Bug #84246 for Devel-REPL: issues with Term::ReadLine::Perl

Wed Mar 27 14:55:08 2013 gortan [...] cpan.org - Ticket created

When trying to install Devel::Repl version 1.3.16 on Gentoo amd64 Linux, cpanminus fails during test with: t/load_core.t ..... ok Unknown error Compilation failed in require at /usr/lib64/perl5/vendor_perl/5.12.4/Term/ReadLine/Perl.pm line 63. # Tests were run but no plan was declared and done_testing() was not seen. t/load_plugins.t .. When I manually execute the testcase: $ perl -Iblib/lib t/load_plugins.t I get: ok 1 - use Devel::REPL; ok 2 - plugin B::Concise loaded Cannot create second readline interface, falling back to dumb. ok 3 - plugin Colors loaded Cannot create second readline interface, falling back to dumb. ... ok 30 - plugin Turtles loaded not ok 31 - no (unexpected) warnings (via done_testing) # Failed test 'no (unexpected) warnings (via done_testing)' # at t/load_plugins.t line 60. 1..31 # Looks like you failed 1 test of 31.

Wed Mar 27 14:55:46 2013 gortan [...] cpan.org - Requestor GORTAN added

Wed Mar 27 16:06:18 2013 ether [...] cpan.org - Correspondence added

Subject:	Re: [rt.cpan.org #84246]
Date:	Wed, 27 Mar 2013 13:05:50 -0700
To:	Philipp Gortan via RT <bug-Devel-REPL [...] rt.cpan.org>
From:	Karen Etheridge <ether [...] cpan.org>

On Wed, Mar 27, 2013 at 02:55:10PM -0400, Philipp Gortan via RT wrote: Show quoted text

> When I manually execute the testcase: > $ perl -Iblib/lib t/load_plugins.t > > I get: > > ok 1 - use Devel::REPL; > ok 2 - plugin B::Concise loaded > Cannot create second readline interface, falling back to dumb. > ok 3 - plugin Colors loaded > Cannot create second readline interface, falling back to dumb. > ... > ok 30 - plugin Turtles loaded > not ok 31 - no (unexpected) warnings (via done_testing) > # Failed test 'no (unexpected) warnings (via done_testing)' > # at t/load_plugins.t line 60. > 1..31 > # Looks like you failed 1 test of 31.

Could you try this again, including -MCarp::Always in the perl command? (Also, you snipped the warning that caused that failing test.) thanks!

Wed Mar 27 16:06:21 2013 The RT System itself - Status changed from 'new' to 'open'

Wed Mar 27 17:31:05 2013 gortan [...] cpan.org - Correspondence added

On Wed Mar 27 16:06:18 2013, ETHER wrote: Show quoted text

> Could you try this again, including -MCarp::Always in the perl command? > (Also, you snipped the warning that caused that failing test.)

Sorry, I didn't see anything of value in the snipped lines ;-) I re-ran the command including -MCarp::Always, complete output is attached.

Subject:

load_plugins.out

Download load_plugins.out
application/octet-stream 3k

Message body not shown because it is not plain text.

Thu Mar 28 19:31:54 2013 ether [...] cpan.org - Ticket #84274: Ticket created

Subject:

issues with Term::ReadLine::Perl

Thu Mar 28 19:33:04 2013 ether [...] cpan.org - Ticket #84274: Dependency on ticket #84275 added

Thu Mar 28 19:35:57 2013 ether [...] cpan.org - Ticket #84274: Merged into ticket #84246

Thu Mar 28 19:35:57 2013 ether [...] cpan.org - Merged into ticket #84246

Thu Mar 28 19:36:47 2013 ether [...] cpan.org - Subject changed from (no value) to 'issues with Term::ReadLine::Perl'

Wed Apr 03 10:44:17 2013 dark3phoenix [...] gmail.com - Correspondence added

From:

dark3phoenix [...] gmail.com

Looks like the problem is that you can't open multiple instances of Term::ReadLine:PERL from the same object. I was able to correct this by changing the test to reuse the same Devel::REPL instance each time. This prevents the warning that causes the test to fail. I don't think this changes the spirit of the test, it's just a byproduct of how Term::ReadLine::Perl works.

