Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-02-01 Thread Pedro Alves
[EMAIL PROTECTED] escreveu: > Hi Nuno, > > Quoting Nuno Lucas <[EMAIL PROTECTED]>: > >> Some functions in math.h (like rint), are inlined in the math.h >> header, but fail to compile with an error when used: >> >> > >> [EMAIL PROTECTED]:~/prjs/wince/cegcc/test/math_h_header$ cat main.c >> #inc

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-02-01 Thread Danny Backx
Looks like Pedro has something better than I do. I'll keep my hands off. Danny On Wed, 2007-01-31 at 08:14 +, [EMAIL PROTECTED] wrote: > I made a first approach to port the src/mingw/mingwex/math dir this weekend, > It is currently mostly i386/i387 asm, like what is found in that peac

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-02-01 Thread pedro_alves
Hi Nuno, Quoting Nuno Lucas <[EMAIL PROTECTED]>: > Some functions in math.h (like rint), are inlined in the math.h > header, but fail to compile with an error when used: > > > [EMAIL PROTECTED]:~/prjs/wince/cegcc/test/math_h_header$ cat main.c > #include > int main( void ) > { > double

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-01-31 Thread Nuno Lucas
On 1/31/07, Danny Backx <[EMAIL PROTECTED]> wrote: > I've extended the patch a bit, see attachment. Does this look > reasonable ? Basically I've changed all the inline assembler definitions > in a simple function definition, under the #if __MINGW32CE__ condition. You made a mistake on your patch:

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-01-31 Thread Danny Backx
I've extended the patch a bit, see attachment. Does this look reasonable ? Basically I've changed all the inline assembler definitions in a simple function definition, under the #if __MINGW32CE__ condition. Danny On Wed, 2007-01-31 at 20:26 +, Nuno Lucas wrote: > On 1/31/07, Danny Bac

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-01-31 Thread Nuno Lucas
On 1/31/07, Danny Backx <[EMAIL PROTECTED]> wrote: > That's only a partial solution : from my scan, the "asm" construction > occurs in these places in math.h : > [...] > Your patch only addresses the lines between 634 and 703. I'm aware of that, but I didn't know if the patch was the right thing t

Re: [Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-01-31 Thread Danny Backx
That's only a partial solution : from my scan, the "asm" construction occurs in these places in math.h : dannypc: {16} fgrep -n asm src/mingw/include/math.h 334: __asm__ ("fxam; fstsw %%ax;" : "=a" (sw): "t" (x)); 355: __asm__ ("fxam;" 364: __asm__ ("fxam;" 373: __asm__ ("fxam;" 390: __asm__

[Cegcc-devel] mingw32ce: Fix some math.h bad inlined functions

2007-01-30 Thread Nuno Lucas
Some functions in math.h (like rint), are inlined in the math.h header, but fail to compile with an error when used: [EMAIL PROTECTED]:~/prjs/wince/cegcc/test/math_h_header$ cat main.c #include int main( void ) { double i = rint( 3.14 ); return 0; } [EMAIL PROTECTED]:~/prjs/wince/cegcc