Re: strsignal.c implementation

[Jeff Johnston <jjohnstn at redhat dot com>]

On 05/11/2010 12:24 PM, Joel Sherrill wrote:
> On 05/11/2010 10:49 AM, Jeff Johnston wrote:
> > On 05/11/2010 09:26 AM, Joel Sherrill wrote:
> > > Hi,
> > >
> > > I have implemented strsignal.c for RTEMS but
> > > wondered if it was worth trying to make the
> > > implementation match the ifdef's in sys/signal.h
> > > and support other targets?
> > >
> > > I have attached the current implementation. I
> > > see two approaches to making it support all
> > > the targets:
> > >
> > > + ifdefs in the table to match the numbers in
> > > sys/signal.h.
> > > + a long series of code something like this:
> > > #ifdef SIGXXX
> > > if ( signal == SIGXXX )
> > > return "string for SIGXXX";
> > > #endif
> > >
> > > Any thoughts or comments?
> > Yes, IMO it is worth it to make it a shared implementation. Option 2 is
> > what is used in libc/string/strerror.c for the errno strings.
> How about the attached version? I went through the list of
> signals and added everyone in sys/signal.h. When the values
> could be the same, I added a little extra conditional magic
> to see if both were defined and the same to avoid
> duplicate case errors.

Looks good.  Any chance you could add some documentation to it to save 
me the trouble?  It does not have to be as detailed as strerror.c.

Thanks,

-- Jeff J.

> 2010-05-11 Joel Sherrill <joel.sherrill@oarcorp.com>
>
> * libc/string/strsignal.c: New file.
>
> > -- Jeff J.