Mailing List Archive

On Thu, Dec 30, 1999 at 11:24:10AM -0400, Marc G. Fournier wrote:

> configure:2050: checking for quad_t
> configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
> configure: In function `main':
> configure:2055: incompatible types in assignment
> configure: failed program was:
> #line 2052 "configure"
> #include "confdefs.h"
> #include <sys/types.h>
> int main() {
> quad_t a; a = 1235;
> ; return 0; }

Does this help anything? (I'd recommend using autoconf if at all possible,
but if you can't, try the configure patch.)

David


--- configure.in.orig Thu Dec 30 10:29:52 1999
+++ configure.in Thu Dec 30 10:30:29 1999
@@ -141,7 +141,7 @@
AC_MSG_CHECKING([for quad_t])
AC_TRY_COMPILE(
[#include <sys/types.h>],
- [quad_t a; a = 1235;],
+ [quad_t a; a = (quad_t) 1235;],
[
AC_DEFINE(HAVE_QUAD_T)
AC_MSG_RESULT(yes)
--- configure.orig Thu Dec 30 10:32:55 1999
+++ configure Thu Dec 30 09:20:47 1999
@@ -2052,7 +2052,7 @@
#include "confdefs.h"
#include <sys/types.h>
int main() {
-quad_t a; a = 1235;
+quad_t a; a = (quad_t) 1235;
; return 0; }
EOF
if { (eval echo configure:2059: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then

Nope, produces the following instead:

configure:2050: checking for quad_t
configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
configure: In function `main':
configure:2055: conversion to non-scalar type requested
configure: failed program was:
#line 2052 "configure"
#include "confdefs.h"
#include <sys/types.h>
int main() {
quad_t a; a = (quad_t) 1235;
; return 0; }


On Thu, 30 Dec 1999, David Rankin wrote:

> On Thu, Dec 30, 1999 at 11:24:10AM -0400, Marc G. Fournier wrote:
>
> > configure:2050: checking for quad_t
> > configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
> > configure: In function `main':
> > configure:2055: incompatible types in assignment
> > configure: failed program was:
> > #line 2052 "configure"
> > #include "confdefs.h"
> > #include <sys/types.h>
> > int main() {
> > quad_t a; a = 1235;
> > ; return 0; }
>
> Does this help anything? (I'd recommend using autoconf if at all possible,
> but if you can't, try the configure patch.)
>
> David
>
>
> --- configure.in.orig Thu Dec 30 10:29:52 1999
> +++ configure.in Thu Dec 30 10:30:29 1999
> @@ -141,7 +141,7 @@
> AC_MSG_CHECKING([for quad_t])
> AC_TRY_COMPILE(
> [#include <sys/types.h>],
> - [quad_t a; a = 1235;],
> + [quad_t a; a = (quad_t) 1235;],
> [
> AC_DEFINE(HAVE_QUAD_T)
> AC_MSG_RESULT(yes)
> --- configure.orig Thu Dec 30 10:32:55 1999
> +++ configure Thu Dec 30 09:20:47 1999
> @@ -2052,7 +2052,7 @@
> #include "confdefs.h"
> #include <sys/types.h>
> int main() {
> -quad_t a; a = 1235;
> +quad_t a; a = (quad_t) 1235;
> ; return 0; }
> EOF
> if { (eval echo configure:2059: \"$ac_compile\") 1>&5; (eval $ac_compile) 2>&5; }; then
>

Marc G. Fournier marc.fournier@acadiau.ca
Senior Systems Administrator Acadia University

"These are my opinions, which are not necessarily shared by my employer"

On Thu, Dec 30, 1999 at 01:20:28PM -0400, Marc G. Fournier wrote:

> Nope, produces the following instead:

> configure:2050: checking for quad_t
> configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
> configure: In function `main':
> configure:2055: conversion to non-scalar type requested
> configure: failed program was:
> #line 2052 "configure"
> #include "confdefs.h"
> #include <sys/types.h>
> int main() {
> quad_t a; a = (quad_t) 1235;
> ; return 0; }
>

Why don't you send out the line in /usr/include/sys/types that defines
quad_t? I'm beginning to suspect that Solaris 7 (I can only access 2.6)
is using quad_t as something != int64_t or long long int.

