Bug 553 - configure fails to acknowledge availability of utimes()
Summary: configure fails to acknowledge availability of utimes()
Status: CLOSED FIXED
Alias: None
Product: Portable OpenSSH
Classification: Unclassified
Component: Build system (show other bugs)
Version: -current
Hardware: HPPA HP-UX
: P2 normal
Assignee: OpenSSH Bugzilla mailing list
URL:
Keywords:
Depends on:
Blocks: 627
  Show dependency treegraph
 
Reported: 2003-05-03 08:51 AEST by Doug Wyatt
Modified: 2004-04-14 12:24 AEST (History)
1 user (show)

See Also:


Attachments
config.log.gz from HP-UX 11.00 (20.42 KB, application/octet-stream)
2003-05-03 11:24 AEST, Darren Tucker
no flags Details
configure from 3.6.1p2 rebuilt with autoconf-2.57a (60.18 KB, application/octet-stream)
2003-05-13 10:26 AEST, Darren Tucker
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Doug Wyatt 2003-05-03 08:51:51 AEST
Running configure for OpenSSH-3.6.1p2 indicates that utimes()
is not found.  This leads to a type conflict error which
prevents Making the package.  The problem did not exist under
3.6p1 and all previous versions that I've built.  It appears
to be a result of a new automake version, but that is only
wild speculation.

Changing line 6187 in the configure file from 'char utimes ();'
to 'int utimes ();' appears to provide a workaround so that
OpenSSH can be configured and built.

I'm working under HP-UX 10.20 using Gcc 3.0 and mostly Gnu tools.
Comment 1 Darren Tucker 2003-05-03 10:11:40 AEST
The test (trimmed) fragments from configure:

3.6.1p1 (autoconf 2.53):
char utimes ();
char (*f) ();

int
main ()
{
#if defined (__stub_utimes) || defined (__stub___utimes)
choke me
#else
f = utimes;
#endif
  return 0;
}

3.6.1p2 (autoconf 2.57):
char utimes ();
#if defined (__stub_utimes) || defined (__stub___utimes)
choke me
#else
char (*f) () = utimes;
#endif

int
main ()
{
return f != utimes;
  return 0;
}

I'll upgrade autoconf on my HP (11.00) and see if I can reproduce the problem.

On a slightly related note, do we need to check for utimes() twice?  There's a 
specific check for it, then it's also in the main AC_CHECK_FUNCS list.
Comment 2 Tim Rice 2003-05-03 10:54:17 AEST
It would be nice to see the section of config.log where the test fails
on HP.

Re: "utimes also in the main AC_CHECK_FUNCS list", it was probably an
oversite on my part. If you yank it out of the AC_CHECK_FUNCS section,
you'll have to add a HAVE_UTIMES template to acconfig.h
Comment 3 Darren Tucker 2003-05-03 11:24:24 AEST
Created attachment 281 [details]
config.log.gz from HP-UX 11.00

Full log attached.  The failing piece is:

configure:6109: result: yes
configure:6175: checking for utimes
configure:6225: gcc -o conftest -g -O2 -Wall -Wpointer-arith -Wno-uninitialized
 -D_HP
UX_SOURCE -D_XOPEN_SOURCE -D_XOPEN_SOURCE_EXTENDED=1  conftest.c -lz -lnsl
-lxnet	-ls
ec >&5
configure:6250: conflicting types for `utimes'
/usr/include/sys/time.h:484: previous declaration of `utimes'
configure:6228: $? = 1
configure: failed program was:
Comment 4 Darren Tucker 2003-05-11 10:39:44 AEST
I was about to submit a bug report to the autoconf folks, but before I did I 
pulled the latest autoconf CVS tree (it calls itself 2.57a).  This problem 
appears to be fixed there.

I rebuilt configure with it and then rebuilt OpenSSH on HP-UX 11.00, which 
worked OK.
Comment 5 Petter Reinholdtsen 2003-05-13 02:23:18 AEST
I see the same problem on HP/UX 11.00 on parisc and 11.22 on ia64
using the native compiler.
Comment 6 Darren Tucker 2003-05-13 10:26:03 AEST
Created attachment 291 [details]
configure from 3.6.1p2 rebuilt with autoconf-2.57a

Updated configure, just replace configure from 3.6.1p2 then "make distclean &&
./configure && make".

This works for me (11.00 w/gcc), I'd be interested in other
versions/configurations.
Comment 7 Darren Tucker 2003-08-24 11:46:53 AEST
Have tested autoconf-2.52 on HP-UX for this and it works OK.
Comment 8 Darren Tucker 2003-09-23 23:07:19 AEST
3.7p1, 3.7.1p1 and 3.7.1p2 are all built with autoconf 2.52 so do not have this
problem.
Comment 9 Damien Miller 2004-04-14 12:24:19 AEST
Mass change of RESOLVED bugs to CLOSED