[ net-snmp-Bugs-1433786 ] BIG_ENDIAN vs LITTLE_ENDIAN - Solaris x86

[SourceForge.net <no...@so...>]

Bugs item #1433786, was opened at 2006-02-17 14:36
Message generated for change (Comment added) made by jvande
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112694&aid=1433786&group_id=12694

Please note that this message will contain a full copy of the comment thread,
including the initial issue submission, for this request,
not just the latest update.
Category: configure
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Jon Vandegrift (jvande)
Assigned to: Nobody/Anonymous (nobody)
Summary: BIG_ENDIAN vs LITTLE_ENDIAN - Solaris x86

Initial Comment:
configure does not correctly determine the BYTE_ORDER 
on Solaris x86.

./configure    # works correctly
./make         # works correctly
./make test    # fails everything with "Segmentation 
Fault - core dumped"

according to this URL, Solaris x86 should be Little-
Endian. 
http://www.rsinc.com/services/techtip.asp?ttid=1804

# grep BIG config.*   # shows this
config.log:| #define WORDS_BIGENDIAN 1
config.log:#define WORDS_BIGENDIAN 1
config.status:${ac_dA}WORDS_BIGENDIAN${ac_dB}
WORDS_BIGENDIAN${ac_dC}1${ac_dD}
config.status:${ac_uA}WORDS_BIGENDIAN${ac_uB}
WORDS_BIGENDIAN${ac_uC}1${ac_uD}

I have both Solaris Sparc 9 & Solaris Intel 9 
platforms.
Extracting chunks of ./configure where the test is 
done for ENDIAN as 1.c, 2.c, 3.c, 4.c.  They are 
enclosed in the attached ZIP file (endian_test.zip) 
file.  They are in the order of testing in ./configure.

1.c fails to compile on either platform.

2.c compiles on both.

Solaris x86 output
# 2 ; echo $?
0

Solaris Sparc output
# 2 ; echo $?
0

3.c compiles on both.

Solaris x86 output
# 3 ; echo $? ; strings 3 | tail -2
0
IBeGDnaiSnSy
LiTTleEnDian

Solaris Sparc output
# 3 ; echo $? ; strings 3 | tail -2
0
BIGenDianSyS
iLTTelnEiDna

4.c compiles on both

Solaris x86 output
# 4 ; echo $?
0

Solaris Sparc output
# 4 ; echo $?
1


Since 2.c compiles and returns the same results on 
Sparc & Intel, both get set to BIG_ENDIAN.

I downloaded the net-snmp-5.1.1.tar.gz & net-snmp-
5.3.0.1.tar.gz source files, and also tried the CVS.  
Both have the same problem.
SunFreeware.com has a prebuilt netsnmp-5.1.1-sol9-
intel-local.gz package installation, but that does not 
include the Perl modules.  I wonder how they got it to 
build???

Please fix.



----------------------------------------------------------------------

>Comment By: Jon Vandegrift (jvande)
Date: 2006-02-17 15:56

Message:
Logged In: YES 
user_id=612950

Checked config.log.  Part that 4.c is based on is failing 
to run with a segmentation fault.

configure:18188: gcc -o conftest -g -O2 -Dsolaris2 -
I/usr/local/ssl/include  -L/usr/local/ssl/lib  conftest.c
 -lcrypto -lkstat -lelf -lm  -lnsl -lsocket -ladm >&5
configure: In function 'main':
configure:18320: warning: incompatible implicit declaration 
of built-in function 'exit'
configure:18191: $? = 0
configure:18193: ./conftest
configure: line 1:  8983 Segmentation Fault      (core 
dumped) ./conftest$ac_exeext
configure:18196: $? = 139
configure: program exited with status 139


this is due to the -lcrypto.

recompile 4.c with -lcrypto and run it.

# gcc 4.c -o 44 -lcrypto
4.c: In function 'main':
4.c:14: warning: incompatible implicit declaration of built-
in function 'exit'
# 44
Segmentation Fault - core dumped



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=112694&aid=1433786&group_id=12694