Constant.t fails 240 of 272 tests and recurs.t fails 1 of 25 tests on HPUX using perl 5.8.7

D

dayo

I am new to perl and am tasked with compiling and installing 5.8.7.
When running make test I see failures associated Constant.t , recurs.t
and 510_ping_upd.t

I am not concerned about the 510_ping_upd.t as other notes say I can
ignore this,
but I would like to resolve the failures for Constant.t and recurs.t

It may be that I am not telling the perl about a library it needs, but
I don't know how to tell what the tests are or why they are failing.

Can somebody tell me how to review the actual tests and failure reason?

Does anybody recognize whether a particular library or other parameter
might be missing from my configuration?
Lastly, being new to perl I'm not sure if maybe its ok just to igore
these failures - though I think I shouldn't because the INSTALL and
README files told me of other errors I might expect under certain
circumstances but said nothing of these - can I ignore this?

The summary report form ./perl harness
Failed Test Stat Wstat Total Fail Failed
List of Failed
.../lib/ExtUtils/t/Constant.t 272 240 88.24%
33-272
.../lib/ExtUtils/t/recurs.t 1 256 25 1 4.00%
14
.../lib/Net/Ping/t/510_ping_udp.t 2 1 50.00%
2

Site configuration information for perl v5.8.7:

Configured by oracle at Wed Dec 7 11:06:45 CST 2005.

Summary of my perl5 (revision 5 version 8 subversion 7) configuration:
Platform:
osname=hpux, osvers=11.11, archname=PA-RISC2.0-LP64
uname='hp-ux dx10host b.11.11 u 9000800 180404600 unlimited-user
license '
config_args='-Duselargefiles -Duse64bitall
-Dman3dir=/u80/dba_workarea/perltest/5.8.7/man/man3'
hint=recommended, useposix=true, d_sigaction=define
usethreads=undef use5005threads=undef useithreads=undef
usemultiplicity=undef
useperlio=define d_sfio=undef uselargefiles=define usesocks=undef
use64bitint=define use64bitall=define uselongdouble=undef
usemymalloc=n, bincompat5005=undef
Compiler:
cc='cc', ccflags =' -Ae -D_HPUX_SOURCE -Wl,+vnocompatwarnings +DD64
-I/opt/local/include +DD64 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64
',
optimize='+O2 +Onolimit',
cppflags='-Aa -D__STDC_EXT__ -D_HPUX_SOURCE -Ae -D_HPUX_SOURCE
-Wl,+vnocompatwarnings +DD64 -I/opt/local/include +DD64'
ccversion='B.11.11.04', gccversion='', gccosandvers=''
intsize=4, longsize=8, ptrsize=8, doublesize=8, byteorder=87654321
d_longlong=define, longlongsize=8, d_longdbl=define, longdblsize=16
ivtype='long', ivsize=8, nvtype='double', nvsize=8, Off_t='off_t',
lseeksize=8
alignbytes=8, prototype=define
Linker and Libraries:
ld='/usr/bin/ld', ldflags =' +DD64 -L/opt/local/lib -L/lib/pa20_64'
libpth=/opt/local/lib /lib/pa20_64 /lib /usr/lib /usr/ccs/lib
libs=-lnsl -lnm -ldl -ldld -lm -lsec -lc
perllibs=-lnsl -lnm -ldl -ldld -lm -lsec -lc
libc=/usr/lib/pa20_64/libc.sl, so=sl, useshrplib=false,
libperl=libperl.a
gnulibc_version=''
Dynamic Linking:
dlsrc=dl_hpux.xs, dlext=sl, d_dlsymun=undef, ccdlflags='-Wl,-E
-Wl,-B,deferred '
cccdlflags='+Z', lddlflags='-b +vnocompatwarnings -L/opt/local/lib
-L/lib/pa20_64'

Locally applied patches:


---
@INC for perl v5.8.7:
lib
/u80/dba_workarea/my587dest/lib/perl5/5.8.7/PA-RISC2.0-LP64
/u80/dba_workarea/my587dest/lib/perl5/5.8.7

/u80/dba_workarea/my587dest/lib/perl5/site_perl/5.8.7/PA-RISC2.0-LP64
/u80/dba_workarea/my587dest/lib/perl5/site_perl/5.8.7
/u80/dba_workarea/my587dest/lib/perl5/site_perl
 
A

Anno Siegel

dayo said:
I am new to perl and am tasked with compiling and installing 5.8.7.
When running make test I see failures associated Constant.t , recurs.t
and 510_ping_upd.t

You shouldn't see the test failures you're seeing, except possibly
the ping_udp failure (not ping_upd). What procedure are you using to
compile perl? Did you accept all defaults during ./Configure? After
../Configure -des, 5.8.7 should compile out of the box on any recent
HP-UX. A few warnings during compilation are okay, but not massive
amounts. Test failures that aren't mentioned in a nearby readme are
a reason not to install.

