Bug #102525 for Filter-Crypto: Tests not passing

Thu Mar 05 08:38:11 2015 ZDM [...] cpan.org - Ticket created

Subject:

Tests not passing

Hello. perl 5.20.2 x64 on centos 7 x64. Pod::Simple::Text and Pod::Perldoc::ToText are installed. Don't know, why builder can't find and load this modules. Does this POD tests really needed in run-time environment? Here is the output: cpanm Filter::Crypto -v ...... PERL_DL_NONLAZY=1 "/usr/local/perl-5.20.2/bin/perl" "-MExtUtils::Command::MM" "-MTest::Harness" "-e" "undef *Test::Harness::Switches; test_harness(0, 'blib/lib', 'blib/arch')" t/*.t t/01_imports_01.t .... ok t/01_imports_02.t .... ok t/01_imports_03.t .... ok t/01_imports_04.t .... ok t/02_function.t ...... ok t/03_script.t ........ 97/105 Can't find any loadable formatter class in Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod?! Aborting at /usr/local/perl-5.20.2/bin/perldoc line 6. # Failed test '-m option works' # at t/03_script.t line 672. # '' # doesn't match '(?^msx:^ NAME .*? # ^ SYNOPSIS .*? # ^ ARGUMENTS .*? # ^ OPTIONS .*? # ^ EXIT\ STATUS .*? # ^ DIAGNOSTICS .*? # ^ EXAMPLES .*? # ^ ENVIRONMENT .*? # ^ SEE\ ALSO .*? # ^ AUTHOR .*? # ^ COPYRIGHT .*? # ^ LICENCE .*? # ^ VERSION .*? # ^ DATE .*? # ^ HISTORY )' # Looks like you failed 1 test of 105. t/03_script.t ........ Dubious, test returned 1 (wstat 256, 0x100) Failed 1/105 subtests t/04_par.t ........... ok t/05_errstr.t ........ ok t/06_changes.t ....... skipped: Author testing only t/07_critic.t ........ skipped: Author testing only t/08_pod.t ........... skipped: Author testing only t/09_pod_coverage.t .. skipped: Author testing only

Wed Mar 11 05:25:37 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 09:25:26 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 5 March 2015 at 13:38, Dmytro Zagashev via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Thu Mar 05 08:38:11 2015: Request 102525 was acted upon. > Transaction: Ticket created by ZDM > Queue: Filter-Crypto > Subject: Tests not passing > Broken in: (no value) > Severity: (no value) > Owner: Nobody > Requestors: ZDM@cpan.org > Status: new > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > > Hello. > > perl 5.20.2 x64 on centos 7 x64. > > Pod::Simple::Text and Pod::Perldoc::ToText are installed. Don't know, why builder can't find and load this modules. > > Does this POD tests really needed in run-time environment? >

It could be skipped unless AUTHOR_TESTING is true, but I'd really like to understand why it has failed if possible. Please could you try the following and send me the output?: Set PERLDOC=-t and PERLDOCDEBUG=5 in your environment, and then run "perldoc script/crypt_file" from the root of the Filter-Crypto directory. That should print the man page, with some verbose debugging output from perldoc. Thank you.

Wed Mar 11 05:25:37 2015 The RT System itself - Status changed from 'new' to 'open'

Wed Mar 11 05:33:25 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 11:33:07 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

