Quoting Richard Guenther <richard.guent...@gmail.com>:
On Tue, Jun 14, 2011 at 11:40 AM, Joern Rennecke <amyl...@spamcop.net> wrote:
Except or the fortran/java bits (committed), this patch hasn't been
reviewed for five weeks:
http://gcc.gnu.org/ml/gcc-patches/2011-05/msg00582.html
A patch doing s/CUMULATIVE_ARGS*/cumulative_args_t/ only
is ok.
It's not quite that simple. The patch makes a distinction between pointers
to the target specific types CUMULATIVE_ARGS, and the target-independent
cumulative_args_t.
Is it still OK if I selectively do the replacement where the
target-independent type is meant, and add a provisional
typedef CUMULATIVE_ARGS *cumulative_args_t to tie it together?
Posting compressed attached patches makes it too easy
to not review things btw ...
The mailing list size limits did't allow this patch to be posted
without compression.
After that patch the "meat" of the patch should be much much smaller
and easier to review (if there is anything left besides the renaming?).
It should be somewhat smaller, but there are lots of places where we have
to convert between cumulative_args_t and CUMULATIVE_ARGS *.
Were a target-independent interface is required, we need cumulative_args_t .
Where a target accesses struct components, it needs CUMULATIVE_ARGS *.
There are some places that just pass CUMULATIVE_ARGS * around, both in
rtl-centric middle-end/ rtl-optimizer code and in target code, which
could be electively converted. In general, I haven't done such optional
conversions. They could be added according to taste once the interface
has been straightened out. There is also a judgement call in each place
how closely the code is tied to the cumulative_args_t side or the
CUMULATIVE_ARGS * side.