Subject:

load_plugins.t

use strict; use warnings; use Test::More; use Test::Warnings; use_ok('Devel::REPL'); my @plugins = qw/ B::Concise Colors Commands Completion CompletionDriver::Globals CompletionDriver::INC CompletionDriver::Keywords CompletionDriver::LexEnv CompletionDriver::Methods CompletionDriver::Turtles DDC DDS DumpHistory FancyPrompt FindVariable History Interrupt LexEnv MultiLine::PPI Nopaste OutputCache PPI Packages Peek ReadLineHistory Refresh ShowClass Timing Turtles /; my $repl = Devel::REPL->new; for my $plugin_name (@plugins) { test_load_plugin($plugin_name, $repl); } sub test_load_plugin { my ($plugin_name, $repl) = @_; my $test_name = "plugin $plugin_name loaded"; eval "use Devel::REPL::Plugin::$plugin_name"; unless($@) { eval { $repl->load_plugin($plugin_name) }; ok(!$@, $test_name); } else { SKIP: { skip "could not eval plugin $plugin_name", 1; } } } done_testing;

Wed Apr 03 17:24:54 2013 ether [...] cpan.org - Correspondence added

RT-Send-CC:

dark3phoenix [...] gmail.com

Thanks, Christopher, I've released this patch as 1.003017! *crosses fingers*

Thu Apr 04 13:16:36 2013 gortan [...] cpan.org - Correspondence added 15 min

Thanks, Ether. If I now manually run the testcase, it works fine: $ cd ~/.cpanm/work/1365095439.20678/Devel-REPL-1.003017 $ perl -I blib/lib/ -I blib/arch/ t/load_plugins.t ... 1..31 so all is fine. However, if I use prove: $ prove -b t/load_core.t ..... ok t/load_plugins.t .. 1/? Unknown error Compilation failed in require at /usr/lib64/perl5/vendor_perl/5.12.4/Term/ReadLine/Perl.pm line 63. # Tests were run but no plan was declared and done_testing() was not seen. t/load_plugins.t .. Dubious, test returned 2 (wstat 512, 0x200) All 1 subtests passed Test Summary Report ------------------- t/load_plugins.t (Wstat: 512 Tests: 1 Failed: 0) Non-zero exit status: 2 Parse errors: No plan found in TAP output Files=2, Tests=28, 0 wallclock secs ( 0.03 usr 0.00 sys + 0.61 cusr 0.05 csys = 0.69 CPU) Result: FAIL Can you reproduce this phenomenon?

Fri Apr 05 14:04:33 2013 ether [...] cpan.org - Correspondence added

Subject:	Re: [rt.cpan.org #84246] issues with Term::ReadLine::Perl
Date:	Fri, 5 Apr 2013 11:04:20 -0700
To:	Philipp Gortan via RT <bug-Devel-REPL [...] rt.cpan.org>
From:	Karen Etheridge <ether [...] cpan.org>

On Thu, Apr 04, 2013 at 01:16:39PM -0400, Philipp Gortan via RT wrote: Show quoted text

> Thanks, Ether. If I now manually run the testcase, it works fine: > However, if I use prove: > ... > Can you reproduce this phenomenon?

Negative -- I have never been able to reproduce this, on either OSX or ubuntu linux. What happens if you do this manually (this is the line that is dying, in Term::ReadLine::Perl): perl -wle'eval {require Term::ReadLine::readline}; die $@ if $@'

Sun Apr 07 16:59:36 2013 gortan [...] cpan.org - Correspondence added 15 min

On Fri Apr 05 14:04:33 2013, ETHER wrote: Show quoted text

> perl -wle'eval {require Term::ReadLine::readline}; die $@ if $@'

I think we have a clue: running the command as you specified: $ perl -wle'eval {require Term::ReadLine::readline}; die $@ if $@' works without an error. However, if I redirect stdout: $ perl -wle'eval {require Term::ReadLine::readline}; die $@ if $@' >/dev/null Unknown error Compilation failed in require at -e line 1.