[root@mediaspider /mnt]$ cpanm --look Filter::Crypto --> Working on Filter::Crypto Fetching http://www.cpan.org/authors/id/S/SH/SHAY/Filter-Crypto-2.07.tar.gz ... OK Entering /root/.cpanm/work/1426066256.6600/Filter-Crypto-2.07 with /bin/bash [root@mediaspider ~/.cpanm/work/1426066256.6600/Filter-Crypto-2.07]$ export PERLDOC=-t [root@mediaspider ~/.cpanm/work/1426066256.6600/Filter-Crypto-2.07]$ export PERLDOCDEBUG=5 [root@mediaspider ~/.cpanm/work/1426066256.6600/Filter-Crypto-2.07]$ perldoc script/crypt_file Parameters to Pod::Perldoc->run: New Pod::Perldoc object Pod::Perldoc=HASH(0xeae018) Considering new formatter classes -MPod::Perldoc::ToPod Adding Pod::Perldoc::ToPod to the list of formatter classes, making them Pod::Perldoc::ToPod. 'opt_o_with: Noting "text" as desired output format... 'opt_o_with: Considering Pod::Perldoc::Totext 'opt_o_with: Considering Pod::Perldoc::Totext 'opt_o_with: Considering Pod::Perldoc::ToText 'opt_o_with: Considering Pod::Perldoc::ToTEXT 'opt_o_with: Considering Pod::Simple::text 'opt_o_with: Considering Pod::Simple::text 'opt_o_with: Considering Pod::Simple::Text 'opt_o_with: Considering Pod::Simple::TEXT 'opt_o_with: Considering Pod::text 'opt_o_with: Considering Pod::text 'opt_o_with: Considering Pod::Text 'opt_o_with: Considering Pod::TEXT Considering new formatter classes -MPod::Perldoc::Totext;Pod::Perldoc::Totext;Pod::Perldoc::ToText;Pod::Perldoc::ToTEXT;Pod::Simple::text;Pod::Simple::text;Pod::Simple::Text;Pod::Simple::TEXT;Pod::text;Pod::text;Pod::Text;Pod::TEXT Adding Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT to the list of formatter classes, making them Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod. 'opt_o_with: Noting "term" as desired output format... 'opt_o_with: Considering Pod::Perldoc::Toterm 'opt_o_with: Considering Pod::Perldoc::Toterm 'opt_o_with: Considering Pod::Perldoc::ToTerm 'opt_o_with: Considering Pod::Perldoc::ToTERM 'opt_o_with: Considering Pod::Simple::term 'opt_o_with: Considering Pod::Simple::term 'opt_o_with: Considering Pod::Simple::Term 'opt_o_with: Considering Pod::Simple::TERM 'opt_o_with: Considering Pod::term 'opt_o_with: Considering Pod::term 'opt_o_with: Considering Pod::Term 'opt_o_with: Considering Pod::TERM Considering new formatter classes -MPod::Perldoc::Toterm;Pod::Perldoc::Toterm;Pod::Perldoc::ToTerm;Pod::Perldoc::ToTERM;Pod::Simple::term;Pod::Simple::term;Pod::Simple::Term;Pod::Simple::TERM;Pod::term;Pod::term;Pod::Term;Pod::TERM Adding Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM to the list of formatter classes, making them Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod. Formatter switches now: [[__bindir /usr/local/perl-5.20.2/bin] [__pod2man pod2man]] Beginning process. Args: script/crypt_file Object contents: [temp_file_list] => [] [bindir] => [/usr/local/perl-5.20.2/bin] [pagers] => [/usr/bin/less -R] [args] => [script/crypt_file] [formatter_classes] => [Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod] [translators] => [] [pod2man] => [pod2man] [target] => [<undef>] [extra_search_dirs] => [] [found] => [] [formatter_switches] => [ARRAY(0x103d938) ARRAY(0x121b558)] [search_path] => [] 'options_reading: Noting env PERLDOC setting of -t Args now: -t script/crypt_file Args right before switch processing: -t script/crypt_file 'GetOptsOO::getopts: Starting switch processing. Scanning arguments [-t script/crypt_file] Truth is YES opt_t is true (YES) 'opt_o_with: Noting "text" as desired output format... 'opt_o_with: Considering Pod::Perldoc::Totext 'opt_o_with: Considering Pod::Perldoc::Totext 'opt_o_with: Considering Pod::Perldoc::ToText 'opt_o_with: Considering Pod::Perldoc::ToTEXT 'opt_o_with: Considering Pod::Simple::text 'opt_o_with: Considering Pod::Simple::text 'opt_o_with: Considering Pod::Simple::Text 'opt_o_with: Considering Pod::Simple::TEXT 'opt_o_with: Considering Pod::text 'opt_o_with: Considering Pod::text 'opt_o_with: Considering Pod::Text 'opt_o_with: Considering Pod::TEXT Considering new formatter classes -MPod::Perldoc::Totext;Pod::Perldoc::Totext;Pod::Perldoc::ToText;Pod::Perldoc::ToTEXT;Pod::Simple::text;Pod::Simple::text;Pod::Simple::Text;Pod::Simple::TEXT;Pod::text;Pod::text;Pod::Text;Pod::TEXT Adding Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT to the list of formatter classes, making them Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod. 'GetOptsOO::getopts: Ending switch processing. Args are [script/crypt_file] with 0 errors. Args after switch processing: script/crypt_file 'pagers_guessing: Pagers: /usr/bin/less -Rmorelesspgviewcat'process: /usr/local/perl-5.20.2/bin/perldoc => Pod::Perldoc v3.25 Attempting to drop privs... OK, I dropped privileges. 'process: Trying to load Pod::Perldoc::Totext... Trying to eval 'require Pod::Perldoc::Totext'... Couldn't load Pod::Perldoc::Totext: No such file or directory Trying to load Pod::Perldoc::Totext... I've tried Pod::Perldoc::Totext before, and it's no good. Skipping. Trying to load Pod::Perldoc::ToText... Trying to eval 'require Pod::Perldoc::ToText'... Settling on Pod::Perldoc::ToText 'find_good_formatter_class: Formatter class Pod::Perldoc::ToText version 3.25 successfully loaded! 'find_good_formatter_class: Will format with the class Pod::Perldoc::ToText 'grand_search_init: Searching for script/crypt_file 'grand_search_init: Found as script/crypt_file Found: [script/crypt_file] That's a non-dynamic pod search. Formatter switches now: [[__bindir /usr/local/perl-5.20.2/bin] [__pod2man pod2man]] Formatter Pod::Perldoc::ToText doesn't support __bindir Formatter Pod::Perldoc::ToText doesn't support __pod2man About to call new Pod::Text (v3.18) with options: (nil) Back from formatting with Pod::Perldoc::ToText script/crypt_file isn't a temp file, so not unlinking. Finished writing to /tmp/dPb6uV0xh6. 'page: About to try calling /usr/bin/less -R /tmp/dPb6uV0xh6 On 11.03.2015 11:25, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> perldoc script/crypt_file

