Re: shadow fails to compile with gcc-2
Mike Frysinger <[email protected]> Sun, 4 Sep 2005 14:06:16 -0400
| Newsgroups | gmane.linux.pld.shadow.general |
|---|---|
| Organization | wh0rd.org |
| Message-ID | <[email protected]> |
On Sunday 04 September 2005 09:33 am, Tomasz Kłoczko wrote: > On Sat, 3 Sep 2005, Mike Frysinger wrote: > > src/newgrp.c declares some variables in the middle of scope and gcc-2 > > rejects this ... the child/pid variables could either be moved up to the > > top of the if scope, or the USE_PAM code could be given a local scope ... > > ive attached a patch by Martin Schlemmer for the former behavior > > but understand > also me: I want keep shadow code in best readable form. > Current form in this area is IMO better. yeah, the current form is def more readable than the version the previous patch changed it to how about this patch instead ? -mike
shadow-newgrp-gcc2.patch
(text/x-diff, 721 B)
Index: src/newgrp.c
===================================================================
RCS file: /cvsroot/shadow/src/newgrp.c,v
retrieving revision 1.40
diff -u -p -r1.40 newgrp.c
--- src/newgrp.c 31 Aug 2005 17:25:00 -0000 1.40
+++ src/newgrp.c 4 Sep 2005 18:04:30 -0000
@@ -459,6 +459,7 @@ int main (int argc, char **argv)
* avoid any possibility of the parent being stopped when it
* receives SIGCHLD from the terminating subshell. -- JWP
*/
+ {
pid_t child, pid;
signal (SIGINT, SIG_IGN);
@@ -502,6 +503,7 @@ int main (int argc, char **argv)
signal (SIGTSTP, SIG_DFL);
signal (SIGTTIN, SIG_DFL);
signal (SIGTTOU, SIG_DFL);
+ }
#endif /* USE_PAM */
}
#endif /* USE_SYSLOG */