Re: Bug in the (Linux x86) binary-UNO--C++ bridge

Stephan Bergmann wrote:
> Hi all,
> 
> I recently discovered that at least the Linux x86 binary-UNO--C++ bridge 
> (see bridges/source/cpp_uno/gcc3_linux_intel) has a bug and does not 
> conform to the platform ABI.  (Due to how GCC generates code, this bug 
> currently does not do any harm.)  See 
> <http://www.openoffice.org/issues/show_bug.cgi?id=41817> for some more 
> information, and a fix.
> 
> This bug may be present in the bridges for other x86 platforms, too, 
> depending on the ABIs used by those platforms.  But except for 
> gcc3_linux_intel and cc50_solaris_intel (which had a similar problem) I 
> did not make any changes.

Also see <http://www.openoffice.org/issues/show_bug.cgi?id=47459> and 
its fix (bridges/source/cpp_uno/gcc3_linux_intel/call.s 1.2.14.2, 
bridges/source/cpp_uno/gcc3_linux_intel/cpp2uno.cxx 1.7.14.2); might be 
interesting for other platforms, too.

> -Stephan