Wed Mar 11 09:31:47 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 13:31:38 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 11 March 2015 at 09:33, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > [root@mediaspider /mnt]$ cpanm --look Filter::Crypto

[...] Show quoted text

> Settling on Pod::Perldoc::ToText > 'find_good_formatter_class: Formatter class Pod::Perldoc::ToText version > 3.25 successfully loaded! > 'find_good_formatter_class: Will format with the class Pod::Perldoc::ToText > 'grand_search_init: Searching for script/crypt_file > 'grand_search_init: Found as script/crypt_file > Found: [script/crypt_file] > That's a non-dynamic pod search. > Formatter switches now: [[__bindir /usr/local/perl-5.20.2/bin] > [__pod2man pod2man]] > Formatter Pod::Perldoc::ToText doesn't support __bindir > Formatter Pod::Perldoc::ToText doesn't support __pod2man > About to call new Pod::Text (v3.18) with options: (nil) > Back from formatting with Pod::Perldoc::ToText > script/crypt_file isn't a temp file, so not unlinking. > Finished writing to /tmp/dPb6uV0xh6. > 'page: About to try calling /usr/bin/less -R /tmp/dPb6uV0xh6 > >

Thanks. That output looks virtually identical to the output on my own (Windows) machine, other than a few OS-specific differences. I'm using the same version of Pod::Perldoc::ToText / Pod::Text as you too. Did this actually work this time for you? There is no sign of the "Can't find any loadable formatter class" error that you had originally.

Wed Mar 11 09:39:10 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 15:38:51 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Message body is not shown because it is too large.

Wed Mar 11 10:05:09 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 14:04:59 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 11 March 2015 at 13:39, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I can install Filter::Crypto only with --force, which isn't suitable for > automated deployment. > > Here is the full log: > > [root@mediaspider /home/zdm]$ cpanm Filter::Crypto --reinstall -v

[...] So is the problem that the test succeeds if you run it manually (as you did with PERLDOCDEBUG=5 earlier), but fails when the tests are run by cpanm? I don't currently have cpanm, but I will try it out if that's the case. I generally just use the standard 'cpan' tool that comes with perl. "install Filter::Crypto" in the standard cpan shell works fine for me.

