We observe the following mod_perl test failures on HP-UX 11.00 on PA-RISC,
with both HP C and GCC in both 32 bit and 64 bit mode. They are not
present on IPF systems in any of these combinations.
Failed Test Total Fail Failed List of Failed
--------------------------------------------------------
api/slurp_filename.t 5 2 40.00% 2 4
modperl/taint.t 4 2 50.00% 1 3
modules/include.t 6 2 33.33% 4-5
modules/include2.t 4 1 25.00% 4
In 64 bit mode, there are 2 additional test failures that I just posted
about 10 minutes ago.
All tests are with Apache 2.0.48, Perl 5.8.2 and mod_perl 1.99_11, all
compiled with the same compiler.
Any ideas what may be failing and/or where to look further?
Thanks,
-Jan
TJ Saunders <[EMAIL PROTECTED]> wrote:
>> api/slurp_filename................1..5
>> # Running under perl version 5.008002 for hpux
>> # Current time local: Thu Dec 4 11:07:29 2003
>> # Current time GMT: Thu Dec 4 19:07:29 2003
>> # Using Test.pm version 1.24
>>
>> [snip]
>>
>> # testing : slurp filename tainted
>> # expected: (?-xism:Insecure dependency in eval)
>> # received:
>> not ok 2
>>
>> [snip]
>>
>> # testing : slurp filename (perl) tainted
>> # expected: (?-xism:Insecure dependency in eval)
>> # received:
>> not ok 4
>
>In both of these test cases, data is being slurped in via some
>slurp_filename() and slurp_filename_perl() methods. That data is then
>eval'd; both of the above tests fail because the eval() failures to emit a
>warning about "Insecure dependency in eval". The difference between the
>two methods is that the second is a pure Perl implementation.
>
>> modperl/taint.....................1..4
>> # Running under perl version 5.008002 for hpux
>> # Current time local: Thu Dec 4 11:07:29 2003
>> # Current time GMT: Thu Dec 4 19:07:29 2003
>> # Using Test.pm version 1.24
>> # testing : ${^TAINT}
>> # expected: 1
>> # received: 0
>> not ok 1
>>
>> [snip]
>>
>> # testing : $Apache::__T
>> # expected: 1
>> # received: 0
>> not ok 3
>
>I'm not sure (yet) why the taint variables are not set to 1.
>
>> modules/include...................1..6
>> # Running under perl version 5.008002 for hpux
>> # Current time local: Thu Dec 4 11:10:58 2003
>> # Current time GMT: Thu Dec 4 19:10:58 2003
>>
>> [snip]
>>
>> # testing : /mod_perl mod_include test/
>> # expected: (?-xism:mod_perl mod_include test)
>> # received: <!--#include virtual="/includes/header.shtml?mod_perl mod_include
>> test" -->
>> #
>> # <!--#include virtual="/includes-registry/test.pl" -->
>> # Failed test 4 in modules/include.t at line 31 fail #2
>> #
>> # <!--#include virtual="/includes-registry/cgipm.pl" -->
>> #
>> # <p align=right>[<a href="../index.html">back</a>]</p>
>> #
>> # <!--#include virtual="/includes/footer.shtml" -->
>> # Failed test 5 in modules/include.t at line 31 fail #3
>> ok 3
>> # testing : /Hello World/
>> # expected: (?-xism:Hello World)
>> # received: <!--#include virtual="/includes/header.shtml?mod_perl mod_include
>> test" -->
>> #
>> # <!--#include virtual="/includes-registry/test.pl" -->
>> #
>> # <!--#include virtual="/includes-registry/cgipm.pl" -->
>> #
>> # <p align=right>[<a href="../index.html">back</a>]</p>
>> #
>> # <!--#include virtual="/includes/footer.shtml" -->
>> not ok 4
>> # testing : /cgi.pm/
>> # expected: (?-xism:cgi.pm)
>> # received: <!--#include virtual="/includes/header.shtml?mod_perl mod_include
>> test" -->
>> #
>> # <!--#include virtual="/includes-registry/test.pl" -->
>> #
>> # <!--#include virtual="/includes-registry/cgipm.pl" -->
>> #
>> # <p align=right>[<a href="../index.html">back</a>]</p>
>> #
>> # <!--#include virtual="/includes/footer.shtml" -->
>> not ok 5
>
>The test failures look to be originating from the include mechanism.
>There are some scripts in t/htdocs/includes-registry/, test.pl and
>cgipm.pl, which simply print out "Hello World\n" and "cgi.pm\n",
>respectively. The tests check for these strings, but does not find them,
>and thus fail. The above output is a little muddled, probably due to
>buffering.
>
>> modules/include2..................1..4
>> # Running under perl version 5.008002 for hpux
>> # Current time local: Thu Dec 4 11:10:59 2003
>> # Current time GMT: Thu Dec 4 19:10:59 2003
>> # Using Test.pm version 1.24
>>
>> [snip]
>>
>> # testing : /mod_perl/
>> # expected: (?-xism:mod_perl)
>> # received: use strict;
>> #
>> # #XXX: this test needs to be more robust.
>> # #various output buffers spread across multiple prints
>> # #more mod_include features mixed and checking that the output
>> # #is *exactly* what we expected, not just matching a few patterns.
>> #
>> # print "Content-type: text/html\n\n";
>> #
>> # my $r = shift;
>> #
>> # my $test_string = 'Perl-SSI';
>> #
>> # $r->subprocess_env->set(MY_TEST => $test_string);
>> #
>> # print <<EOF;
>> # Hello World from <!--#echo var="SERVER_ADMIN" --><br>
>> # Local date is <!--#echo var="DATE_LOCAL" --><br>
>> # Brought to you by <!--#echo var="MY_TEST" -->
>> # <!--#echo var="SERVER_SOFTWARE" -->
>> # <br>
>> # EOF
>> not ok 4
>
>Similar to the include.t test, something does not seem to be properly
>pulling in/handling .pl files.
--
Reporting bugs: http://perl.apache.org/bugs/
Mail list info: http://perl.apache.org/maillist/modperl.html