Thanks,
David

--
David W. Rankin, Jr. Husband, Father, and UNIX Sysadmin.
Email: drankin@bohemians.lexington.ky.us Address/Phone Number: Ask me.
"It is no great thing to be humble when you are brought low; but to be humble
when you are praised is a great and rare accomplishment." St. Bernard

dragon:/var/patches/OpenWindows_3.6.1_x86> grep quad_t
/usr/include/sys/types.h
typedef struct _quad { int val[2]; } quad_t; /* used by UFS */
typedef quad_t quad; /* used by UFS */


On Thu, 30 Dec 1999, David Rankin wrote:

> On Thu, Dec 30, 1999 at 01:20:28PM -0400, Marc G. Fournier wrote:
>
> > Nope, produces the following instead:
>
> > configure:2050: checking for quad_t
> > configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
> > configure: In function `main':
> > configure:2055: conversion to non-scalar type requested
> > configure: failed program was:
> > #line 2052 "configure"
> > #include "confdefs.h"
> > #include <sys/types.h>
> > int main() {
> > quad_t a; a = (quad_t) 1235;
> > ; return 0; }
> >
>
> Why don't you send out the line in /usr/include/sys/types that defines
> quad_t? I'm beginning to suspect that Solaris 7 (I can only access 2.6)
> is using quad_t as something != int64_t or long long int.
>
> Thanks,
> David
>
> --
> David W. Rankin, Jr. Husband, Father, and UNIX Sysadmin.
> Email: drankin@bohemians.lexington.ky.us Address/Phone Number: Ask me.
> "It is no great thing to be humble when you are brought low; but to be humble
> when you are praised is a great and rare accomplishment." St. Bernard
>

Marc G. Fournier marc.fournier@acadiau.ca
Senior Systems Administrator Acadia University

"These are my opinions, which are not necessarily shared by my employer"

On Thu, Dec 30, 1999 at 03:12:18PM -0400, Marc G. Fournier wrote:

> dragon:/var/patches/OpenWindows_3.6.1_x86> grep quad_t
> /usr/include/sys/types.h
> typedef struct _quad { int val[2]; } quad_t; /* used by UFS */
> typedef quad_t quad; /* used by UFS */

I was afraid of something like this. I did a grep of the entire source
code, and I didn't notice a single place quad_t was used withing OpenSSH
itself. Is there a reason why it's being checked? If not, then perhaps
the best "fix" is to just comment this out.

If a quad_t check is necessary, this configure.in patch might help. If
we eventually need to tell the difference between Solaris' quad_t and
the int64_t equivalent, do it here.

Thanks,
David


--- configure.in.orig Thu Dec 30 14:50:11 1999
+++ configure.in Thu Dec 30 14:52:08 1999
@@ -149,6 +149,18 @@
[AC_MSG_RESULT(no)]
)

+dnl XXX on Solaris 7 at least, quad_t is actually a struct.
+AC_MSG_CHECKING([for quad_t Solaris 7 style])
+AC_TRY_COMPILE(
+ [#include <sys/types.h>],
+ [quad_t a; a.val[0] = 1235;],
+ [
+ AC_DEFINE(HAVE_QUAD_T)
+ AC_MSG_RESULT(yes)
+ ],
+ [AC_MSG_RESULT(no)]
+)
+
AC_MSG_CHECKING([for intXX_t types])
AC_TRY_COMPILE(
[#include <sys/types.h>],

"Marc G. Fournier" wrote:
>
> configure:2050: checking for quad_t
> configure:2059: gcc -c -g -O2 -Wall -I/usr/slocal/include conftest.c 1>&5
> configure: In function `main':
> configure:2055: incompatible types in assignment
> configure: failed program was:
> #line 2052 "configure"
> #include "confdefs.h"
> #include <sys/types.h>
> int main() {
> quad_t a; a = 1235;
> ; return 0; }

How is quad_t defined in sys/types.h?

Regards,
Damien

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Thu, 30 Dec 1999, David Rankin wrote:

> I was afraid of something like this. I did a grep of the entire source
> code, and I didn't notice a single place quad_t was used withing OpenSSH
> itself. Is there a reason why it's being checked? If not, then perhaps
> the best "fix" is to just comment this out.

You are correct. quad_t was used in the past for something in scp
(don't remember what), but is no longer.

I have removed the test and the typedef.

Regards,
Damien

- --
| "Bombay is 250ms from New York in the new world order" - Alan Cox
| Damien Miller - http://www.mindrot.org/
| Email: djm@mindrot.org (home) -or- djm@ibs.com.au (work)


-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.0 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE4a+CiormJ9RG1dI8RApewAKDQMv4Kie6GDXMkOe6sFA3EHHtKkACfY9In
dM9yd4ih2IdjJ9nD6Y6VoyM=
=R5vL
-----END PGP SIGNATURE-----