Re: [AD] SF.net SVN: alleg:[11393] allegro/branches/4.9

[Elias Pschernig <elias.pschernig@xxxxxxxxxx>]

On Tue, 2009-01-06 at 07:11 -0700, Thomas Fjellstrom wrote:
> On January 6, 2009, Evert Glebbeek wrote:
> > On 6-Jan-09, at 2:01 PM, Thomas Fjellstrom wrote:
> > > As the code says, 1 for true, 0 for false, -1 on error.
> >
> > That's horrible.
> > It should be possible to do "if (al_path_exists()) ...". When does it
> > fail (I know, I could check the code)? Is there a better way to detect
> > when an error occurs (eg, is there an "errno" type variable)?
> 
> There is currently no fshook specific errno variable. in the stdio driver's 
> "exists" hook, if stat fails, it checks libc's errno for ENOENT and returns 0, 
> otherwise it returns -1.
> 
> I have no problems with changing it.. But I thought actually being able to 
> distinguish between a negative condition and a failure was important. I'm not 
> so sure now.
> 
> If we add some form of errno variable, should it be fshook only? or should we 
> add back allegro_error? (iirc, it was removed...)

Instead of a global variable it should be al_get_error() though, and it
would be thread local.

-- 
Elias Pschernig <elias@xxxxxxxxxx>