sorry
the alpha was tested on linux
the coordination of SOCK_CLOEXEC etc. is still in progress
note that the workaround below may get ksh to build but ignores
the close-on-exec flag
the next alpha will be better on portability
On Wed, 12 Sep 2012 14:17:55 +0200 Irek Szczesniak wrote:
> Our build of ast-ksh.20120911 fails with src/cmd/ksh93/sh/io.c", line
> 1008: undefined symbol: SOCK_CLOEXEC
> I used the patch below to fix the problem, but then ran into the old
> bug with shtests: line 307: write to 1 failed [Bad file number].
> Now the question is: Is Solaris still going to be supported? Or is
> support fading away? I've seen many other opensource projects
> abandoning Solaris after Oracle killed Opensolaris...
> diff -r -u src/cmd/ksh93/sh/io.c src/cmd/ksh93/sh/io.c
> --- src/cmd/ksh93/sh/io.c 2012-09-05 23:53:53.000000000 +0200
> +++ src/cmd/ksh93/sh/io.c 2012-09-12 14:06:04.632100936 +0200
> @@ -973,13 +973,14 @@
> return(0);
> }
> -#ifndef _lib_accept4
> -# define accept4(a,b,c,d) accept(a,b,c)
> -#endif
> #if SHOPT_COSHELL
> int sh_coaccept(Shell_t *shp,int *pv,int out)
> {
> +#ifdef _lib_accept4
> int fd = accept4(pv[0],(struct
> sockaddr*)0,(socklen_t*)0,SOCK_CLOEXEC);
> +#else
> + int fd = accept(pv[0],(struct sockaddr*)0,(socklen_t*)0);
> +#endif
> sh_close(pv[0]);
> pv[0] = -1;
> if(fd<0)
> @@ -1005,7 +1006,11 @@
> int r,port=20000;
> struct sockaddr_in sin;
> socklen_t slen;
> +#ifdef SOCK_CLOEXEC
> if ((pv[out] = socket (AF_INET, SOCK_STREAM|SOCK_CLOEXEC, 0)) < 0)
> +#else
> + if ((pv[out] = socket (AF_INET, SOCK_STREAM, 0)) < 0)
> +#endif
> errormsg(SH_DICT,ERROR_system(1),e_pipe);
> do
> {
> Irek
> _______________________________________________
> ast-developers mailing list
> ast-developers@research.att.com
> https://mailman.research.att.com/mailman/listinfo/ast-developers
_______________________________________________
ast-developers mailing list
ast-developers@research.att.com
https://mailman.research.att.com/mailman/listinfo/ast-developers
