Re: [Mod-security-developers] [Vote] 2.7.5

[Breno S. <bre...@gm...>]

I will update this rx test

Thanks

Breno


On Tue, Jul 16, 2013 at 5:15 PM, Rainer Jung <rai...@ki...>wrote:

> On 16.07.2013 21:05, Christian Folini wrote:
> > Hello Rainer,
> >
> > On Tue, Jul 16, 2013 at 01:05:08PM +0200, Rainer Jung wrote:
> >> Hi Christian,
> >>
> >> I think when running the unit tests via "make test" you should add
> >> -DMSC_TEST to CFLAGS. That flag comments all httpd calls and produces
> >> standalone binaries.
> >
> > Thanks for the info. True, it's mentioned in the reference guide.
> > It's just that I used outdated documentation.
> >
> > Here we have the culprit:
> > $> make CFLAGS=-DMSC_TEST test
> > ...
> > Loaded 8 tests from ./op/rx.t
> >      1) op "rx": passed (Pattern match "" at UNIT_TEST.)
> >      2) op "rx": passed
> >      3) op "rx": passed (Pattern match "" at UNIT_TEST.)
> >      4) op "rx": passed (Pattern match "abc" at UNIT_TEST.)
> >      5) op "rx": passed (Pattern match "def" at UNIT_TEST.)
> >      6) op "rx": passed (Pattern match "ghi" at UNIT_TEST.)
> >      7) op "rx": passed
> > ERROR: Failed to create rule for op "rx": Error creating rule: Error
> compiling pattern (offset 2): unrecognized character after (? or (?-
> > Test exited with signal 11.
> > Executed: ./msc_test "-t" "op" "-n" "rx" "-p"
> "(?^i:^([^=])\s*=\s*((?:abc)+(?:def|ghi){2})$)" "-D" "0" "-r" "1"
> >      8) op "rx": failed
> > Passed:  7; Failed:  1
> > ...
> > 576/577 tests passed.
> > FAIL: run-unit-tests.pl
> > ========================================
> > 1 of 1 test failed
> > Please report to su...@mo...
> > ========================================
> > make[2]: *** [check-TESTS] Error 1
> > make[2]: Leaving directory
> `/usr/src/modsecurity/ModSecurity-remotes-2.7.x/tests'
> > make[1]: *** [check-am] Error 2
> > make[1]: Leaving directory
> `/usr/src/modsecurity/ModSecurity-remotes-2.7.x/tests'
> > make: *** [check-recursive] Error 1
>
> I was curious why this didn't fail for me. I vaguely remember other
> users had reported the same problem before. So here's what I think is
> the explanation:
>
> The original pattern in the unit test is in op/rx.t:
>
>     qr/^([^=])\s*=\s*((?:abc)+(?:def|ghi){2})$/i
>
> The pattern is read by the perl script run-unit-tests.pl. Before Perl
> 5.14 the qr was converted to:
>
> (?i-xsm:^([^=])\s*=\s*((?:abc)+(?:def|ghi){2})$)
>
> Starting with Perl 5.14 there's a new syntax and the pattern results in
>
> (?^i:^([^=])\s*=\s*((?:abc)+(?:def|ghi){2})$)
>
> (See e.g. http://perldoc.perl.org/perlre.html#Extended-Patterns, there
> look for "Starting in Perl 5.14").
>
> Now the unit tests feed the internal representation to the compiled
> msc_test binary, which calls the PCRE library to handle the regexp.
> Unfortunately the PCRE library only accepts the old way of expressing
> this pattern, not the new ?^ syntax.
>
> So if you run the unit tests with a Perl before 5.14 they succeed, with
> a newer Perl they fail, because a Perl pre-compiled pattern might no
> longer be compatible with PCRE (despite what PCRE means).
>
> I didn't find a bug report for the missing ?^ syntax in the PCRE bug
> tracker, but "?^" is not the best token to start a search with.
>
> Regards,
>
> Rainer
>
>
> ------------------------------------------------------------------------------
> See everything from the browser to the database with AppDynamics
> Get end-to-end visibility with application monitoring from AppDynamics
> Isolate bottlenecks and diagnose root cause in seconds.
> Start your free trial of AppDynamics Pro today!
> http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
> _______________________________________________
> mod-security-developers mailing list
> mod...@li...
> https://lists.sourceforge.net/lists/listinfo/mod-security-developers
> ModSecurity Services from Trustwave's SpiderLabs:
> https://www.trustwave.com/spiderLabs.php
>