Wed Mar 11 10:11:30 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 16:11:16 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

There is no differences between cpan and cpanm. Both just download archive and run standard commands. I try to install manually, download and unpack tar and perform: perl Makefile.PL make make test and error still occurs. On 11.03.2015 16:05, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 13:39, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> I can install Filter::Crypto only with --force, which isn't suitable for >> automated deployment. >> >> Here is the full log: >> >> [root@mediaspider /home/zdm]$ cpanm Filter::Crypto --reinstall -v

> [...] > > So is the problem that the test succeeds if you run it manually (as > you did with PERLDOCDEBUG=5 earlier), but fails when the tests are run > by cpanm? > > I don't currently have cpanm, but I will try it out if that's the > case. I generally just use the standard 'cpan' tool that comes with > perl. "install Filter::Crypto" in the standard cpan shell works fine > for me. >

Wed Mar 11 10:16:52 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 14:16:42 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 11 March 2015 at 14:11, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > There is no differences between cpan and cpanm. Both just download > archive and run standard commands. > > I try to install manually, download and unpack tar and perform: > > perl Makefile.PL > make > make test > > and error still occurs. >

Do you still get the error when running "perldoc crypt/script_file" manually with PERLDOC=-t ? It looked like that had worked ok when you tried it before with PERLDOCDEBUG=5.

Wed Mar 11 10:34:30 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 16:34:15 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

No errors when run perldoc crypt/script_file On 11.03.2015 16:16, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 14:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> There is no differences between cpan and cpanm. Both just download >> archive and run standard commands. >> >> I try to install manually, download and unpack tar and perform: >> >> perl Makefile.PL >> make >> make test >> >> and error still occurs. >>

> Do you still get the error when running "perldoc crypt/script_file" > manually with PERLDOC=-t ? It looked like that had worked ok when you > tried it before with PERLDOCDEBUG=5. >

Wed Mar 11 10:36:42 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 16:36:28 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

This error only appears during make test. And prevent to install package. On 11.03.2015 16:16, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 14:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> There is no differences between cpan and cpanm. Both just download >> archive and run standard commands. >> >> I try to install manually, download and unpack tar and perform: >> >> perl Makefile.PL >> make >> make test >> >> and error still occurs. >>

> Do you still get the error when running "perldoc crypt/script_file" > manually with PERLDOC=-t ? It looked like that had worked ok when you > tried it before with PERLDOCDEBUG=5. >

Wed Mar 11 10:50:07 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 16:49:49 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Also, as I correctly understand, there is nothing that can prevent somebody to manually pass encrypted sources via Filter::Crypto::Decrypt (that packed in PAR) and get decrypted data? On 11.03.2015 16:16, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 14:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> There is no differences between cpan and cpanm. Both just download >> archive and run standard commands. >> >> I try to install manually, download and unpack tar and perform: >> >> perl Makefile.PL >> make >> make test >> >> and error still occurs. >>

> Do you still get the error when running "perldoc crypt/script_file" > manually with PERLDOC=-t ? It looked like that had worked ok when you > tried it before with PERLDOCDEBUG=5. >

Wed Mar 11 14:12:45 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 18:12:35 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 11 March 2015 at 14:36, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > This error only appears during make test. > > And prevent to install package. >

Please could you run the attached test.t from the root Filter-Crypto folder and send the output? I'm just trying to figure out why "$^X script/crypt_file -m" fails (in make test), but a simple "perldoc script/crypt_file" works...

Message body is not shown because sender requested not to inline it.

Wed Mar 11 14:18:06 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 18:17:57 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 11 March 2015 at 14:50, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Also, as I correctly understand, there is nothing that can prevent > somebody to manually pass encrypted sources via Filter::Crypto::Decrypt > (that packed in PAR) and get decrypted data? >

That's correct. At the end of the day, the encrypted source file has to be decrypted somewhere in memory before being passed to the Perl interpreter since that (obviously) can't interpret encrypted source code. There are warnings in the man pages about this. Using a filter mechanism like this does make it harder to get at the decrypted data than other mechanisms would, though, and static-linking the Decrypt DLL (or whatever it is on your OS) makes it a little more robust too. The main thing is not to distribute the CryptFile DLL with your encrypted source files since that makes it trivial to decrypt them.

Wed Mar 11 14:26:47 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 20:26:32 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

From the root of unpacked distribution? On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 14:36, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> This error only appears during make test. >> >> And prevent to install package. >>

> Please could you run the attached test.t from the root Filter-Crypto > folder and send the output? > > I'm just trying to figure out why "$^X script/crypt_file -m" fails (in > make test), but a simple "perldoc script/crypt_file" works... >

Wed Mar 11 14:38:19 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 20:38:03 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Seems I understand what happens. You have shebang #!perl, but more wide used practice (and more correct) is to use #!/usr/bin/env perl. In my case invalid perl interpreter used. On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 11 March 2015 at 14:36, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> This error only appears during make test. >> >> And prevent to install package. >>

> Please could you run the attached test.t from the root Filter-Crypto > folder and send the output? > > I'm just trying to figure out why "$^X script/crypt_file -m" fails (in > make test), but a simple "perldoc script/crypt_file" works... >

Wed Mar 11 14:52:41 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-03-11 14:38:19, dzagashev@gmail.com wrote: Show quoted text

> Seems I understand what happens. > > You have shebang #!perl, but more wide used practice (and more correct) > is to use #!/usr/bin/env perl.

Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. Show quoted text

> > In my case invalid perl interpreter used. > > On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

> > <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > > > On 11 March 2015 at 14:36, dzagashev@gmail.com via RT > > <bug-Filter-Crypto@rt.cpan.org> wrote:

> >> Queue: Filter-Crypto > >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > >> > >> This error only appears during make test. > >> > >> And prevent to install package. > >>

> > Please could you run the attached test.t from the root Filter-Crypto > > folder and send the output? > > > > I'm just trying to figure out why "$^X script/crypt_file -m" fails (in > > make test), but a simple "perldoc script/crypt_file" works... > >

> >

Wed Mar 11 15:07:35 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 11 Mar 2015 21:07:18 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

I have two interpreters on this server. Standard centos perl 5.14. and working 5.20. Working interpreter is in path before /usr/bin/. So, shebang #!perl should work correctly, my mistake. No, it not works at all. I run small script: #!perl print $^X; and get error - bash: ./1.pl: perl: bad interpreter: No such file or directory On 11.03.2015 20:52, Slaven_Rezic via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>> Seems I understand what happens. >> >> You have shebang #!perl, but more wide used practice (and more correct) >> is to use #!/usr/bin/env perl.

> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >

>> In my case invalid perl interpreter used. >> >> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> This error only appears during make test. >>>> >>>> And prevent to install package. >>>>

>>> Please could you run the attached test.t from the root Filter-Crypto >>> folder and send the output? >>> >>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>> make test), but a simple "perldoc script/crypt_file" works... >>>

>>

> >

Fri Mar 13 06:16:17 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Fri, 13 Mar 2015 12:15:58 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

So, this is not fixable? Other packages, based on ExtUtils::MakeMaker, installed without problems. On 11.03.2015 20:52, Slaven_Rezic via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>> Seems I understand what happens. >> >> You have shebang #!perl, but more wide used practice (and more correct) >> is to use #!/usr/bin/env perl.

> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >

>> In my case invalid perl interpreter used. >> >> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> This error only appears during make test. >>>> >>>> And prevent to install package. >>>>

>>> Please could you run the attached test.t from the root Filter-Crypto >>> folder and send the output? >>> >>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>> make test), but a simple "perldoc script/crypt_file" works... >>>

>>

> >

Mon Mar 16 17:30:53 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Mon, 16 Mar 2015 21:30:41 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

(I just want to note that I haven't given up with this; I'm just tied up with other things at the moment. I will return to this soon, and your other rt.cpan.org (#102788). Thank you for your interest in this module.) On 13 March 2015 at 10:16, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > So, this is not fixable? > > Other packages, based on ExtUtils::MakeMaker, installed without problems. > > On 11.03.2015 20:52, Slaven_Rezic via RT wrote:

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>>> Seems I understand what happens. >>> >>> You have shebang #!perl, but more wide used practice (and more correct) >>> is to use #!/usr/bin/env perl.

>> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >>

>>> In my case invalid perl interpreter used. >>> >>> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>> Queue: Filter-Crypto >>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> This error only appears during make test. >>>>> >>>>> And prevent to install package. >>>>>

>>>> Please could you run the attached test.t from the root Filter-Crypto >>>> folder and send the output? >>>> >>>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>>> make test), but a simple "perldoc script/crypt_file" works... >>>>

>>>

>> >>

> > >

Tue Mar 17 18:18:10 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Wed, 18 Mar 2015 00:17:56 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Hi. Thanks for your work. Do you think about to encode encrypted file to base64 instead of hex? Size will be smaller about 30%. On 16.03.2015 23:30, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > (I just want to note that I haven't given up with this; I'm just tied > up with other things at the moment. I will return to this soon, and > your other rt.cpan.org (#102788). Thank you for your interest in this > module.) > > On 13 March 2015 at 10:16, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> So, this is not fixable? >> >> Other packages, based on ExtUtils::MakeMaker, installed without problems. >> >> On 11.03.2015 20:52, Slaven_Rezic via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>>>> Seems I understand what happens. >>>> >>>> You have shebang #!perl, but more wide used practice (and more correct) >>>> is to use #!/usr/bin/env perl.