[details snipped, nothing blatant there]

Anno
 
D

dayo

Anno Siegel wrote:
What procedure are you using to compile perl?
Did you accept all defaults during ./Configure?

I ran Configure interactively but supplied the command shown here when
invoking it...
sh Configure -Duselargefiles -Duse64bitall
-Dman3dir=/u80/dba_workarea/perltest/5.8.7/man/man3

I did not accept all the defaults.

I had to specify the location of my C library
Where is your C library? [/lib/pa20_64/libc.sl]
/usr/lib/pa20_64/libc.sl

And I provided an installation prefix
Installation prefix to use? (~name ok) [/opt/perl5]
/u80/dba_workarea/my587dest

While Configure was running I also got 4 WHOA there messages for which
I accepted the default at that time:

The recommended value for $d_fpclassify on this machine was
"define"!
The recommended value for $d_isfinite on this machine was "define"!
The recommended value for $d_isinf on this machine was "define"!
The recommended value for $d_unordered on this machine was
"define"!

To all of the above I hit enter at Keep the recommended value?
[y]

I will try accepting all of the defaults except for "Installation
prefix to use?" to see if the test failures are fixed.
 
D

dayo

I continue to have this issue.
I have tried compliling with all of the default options and the results
are the same.
I have downloaded perl 5.8.5 and compiled with the default options and
with options I felt more appropriate and the results are the same.
Constant.t tests almost all fail.

Can somebody tell me how to see what each failing test is and how I can
run the failing tests individually to try to further understand the
issue?
 
A

Anno Siegel

dayo said:
I continue to have this issue.
I have tried compliling with all of the default options and the results
are the same.
I have downloaded perl 5.8.5 and compiled with the default options and
with options I felt more appropriate and the results are the same.
Constant.t tests almost all fail.

Something must be unusual about your system. I don't have HP-UX around
anywhere, but it's mainstream Unix -- perl should compile.
Can somebody tell me how to see what each failing test is and how I can
run the failing tests individually to try to further understand the
issue?

In the build directory:

setenv $X `pwd`
cd lib/ExtUtils/t
$X/perl -I$X Constant.t

This runs the tests in Constant.t directly.

Anno
 
D

dayo

I was able to run Constant.t as described thanks for that info...

In looking at all of the modules in the $X/lib/ExtUtils/t directory I
see that all the test scripts have all got #!/usr/bin/perl hard coded
at the top of them.
An earlier version of perl is installed in the /usr/bin/ directory (
version 5.8.0 ) so I presumed that might be my problem with the tests.

I edited all of the scripts in $X/lib/ExtUtils/t to point at $X/perl
(I did not code $X but the actual path).
When I made this change Constant.t succeeded on the first 32 tests and
failed on test 33, prior to this point it was failing on test #3.

Are there other places I should look to determine if any components for
the tests being run are pointing at /usr/bin/perl?
 
D

dayo

I ran Constant.t 8 times in succession over a few minutes using $X/perl
-I$X Constant.t

Instead of repeatedly failing at the same place, I found that the test
which could change each time I ran despite no other changes to my
setup. All of the failures were very similar indicating the
Makefile.PL is 0 second(s) older than Makefile. Any suggestions?

1st execution failed on test 33
2nd execution failed on test 103
3rd on 3
4th thru 6th on 33
7th on 229
and 8th execution failed test 66

each failure produced a message similar to the following message:

# Makefile.PL is 0 second(s) older than Makefile
# make = 'make'

Stop.
not ok 229 # make failed: 256
# Makefile out-of-date with respect to Makefile.PL
# Cleaning current config before rebuilding Makefile...
# make -f Makefile.old clean > /dev/null 2>&1 || /bin/sh -c true
# /u80/dba_workarea/perltest/perl-5.8.7/perl Makefile.PL
# Checking if your kit is complete...
# Looks good
# Writing Makefile for ExtTest
# ==> Your Makefile has been rebuilt. <==
# ==> Please rerun the make command. <==
# false
# *** Error exit code 1
# ext-5097 being removed...
 
