You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@httpd.apache.org by Martin Kraemer <Ma...@mch.sni.de> on 1998/05/10 00:13:25 UTC
[PATCH] own sys_siglist on SVR4 & others
SVR4 has sys_siglist, but only in the UCB compatibility library,
/usr/ucblib/libucb.a -- and there are subtle incompatibility bugs in the
readdir() interface if you link it in before/after libc (depending on the
header file include order: /usr/ucbinclude after/before /usr/include).
So, to avoid trouble, it's best to not link against libucb. But without
it, we can not print the signal names. Dean already said it,
it's mainly a bunch of #ifdef's. How about the attached patch to add
signal names to every platform without sys_siglist?
Should the name sys_siglist be changed to ap_sys_siglist in the patch?
Martin
--
| S I E M E N S | <Ma...@mch.sni.de> | Siemens Nixdorf
| ------------- | Voice: +49-89-636-46021 | Informationssysteme AG
| N I X D O R F | FAX: +49-89-636-44994 | 81730 Munich, Germany
~~~~~~~~~~~~~~~~My opinions only, of course; pgp key available on request
Re: [PATCH] own sys_siglist on SVR4 & others
Posted by Martin Kraemer <Ma...@mch.sni.de>.
Oops, I just see that Linux has another NSIG variant, with _one_
leading underscore. Thus the name _NSIG should be tried as well.
Martin
On Sun, May 10, 1998 at 01:13:25AM +0200, Martin Kraemer wrote:
> +#if defined(NSIG)
> +# define NumSIG NSIG
> +#elif defined(__NSIG)
> +# define NumSIG __NSIG
> +#else
> +# define NumSIG 32
> +#endif
--
| S I E M E N S | <Ma...@mch.sni.de> | Siemens Nixdorf
| ------------- | Voice: +49-89-636-46021 | Informationssysteme AG
| N I X D O R F | FAX: +49-89-636-44994 | 81730 Munich, Germany
~~~~~~~~~~~~~~~~My opinions only, of course; pgp key available on request
Re: [PATCH] own sys_siglist on SVR4 & others
Posted by Dean Gaudet <dg...@arctic.org>.
Looks good... I wonder if you should move the section:
+#if defined(NSIG)
+#define NumSIG NSIG
+#elif defined(__NSIG)
+#define NumSIG __NSIG
+#else
+#define NumSIG 32
+#endif
outside, because NSIG seems to be present on linux/solaris as well... and
it'd be good to make sure the signal is within the size of the array.
On Sun, 10 May 1998, Martin Kraemer wrote:
> So, to avoid trouble, it's best to not link against libucb. But without
> it, we can not print the signal names. Dean already said it,
> it's mainly a bunch of #ifdef's. How about the attached patch to add
> signal names to every platform without sys_siglist?
>
> Should the name sys_siglist be changed to ap_sys_siglist in the patch?
Probably... avoids any chance of conflict.
Dean