Sun May 19 09:21:19 2013 mephinet [...] gmx.net - Correspondence added

From:

mephinet [...] gmx.net

On Sun Apr 07 16:59:36 2013, gortan@gmail.com wrote: Show quoted text

> However, if I redirect stdout: > $ perl -wle'eval {require Term::ReadLine::readline}; die $@ if $@'

> >/dev/null

> Unknown error > Compilation failed in require at -e line 1.

I think I found a work-around which makes the testsuite pass again: changing the definition of Devel::REPL's term attribute to lazy_build. has 'term' => (is => 'rw', lazy_build => 1); sub _build_term {Term::ReadLine->new('Perl REPL')}; Can you give this one a try please? PS: does it make sense to define an attribute as required _and_ specify a default?

Sun May 19 11:52:26 2013 ether [...] cpan.org - Correspondence added

Subject:	Re: [rt.cpan.org #84246] issues with Term::ReadLine::Perl
Date:	Sun, 19 May 2013 08:52:09 -0700
To:	"mephinet [...] gmx.net via RT" <bug-Devel-REPL [...] rt.cpan.org>
From:	Karen Etheridge <ether [...] cpan.org>

On Sun, May 19, 2013 at 09:21:21AM -0400, mephinet@gmx.net via RT wrote: Show quoted text

> I think I found a work-around which makes the testsuite pass again: changing the definition of Devel::REPL's term attribute to lazy_build.

Hmmm interesting! why would delaying the building of the term object make a difference? But it's certainly worth a try. I've also poked the p5p folks to release the newest Term::ReadLine to cpan -- the version in 5.18.0 is two versions ahead of what's on PAUSE. Show quoted text

> PS: does it make sense to define an attribute as required _and_ specify a default?

No, if both fields are defined, 'required' is ignored.

Sun May 19 11:54:53 2013 devel.chm.01 [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #84246] issues with Term::ReadLine::Perl
Date:	Sun, 19 May 2013 11:54:41 -0400
To:	bug-Devel-REPL [...] rt.cpan.org
From:	Chris Marshall <devel.chm.01 [...] gmail.com>

It may be because Term::ReadLine::Perl can't instantiate more than one instance. -chm On Sun, May 19, 2013 at 11:52 AM, Karen Etheridge via RT < bug-Devel-REPL@rt.cpan.org> wrote: Show quoted text

> Queue: Devel-REPL > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=84246 > > > On Sun, May 19, 2013 at 09:21:21AM -0400, mephinet@gmx.net via RT wrote:

> > I think I found a work-around which makes the testsuite pass again:

> changing the definition of Devel::REPL's term attribute to lazy_build. > > Hmmm interesting! why would delaying the building of the term object make > a difference? But it's certainly worth a try. > > I've also poked the p5p folks to release the newest Term::ReadLine to cpan > -- the version in 5.18.0 is two versions ahead of what's on PAUSE. >

> > PS: does it make sense to define an attribute as required _and_ specify

> a default? > > No, if both fields are defined, 'required' is ignored. > >

Sun May 19 12:05:13 2013 ether [...] cpan.org - Correspondence added

Subject:	Re: [rt.cpan.org #84246] issues with Term::ReadLine::Perl
Date:	Sun, 19 May 2013 09:04:56 -0700
To:	"devel.chm.01 [...] gmail.com via RT" <bug-Devel-REPL [...] rt.cpan.org>
From:	Karen Etheridge <ether [...] cpan.org>

On Sun, May 19, 2013 at 11:54:54AM -0400, devel.chm.01@gmail.com via RT wrote: Show quoted text

> Queue: Devel-REPL > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=84246 > > > It may be because Term::ReadLine::Perl can't instantiate more than one > instance. -chm

Ok, Devel-REPL-1.003018 has just been uploaded... let another round of failing cpantesters reports begin :D

Thu May 23 06:44:43 2013 gortan [...] cpan.org - Correspondence added

On Sun May 19 12:05:13 2013, ETHER wrote: Show quoted text

> Ok, Devel-REPL-1.003018 has just been uploaded... let another round of > failing cpantesters reports begin :D

Works for me, thanks! And the cpantester reports are looking good, too...

