a patch file....

David> Any volunteer ? I'm at work. (and don't forget to update
David> BRANCH-2-1-5 _and_ HEAD branches on the CVS)

I don't have commit access, but here is the diff....

-JimC

diff -uNr a/include/freetype/config/ftoption.h 
b/include/freetype/config/ftoption.h
--- a/include/freetype/config/ftoption.h        2003-08-17 19:03:36 -0400
+++ b/include/freetype/config/ftoption.h        2003-09-02 15:57:34 -0400
@@ -399,7 +399,7 @@
   /*   Do not #undef this macro here, since the build system might         */
   /*   define it for certain configurations only.                          */
   /*                                                                       */
-#undef  TT_CONFIG_OPTION_BYTECODE_INTERPRETER
+/* #define  TT_CONFIG_OPTION_BYTECODE_INTERPRETER */
 
 
   /*************************************************************************/