>>> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >>>

>>>> In my case invalid perl interpreter used. >>>> >>>> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>> Queue: Filter-Crypto >>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> This error only appears during make test. >>>>>> >>>>>> And prevent to install package. >>>>>>

>>>>> Please could you run the attached test.t from the root Filter-Crypto >>>>> folder and send the output? >>>>> >>>>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>>>> make test), but a simple "perldoc script/crypt_file" works... >>>>>

>>>

>> >>

Sun Mar 22 09:41:15 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 13:41:04 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

Coming back to this now. Apologies for the delay... I'm still trying to understand why "perldoc script/crypt_file" works fine in the shell, but "$^X script/crypt_file -m" fails in the test script, given that the latter basically runs the former! We've wondered whther the shebang line in script/crypt_file is faulty, but it doesn't appear that way. It looks perfectly normal to write either #!perl or #!/usr/bin/perl. Also, the test script itself (t/03_script.t) has the same #!perl in it and runs fine, and the two previous tests in that script (which run "$^X script/crypt_file -v" and "$^X script/crypt_file -h") pass without trouble. Previous debug output has shown that you're running Pod::Usage 1.67, which means that the following code is being run for the -m option in script/crypt_file: my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; my $perldoc = catfile($script_dir, 'perldoc'); pod2usage( -exitval => 1, -verbose => 2, -noperldoc => MM->maybe_command($perldoc) ? 0 : 1 ); So, in order to see better exactly what is happening here, please could you put the following into a script, run it, and send the output here?: use strict; use warnings; use Config; use ExtUtils::MakeMaker; use File::Spec::Functions qw(catfile); my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; my $perldoc = catfile($script_dir, 'perldoc'); my $noperldoc = MM->maybe_command($perldoc) ? 0 : 1; print "script_dir=$script_dir\n"; print "perldoc=$perldoc\n"; print "noperldoc=$noperldoc\n"; Thank you. On 16 March 2015 at 21:30, Steve Hay <steve.m.hay@googlemail.com> wrote: Show quoted text

> (I just want to note that I haven't given up with this; I'm just tied > up with other things at the moment. I will return to this soon, and > your other rt.cpan.org (#102788). Thank you for your interest in this > module.) > > On 13 March 2015 at 10:16, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> So, this is not fixable? >> >> Other packages, based on ExtUtils::MakeMaker, installed without problems. >> >> On 11.03.2015 20:52, Slaven_Rezic via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>>>> Seems I understand what happens. >>>> >>>> You have shebang #!perl, but more wide used practice (and more correct) >>>> is to use #!/usr/bin/env perl.

>>> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >>>

>>>> In my case invalid perl interpreter used. >>>> >>>> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>> Queue: Filter-Crypto >>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> This error only appears during make test. >>>>>> >>>>>> And prevent to install package. >>>>>>

>>>>> Please could you run the attached test.t from the root Filter-Crypto >>>>> folder and send the output? >>>>> >>>>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>>>> make test), but a simple "perldoc script/crypt_file" works... >>>>>

>>>>

>>> >>>

>> >> >>

Sun Mar 22 09:46:19 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 13:46:10 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

I hadn't considered using base64, no :-/ There are doubtless other encodings that are even smaller too, but the trouble with changing it now is that users will have to re-encrypt any previously encrypted files if they upgrade their Filter-Crypto. So I'm not keen on doing it now unless there is a significant speed-up to be had. I will certainly consider it for the future, though, and definitely if there is ever a need to make some other incompatible change that would require files to be re-encrypted anyway. On 17 March 2015 at 22:18, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Hi. > > Thanks for your work. > > Do you think about to encode encrypted file to base64 instead of hex? > > Size will be smaller about 30%. > > On 16.03.2015 23:30, steve.m.hay@googlemail.com via RT wrote:

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> (I just want to note that I haven't given up with this; I'm just tied >> up with other things at the moment. I will return to this soon, and >> your other rt.cpan.org (#102788). Thank you for your interest in this >> module.) >> >> On 13 March 2015 at 10:16, dzagashev@gmail.com via RT >> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>> Queue: Filter-Crypto >>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> So, this is not fixable? >>> >>> Other packages, based on ExtUtils::MakeMaker, installed without problems. >>> >>> On 11.03.2015 20:52, Slaven_Rezic via RT wrote:

>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

>>>>> Seems I understand what happens. >>>>> >>>>> You have shebang #!perl, but more wide used practice (and more correct) >>>>> is to use #!/usr/bin/env perl.

