Re: [ft-devel] ftjam segfault building with GCC 4.2

2007-06-10 Thread David Turner
Hello Ryan and all, I just released a second release candidate for FT-Jam 2.5.3 that contains a fix for these aliasing issues. It's available from: http://david.freetype.org/jam/ be sure to get 2.5.3rc2 packages, not previous versions note that I hope to make it a real release soon, testing

Re: [ft-devel] ftjam segfault building with GCC 4.2

2007-06-10 Thread Ismail Dönmez
On Sunday 10 June 2007 20:28:50 David Turner wrote: Hello Ryan and all, I just released a second release candidate for FT-Jam 2.5.3 that contains a fix for these aliasing issues. It's available from: http://david.freetype.org/jam/ be sure to get 2.5.3rc2 packages, not previous versions

[ft-devel] ftjam segfault building with GCC 4.2

2007-06-09 Thread Ryan Hill
Building ftjam with GCC 4.2 using -O2 optimizations fails with a segfault in jam0. This is caused by a violation of strict-aliasing rules which GCC 4.2 assumes are followed with optimization levels of -O2 and above. It can be worked around with -fno-strict-aliasing. All jam implementations

Re: [ft-devel] ftjam segfault building with GCC 4.2

2007-06-09 Thread Ismail Dönmez
On Saturday 09 June 2007 22:59:54 Ryan Hill wrote: Building ftjam with GCC 4.2 using -O2 optimizations fails with a segfault in jam0. This is caused by a violation of strict-aliasing rules which GCC 4.2 assumes are followed with optimization levels of -O2 and above. It can be worked around

Re: [ft-devel] ftjam segfault building with GCC 4.2

2007-06-09 Thread Ryan Hill
Ismail Dönmez wrote: You should note that gcc 4.2 has an ugly strict aliasing bug in itself [0]. Not sure if the jam case is related though. [0] http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30252 I believe that bug affects only C++. -- dirtyepic salesman said this vacuum's