Thread: [gobo-eiffel-develop] No -O2 flag in gcc.cfg

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

I see that in gcc.cfg, the -O2 flag is commented out.
What's the reason for this?

Colin Adams wrote:
> I see that in gcc.cfg, the -O2 flag is commented out.
> What's the reason for this?

I don't remember. I guess it was back when I tried it in
SourceForge's compile farm. You can try uncommenting it
and see what happens.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

Eric Bezault wrote:
> Colin Adams wrote:
>> I see that in gcc.cfg, the -O2 flag is commented out.
>> What's the reason for this?
> 
> I don't remember. I guess it was back when I tried it in
> SourceForge's compile farm. You can try uncommenting it
> and see what happens.

I think I remember now: at the time I was testing it in
the SourceForge's compile farm, gec was generating C files
that were too big for the C compiler to compile with
the option -O2. Since then I modified gec to be able to
generate several smaller C files rather than one big file.

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

On 26/12/2007, Eric Bezault <er...@go...> wrote:
> Colin Adams wrote:
> > I see that in gcc.cfg, the -O2 flag is commented out.
> > What's the reason for this?
>
> I don't remember. I guess it was back when I tried it in
> SourceForge's compile farm. You can try uncommenting it
> and see what happens.

I timed one long-running test and the elapsed time became 2/3 of that
without -O2.
I than ran the entire W3C test suite. There was only one difference:

< <v toobig="yes">1.9345000000000003E+122</v>
---
> <v toobig="yes">1.9344999999999999E+122</v>

Since there are already differences on this test case with the W3C
expected results, and these are only floating point rounding
differences, I am happy, and will check in the change.

Colin Adams wrote:
> Since there are already differences on this test case with the W3C
> expected results, and these are only floating point rounding
> differences, I am happy, and will check in the change.

Can you run the Gobo bootstrap and test suite with this change first?

-- 
Eric Bezault
mailto:er...@go...
http://www.gobosoft.com

>>>>> "Eric" == Eric Bezault <er...@go...> writes:

    Eric> Colin Adams wrote:
    >> Since there are already differences on this test case with the
    >> W3C expected results, and these are only floating point
    >> rounding differences, I am happy, and will check in the change.

    Eric> Can you run the Gobo bootstrap and test suite with this
    Eric> change first?

Too late!
But I will do it now.
-- 
Colin Adams
Preston Lancashire

>>>>> "Colin" == Colin Paul Adams <co...@co...> writes:

>>>>> "Eric" == Eric Bezault <er...@go...> writes:
    Eric> Colin Adams wrote:
    >>> Since there are already differences on this test case with the
    >>> W3C expected results, and these are only floating point
    >>> rounding differences, I am happy, and will check in the
    >>> change.

    Eric> Can you run the Gobo bootstrap and test suite with this
    Eric> change first?

    Colin> Too late!  But I will do it now.

All tests pass.
-- 
Colin Adams
Preston Lancashire