>>>> Sorry, I have to disagree. Currently it's incorrect to specify #!/usr/bin/env perl for scripts which are installed normally via blib/scripts. The reason is that ExtUtils::MakeMaker's _fixin_replace_shebang does not work correctly otherwise --- it does not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl or so. >>>>

>>>>> In my case invalid perl interpreter used. >>>>> >>>>> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

>>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT >>>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>>> Queue: Filter-Crypto >>>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>>> >>>>>>> This error only appears during make test. >>>>>>> >>>>>>> And prevent to install package. >>>>>>>

>>>>>> Please could you run the attached test.t from the root Filter-Crypto >>>>>> folder and send the output? >>>>>> >>>>>> I'm just trying to figure out why "$^X script/crypt_file -m" fails (in >>>>>> make test), but a simple "perldoc script/crypt_file" works... >>>>>>

>>>>

>>> >>>

> > >

Sun Mar 22 11:35:56 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 17:35:44 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Hi. with !#perl shebang: bash: /var/local/1.pl: perl: bad interpreter: No such file or directory with #!/usr/bin/env perl shebang: script_dir=/usr/local/perl-5.20.2/bin perldoc=/usr/local/perl-5.20.2/bin/perldoc noperldoc=0 On 22.03.2015 15:41, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> use strict; > use warnings; > use Config; > use ExtUtils::MakeMaker; > use File::Spec::Functions qw(catfile); > my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; > my $perldoc = catfile($script_dir, 'perldoc'); > my $noperldoc = MM->maybe_command($perldoc) ? 0 : 1; > print "script_dir=$script_dir\n"; > print "perldoc=$perldoc\n"; > print "noperldoc=$noperldoc\n";

Sun Mar 22 11:41:04 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 17:40:54 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Download (untitled)
application/pgp-encrypted 11b

Message body not shown because it is not plain text.

Download encrypted.asc
application/octet-stream 1.6k

Message body not shown because it is not plain text.

Sun Mar 22 11:43:43 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 17:43:33 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Maybe add additional option to crypt_file function? Or even add new function crypt_file_b64? This is not so important, but will be useful. On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> I hadn't considered using base64, no :-/ There are doubtless other > encodings that are even smaller too, but the trouble with changing it > now is that users will have to re-encrypt any previously encrypted > files if they upgrade their Filter-Crypto. So I'm not keen on doing it > now unless there is a significant speed-up to be had. I will certainly > consider it for the future, though, and definitely if there is ever a > need to make some other incompatible change that would require files > to be re-encrypted anyway.

Sun Mar 22 12:08:26 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 16:08:16 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

Ok. It still seems weird to me that scripts with the #!perl shebang line are not running on your system, especially given that all the t/*.t scripts have that in them! My system's perldoc script also has #!perl in it. Does yours too? How did you run that test script (1.pl)? Did you run "perl 1.pl" or just "1.pl"? Please could you also try this for me?: Replace the following line: chomp($data = qx{$perl $crypt_file -m}); near the end of t/03_script.t with this: chomp($data = qx{/usr/local/perl-5.20.2/bin/perldoc $crypt_file}); and then try running "make test" again. On 22 March 2015 at 15:35, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Hi. > > with !#perl shebang: > > bash: /var/local/1.pl: perl: bad interpreter: No such file or directory > > with #!/usr/bin/env perl shebang: > > script_dir=/usr/local/perl-5.20.2/bin > perldoc=/usr/local/perl-5.20.2/bin/perldoc > noperldoc=0 > > > On 22.03.2015 15:41, steve.m.hay@googlemail.com via RT wrote:

>> use strict; >> use warnings; >> use Config; >> use ExtUtils::MakeMaker; >> use File::Spec::Functions qw(catfile); >> my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; >> my $perldoc = catfile($script_dir, 'perldoc'); >> my $noperldoc = MM->maybe_command($perldoc) ? 0 : 1; >> print "script_dir=$script_dir\n"; >> print "perldoc=$perldoc\n"; >> print "noperldoc=$noperldoc\n";

