#2352: POSIX.1 unsetenv returns int
----------------------------+-----------------------------------------------
Reporter: donn | Owner: igloo
Type: bug | Status: new
Priority: normal | Milestone: 6.10.1
Component: libraries/unix | Version: 6.8.2
Severity: minor | Resolution:
Keywords: | Difficulty: Unknown
Testcase: | Architecture: x86_64 (amd64)
Os: NetBSD |
----------------------------+-----------------------------------------------
Changes (by igloo):
* owner: => igloo
* difficulty: => Unknown
* milestone: => 6.10.1
Old description:
> While plenty of C libraries declare void unsetenv(), the POSIX.1003.1
> standard is int unsetenv(), with return values 0 or -1. In the latter
> case, errno is set to EINVAL, for causes including the present of "=" in
> the value. Of course this is not of earthshaking importance.
>
> The following is the autoconf test (just a copy of the same test for
> usleep().)
>
> --- libraries/unix/configure.ac.dist 2008-06-03 10:39:45.000000000
> -0700
> +++ libraries/unix/configure.ac 2008-06-06 21:43:19.000000000 -0700
> @@ -75,6 +75,19 @@
> ;;
> esac
>
> +### POSIX.1003.1 unsetenv returns 0 or -1 (EINVAL), but older
> implementations
> +### in common use return void.
> +AC_CACHE_CHECK([return type of unsetenv], cv_func_unsetenv_return_type,
> + [AC_EGREP_HEADER(changequote(<, >)<void[
> ]+unsetenv>changequote([, ]),
> + /usr/include/stdlib.h,
> + [cv_func_unsetenv_return_type=void],
> + [cv_func_unsetenv_return_type=int])])
> +case "$cv_func_unsetenv_return_type" in
> + "void" )
> + AC_DEFINE([UNSETENV_RETURNS_VOID], [1], [Define if stdlib.h declares
> unsetenv to return void.])
> + ;;
> +esac
> +
> dnl ** sometimes RTLD_NEXT is hidden in #ifdefs we really don't wan to
> set
> AC_MSG_CHECKING(for RTLD_NEXT from dlfcn.h)
> AC_EGREP_CPP(yes,
New description:
While plenty of C libraries declare void unsetenv(), the POSIX.1003.1
standard is int unsetenv(), with return values 0 or -1. In the latter
case, errno is set to EINVAL, for causes including the present of "=" in
the value. Of course this is not of earthshaking importance.
The following is the autoconf test (just a copy of the same test for
usleep().)
{{{
--- libraries/unix/configure.ac.dist 2008-06-03 10:39:45.000000000
-0700
+++ libraries/unix/configure.ac 2008-06-06 21:43:19.000000000 -0700
@@ -75,6 +75,19 @@
;;
esac
+### POSIX.1003.1 unsetenv returns 0 or -1 (EINVAL), but older
implementations
+### in common use return void.
+AC_CACHE_CHECK([return type of unsetenv], cv_func_unsetenv_return_type,
+ [AC_EGREP_HEADER(changequote(<, >)<void[ ]+unsetenv>changequote([,
]),
+ /usr/include/stdlib.h,
+ [cv_func_unsetenv_return_type=void],
+ [cv_func_unsetenv_return_type=int])])
+case "$cv_func_unsetenv_return_type" in
+ "void" )
+ AC_DEFINE([UNSETENV_RETURNS_VOID], [1], [Define if stdlib.h declares
unsetenv to return void.])
+ ;;
+esac
+
dnl ** sometimes RTLD_NEXT is hidden in #ifdefs we really don't wan to
set
AC_MSG_CHECKING(for RTLD_NEXT from dlfcn.h)
AC_EGREP_CPP(yes,
}}}
Comment:
Thanks for the patch, we'll take a look
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/2352#comment:1>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler_______________________________________________
Glasgow-haskell-bugs mailing list
Glasgow-haskell-bugs@haskell.org
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
