Andy Dougherty schrieb:
> On Mon, 14 Jul 2008, Reini Urban wrote:
>
>> Avoid to load a possibly binary incompatible Term::ReadLine::Gnu
>>
>> leading to
>> not ok 1 - The ${main::_<filename} variable in the debugger was not destroyed
>> # Failed at lib/perl5db.t line 71
>> # got 'main::(../lib/perl5db/t/eval-line-bug:6):
>> # 6: my $i = 5;
>> # support available.
>> #
>> # Enter h or `h h' for help, or `man perldebug' for more help.
>> #
>> # Signal SEGV at
>> /usr/lib/perl5/vendor_perl/5.11/i686-debug-cygwin/Term/ReadLine/Gnu.pm line
>> 100
>> # require Term/ReadLine/Gnu.pm called at (eval
>> 4)[../lib/Term/ReadLine.pm:320] line 1
>> # Term::ReadLine::BEGIN() called at
>> /usr/lib/perl5/vendor_perl/5.11/i686-debug-cygwin/Term/ReadLine/Gnu.pm line 0
>> # eval {...} called at
>> /usr/lib/perl5/vendor_perl/5.11/i686-debug-cygwin/Term/ReadLine/Gnu.pm line 0
>> # eval 'use Term::ReadLine::Gnu; 1
>> # ;' called at ../lib/Term/ReadLine.pm line 320
>> # require Term/ReadLine.pm called at ../lib/perl5db.pl line 5957
>> # eval {...} called at ../lib/perl5db.pl line 5957
>> # DB::setterm called at ../lib/perl5db.pl line 2229
>> # DB::DB called at ../lib/perl5db/t/eval-line-bug line 6
>> # '
>> # expected /(?-xism:sub factorial)/
>
> But doesn't this just paper over the problem? If the user then installs
> the version of perl being installed and tries to run the debugger, won't
> this exact same problem arise?
No. First core is tested, then installed, then the cpan modules updated.
The problem is the incompatibility new core against old cpan package
which breaks the core test.
My fix is just for make test,
and just ignores the Term::ReadLine::Gnu module,
because it is not on core and should not be tested.
Then in real life the available Term::ReadLine::* modules are used as
fit, which means Gnu comes first.
--
Reini Urban
http://phpwiki.org/ http://murbreak.at/
Thread Previous
|
Thread Next