D

dayo

I have now found on the web what seems to be an idenitcle match to what
I am seeing with sporadic failures of tests in constant.t.
Even the failing test numbers match mine.

I don't understand though what SMOKE is?
Is this saying there is a patch ( 23856 ) I can download to resolve
this issue?
Does somebody know how I get that patch?

Re: Smoke [5.8.6] 23856 FAIL(XF) hp-ux 11.11/64 (PA-2.0/32/1 cpu)

http://www.gossamer-threads.com/lists/perl/porters/189857?do=post_view_threaded#189857
 
I

Ilya Zakharevich

[A complimentary Cc of this posting was sent to
dayo
I ran Constant.t 8 times in succession over a few minutes using $X/perl
-I$X Constant.t

First of all, you need -I $X/lib

Second, you need to do this earlier.

env PERL5LIB=$X/lib $X/perl Makefile.PL
env PERL5LIB=$X/lib make
env PERL5LIB=$X/lib make test

Do not forget to run

make clean

before resetting the Perl version.

Hope htis helps,
Ilya
 
A

Anno Siegel

Ilya Zakharevich said:
[A complimentary Cc of this posting was sent to
dayo
I ran Constant.t 8 times in succession over a few minutes using $X/perl
-I$X Constant.t

First of all, you need -I $X/lib

Oh yes, my mistake (in an earlier posting). "-I $X" is useless, pardon
the misinformation.

Anno
 
D

dayo

Actually I got the same results using

setenv $X `pwd`
cd lib/ExtUtils/t
$X/perl -I$X Constant.t

As I got using alternate syntax you suggested.
In any case, I have been able to run Constant.t tests multiple times
and I found a SMOKE patch (though I don't know what SMOKE is) that is
addressing an identicle scenario. I issued a different reply in this
thread yesterday asking about this patch.

Thanks for your advice on how to run the tests.

Ilya said:
[A complimentary Cc of this posting was sent to
dayo
I ran Constant.t 8 times in succession over a few minutes using $X/perl
-I$X Constant.t

First of all, you need -I $X/lib

Second, you need to do this earlier.

env PERL5LIB=$X/lib $X/perl Makefile.PL
env PERL5LIB=$X/lib make
env PERL5LIB=$X/lib make test

Do not forget to run

make clean

before resetting the Perl version.

Hope htis helps,
Ilya
 
I

Ilya Zakharevich

[A complimentary Cc of this posting was sent to
dayo
each failure produced a message similar to the following message:

# Makefile.PL is 0 second(s) older than Makefile
# make = 'make'

Stop.
not ok 229 # make failed: 256
# Makefile out-of-date with respect to Makefile.PL
# Cleaning current config before rebuilding Makefile...
# make -f Makefile.old clean > /dev/null 2>&1 || /bin/sh -c true
# /u80/dba_workarea/perltest/perl-5.8.7/perl Makefile.PL
# Checking if your kit is complete...
# Looks good
# Writing Makefile for ExtTest
# ==> Your Makefile has been rebuilt. <==
# ==> Please rerun the make command. <==
# false
# *** Error exit code 1
# ext-5097 being removed...

These are bogus. I suspect this is a combination of several bugs:

a) Makefile is not allowed to be of the same age as Makefile.PL;

b) Test suite does not wait 3 sec between creating Makefile.PL and
running Perl on it (some filesystems have 2sec granularity, +1sec
of savety margin);

Of these, only "a" is a real bug. "b" would be just a workaround
against "a".

Actually, maybe "a" is a feature too (since something might have
edited/recreated Makefile.PL after a run of `perl Makefile.PL', and
all this happened inside a filesystem timestamp granularity). One
should decide... But one of "a" and "b" *must* be implemented for the
test to work on quickier machines.

Try looking for the string "Makefile out-of-date with respect to" in
the source of lib/ExtUtils/* and check the test from < to <= (or visa
versa ;-).

Hope this helps,
Ilya
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

No members online now.

Forum statistics

Threads
473,981
Messages
2,570,188
Members
46,731
Latest member
MarcyGipso

Latest Threads

Top