On Mon, 2 May 2011, Rainer Orth wrote:

> * Another wart was that gcc/ada/gcc-interface/Makefile.in requires
>   TARGET_LIBGCC2_CFLAGS, which is otherwise only used in libgcc.  It
>   often contains -fPIC, and if the variable is removed, libgnat.so won't
>   link (text relocations remain).
> 
>   This is probably another ada PR?

It would be good for the Ada target library build to rely only on the 
flags in the toplevel libada directory (and if necessary on flags in the 
toplevel libgcc directory if there is something that genuinely should be 
shared), though I don't know how hard that is to fix.  (In general more of 
the logic and source code relating to building Ada files for the target 
should go in the toplevel libada directory.  But there are quite a lot of 
files built for both the host and the target which complicates things, and 
those contribute a few target macros to the list in 
<http://gcc.gnu.org/wiki/Top-Level_Libgcc_Migration> of those still used 
in target code.)

As a related issue, there are too many macros used to indicate that code 
is being built for the target (USED_FOR_TARGET, IN_TARGET_LIBS, IN_RTS, 
IN_LIBGCC2, __GCC_FLOAT_NOT_NEEDED) - see my comments in 
<http://gcc.gnu.org/ml/gcc-patches/2010-10/msg00947.html>.

> * There are several one- or two-line makefile fragments in gcc and
>   libgcc.  What is our policy for adding copyright notices to them.  In
>   a couple of cases, the files are the only remainder/extract of a
>   previously far longer file.  Should they keep the copyright dates of
>   the former ones?

The FSF policy is that any file more than ten lines long should have a 
copyright and license notice; files up to ten lines don't need them.

http://www.gnu.org/prep/maintain/html_node/Copyright-Notices.html

-- 
Joseph S. Myers
jos...@codesourcery.com

Reply via email to