Thu May 23 06:44:45 2013 gortan [...] cpan.org - Fixed in 1.003018 added

Fri Jun 14 22:41:34 2013 ether [...] cpan.org - Correspondence added

RT-Send-CC:

devel.chm.01 [...] gmail.com, mephinet [...] gmx.net, dark3phoenix [...] gmail.com

On 2013-05-23 03:44:43, gortan@gmail.com wrote: Show quoted text

> On Sun May 19 12:05:13 2013, ETHER wrote: >

> > Ok, Devel-REPL-1.003018 has just been uploaded... let another round of > > failing cpantesters reports begin :D

> > Works for me, thanks! And the cpantester reports are looking good, too...

If anyone is still having issues, please see: Show quoted text

> From: Rocky Bernstein <rocky@cpan.org> > > I've just released > Term:ReadLine::Perl5<https://metacpan.org/module/ROCKY/Term-ReadLine-Perl5-1.08/ > lib/Term/ReadLine/Perl5.pm> > and > things don't > fail<http://www.cpantesters.org/distro/T/Term-ReadLine-Perl5.html#Term-ReadLine- > Perl5-1.08>to > the point where I can start considering outstanding bugs. (Slaven's > failures don't count.) > > You mentioned that Term::ReadLine::Perl messes up Devel::REPL. (I'm a > casual user of Devel::REPL and even have a Devel::Trepan > plugin<https://metacpan.org/module/ROCKY/Devel-Trepan-Shell-1.4/lib/Devel/Trepan > /Shell.pm>for > it). > > If you could file an github > issue<https://github.com/rocky/p5-Term-ReadLine-Perl5/issues>with > enough detail to explain what's wrong, ideally paring things down if > possible, I'll look at it when I get a chance. (Please *do not *use the RT > tracker for this project and if I knew how to disable it I would.)

Sat Jun 15 02:01:24 2013 Rocky Bernstein - Correspondence added

On Fri Jun 14 22:41:34 2013, ETHER wrote: Show quoted text

> On 2013-05-23 03:44:43, gortan@gmail.com wrote:

> > On Sun May 19 12:05:13 2013, ETHER wrote: > >

> > > Ok, Devel-REPL-1.003018 has just been uploaded... let another

> round of

> > > failing cpantesters reports begin :D

> > > > Works for me, thanks! And the cpantester reports are looking good,

> too... > > > If anyone is still having issues, please see: >

> > From: Rocky Bernstein <rocky@cpan.org> > > > > I've just released > > Term:ReadLine::Perl5<https://metacpan.org/module/ROCKY/Term-

> ReadLine-Perl5-1.08/

> > lib/Term/ReadLine/Perl5.pm> > > and > > things don't > > fail<http://www.cpantesters.org/distro/T/Term-ReadLine-

> Perl5.html#Term-ReadLine-

> > Perl5-1.08>to > > the point where I can start considering outstanding bugs. (Slaven's > > failures don't count.) > > > > You mentioned that Term::ReadLine::Perl messes up Devel::REPL. (I'm

> a

> > casual user of Devel::REPL and even have a Devel::Trepan > > plugin<https://metacpan.org/module/ROCKY/Devel-Trepan-Shell-

> 1.4/lib/Devel/Trepan

> > /Shell.pm>for > > it). > > > > If you could file an github > > issue<https://github.com/rocky/p5-Term-ReadLine-Perl5/issues>with > > enough detail to explain what's wrong, ideally paring things down if > > possible, I'll look at it when I get a chance. (Please *do not *use

> the RT

> > tracker for this project and if I knew how to disable it I would.)

>

An important clarification. I do NOT work on Term::ReadLine::Perl but Term::ReadLine::Perl5. So if you are using this and have a problem, then by all means open an issue.