Mail Archives: djgpp/2009/05/26/02:30:26
Hi,
On May 26, 12:39=A0am, Andris Pavenis <andris DOT pave DOT DOT DOT AT iki DOT fi> wrote:
> On 05/26/2009 12:17 AM, Rugxulo wrote:
>
> > On May 25, 7:56 am, Andris Pavenis<andris DOT pave DOT DOT DOT AT iki DOT fi> =A0wrote:
>
> >> Perhaps the best source is the documentation (archives
> >> present at specified URL even if not montioned in earlier post) or gcc=
-help
> >> mailing list if the information is not found in documentation.
>
> > Yes, I know. I tried (and tried and tried ...). What I discovered is
> > that "__attribute__((optimize(2)))" works on function prototypes
> > (only), good!, but "#pragma GCC optimize" seems to be a complete no-
> > op, does nothing (or else I did it wrong, searching didn't really help
> > for that). Same results with unofficial MinGW GCC 4.4.0, also.
>
> If there is no help from the documentation then You can ask in gcc-help
> mailing list (search archives at first:http://gcc.gnu.org/lists.html).
I did search there (well, May at least), not easy to find otherwise. I
also found an old (11-month old?) patch regarding this, didn't shed a
lot of light. I tried a bunch of variations, still nothing. And also
read some hint somewhere that some ICEs were a result of this new
feature, so maybe they did disable it (partially / fully) on some /
all targets, who knows. Anyways, it's not a big deal. Although, for
the record, I did find my app that previously ran fastest with -O3
(4.3.2) now runs fastest with 4.4.0's -O2 (12% slower with -O3), hence
some interest on my part. But I'm just lucky I benchmarked it first
instead of blindly assuming anything!
> >> When I tried to specify that sockets are not
> >> supported, I got build errors for libada. I do not have enough time
> >> for fixing that (I do not know Ada well enough either), so the help is
> >> appreciated.
>
> > I wish I could help, but I don't know Ada at all, and obviously don't
> > have the skills to participate more. =A0 :-(
>
> The message was sent to mailing list, so maybe somebody who can and want
> it will read the message. I'm afraid however that the probability of this
> happening is rather small unfortunatelly.
BTW, thanks again for building 4.4.0, I actually was quite eager for
it for no huge reason (although I had already heard about "target" and
"optimize" pragmas), even downloading the unofficial MinGW 4.4.0
(blech, MSVCRT.DLL). ;-)
In short, being "stuck" to 4.3.2 / 4.3.3 for Ada wouldn't be the end
of the world. I mean, hey, it's Ada 83 / 95 / partial 2005, right? Not
too shabby. So thanks for what you've built so far. At least with
Gautier's 3d engine, DOS has at least one good publicly available
DJGPP Ada app! I mean, if people can still use TC++ 1.01 (and FreeDOS
does, somewhat), they can use "old" GCC 4.3.3 for many years. Worst
case scenario. :-)
- Raw text -