[webkit-dev] Linux port bug 8515

[Mike Emmel]

On 6/7/06, Darin Adler <darin at apple.com> wrote:
> On Jun 6, 2006, at 8:44 PM, Mike Emmel wrote:
>
> > +#Not sure what the earliest version with the defines is
> > +$flex_version=`flex --version`;
> > +if( $flex_version = ~/2\.5/ ) {
> > +print<<END
> > +/* C99 systems have <inttypes.h>. Non-C99 systems may or may not. */
> > +
> > +#if defined __STDC_VERSION__ && __STDC_VERSION__ >= 199901L
> > +#include <inttypes.h>
> > +typedef int8_t flex_int8_t;
> > +typedef uint8_t flex_uint8_t;
> > +typedef int16_t flex_int16_t;
> > +typedef uint16_t flex_uint16_t;
> > +typedef int32_t flex_int32_t;
> > +typedef uint32_t flex_uint32_t;
> > +#else
> > +typedef signed char flex_int8_t;
> > +typedef short int flex_int16_t;
> > +typedef int flex_int32_t;
> > +typedef unsigned char flex_uint8_t;
> > +typedef unsigned short int flex_uint16_t;
> > +typedef unsigned int flex_uint32_t;
> > +#endif /* ! C99 */
> > +END
> > +}
>
> This really looks to me like a problem on the Linux side. Why would
> you need to add all this code just to get the flex-generated code to
> compile? Are the other projects that use flex on Linux doing this?
>
>      -- Darin
>
>
The problem is the perl script actually only takes part of the flex output
if you used the flex generated file as is it would not matter.
I recall its doing this for unicode support.


Mike