> >

Sun Mar 22 12:09:43 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 16:09:35 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

Sure, I will certainly look into adding an option to use base64 encoded files in the future. Thanks for the suggestion. On 22 March 2015 at 15:43, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Maybe add additional option to crypt_file function? > > Or even add new function crypt_file_b64? > > This is not so important, but will be useful. > > On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>> I hadn't considered using base64, no :-/ There are doubtless other >> encodings that are even smaller too, but the trouble with changing it >> now is that users will have to re-encrypt any previously encrypted >> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >> now unless there is a significant speed-up to be had. I will certainly >> consider it for the future, though, and definitely if there is ever a >> need to make some other incompatible change that would require files >> to be re-encrypted anyway.

>

Sun Mar 22 12:16:02 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 18:15:51 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Or, better, Base85. ;-) On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Sure, I will certainly look into adding an option to use base64 > encoded files in the future. Thanks for the suggestion. > > On 22 March 2015 at 15:43, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Maybe add additional option to crypt_file function? >> >> Or even add new function crypt_file_b64? >> >> This is not so important, but will be useful. >> >> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>> I hadn't considered using base64, no :-/ There are doubtless other >>> encodings that are even smaller too, but the trouble with changing it >>> now is that users will have to re-encrypt any previously encrypted >>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>> now unless there is a significant speed-up to be had. I will certainly >>> consider it for the future, though, and definitely if there is ever a >>> need to make some other incompatible change that would require files >>> to be re-encrypted anyway.

Sun Mar 22 12:21:07 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 18:20:56 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Download (untitled)
application/pgp-encrypted 11b

Message body not shown because it is not plain text.

Download encrypted.asc
application/octet-stream 2.7k

Message body not shown because it is not plain text.

Sun Mar 22 12:23:11 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 18:22:56 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

On 22.03.2015 18:08, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Ok. It still seems weird to me that scripts with the #!perl shebang > line are not running on your system, especially given that all the > t/*.t scripts have that in them! My system's perldoc script also has > #!perl in it. Does yours too? > > How did you run that test script (1.pl)? Did you run "perl 1.pl" or

just "1.pl"? I run ./1.pl, without perl. Show quoted text

> Please could you also try this for me?: Replace the following line: > > chomp($data = qx{$perl $crypt_file -m}); > > near the end of t/03_script.t with this: > > chomp($data = qx{/usr/local/perl-5.20.2/bin/perldoc

$crypt_file}); Show quoted text

> > and then try running "make test" again.

t/03_script.t ........ 94/105 Can't find any loadable formatter class in Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text Pod::TEXT Pod::Perldoc::ToPod?! Aborting at /usr/local/perl-5.20.2/bin/perldoc line 6. # Failed test '-m option works' # at t/03_script.t line 673. # '' # doesn't match '(?^msx:^ NAME .*? # ^ SYNOPSIS .*? # ^ ARGUMENTS .*? # ^ OPTIONS .*? # ^ EXIT\ STATUS .*? # ^ DIAGNOSTICS .*? # ^ EXAMPLES .*? # ^ ENVIRONMENT .*? # ^ SEE\ ALSO .*? # ^ AUTHOR .*? # ^ COPYRIGHT .*? # ^ LICENCE .*? # ^ VERSION .*? # ^ DATE .*? # ^ HISTORY )' # Looks like you failed 1 test of 105. t/03_script.t ........ Dubious, test returned 1 (wstat 256, 0x100) Failed 1/105 subtests Show quoted text

> On 22 March 2015 at 15:35, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Hi. >> >> with !#perl shebang: >> >> bash: /var/local/1.pl: perl: bad interpreter: No such file or directory >> >> with #!/usr/bin/env perl shebang: >> >> script_dir=/usr/local/perl-5.20.2/bin >> perldoc=/usr/local/perl-5.20.2/bin/perldoc >> noperldoc=0 >> >> >> On 22.03.2015 15:41, steve.m.hay@googlemail.com via RT wrote:

>>> use strict; >>> use warnings; >>> use Config; >>> use ExtUtils::MakeMaker; >>> use File::Spec::Functions qw(catfile); >>> my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; >>> my $perldoc = catfile($script_dir, 'perldoc'); >>> my $noperldoc = MM->maybe_command($perldoc) ? 0 : 1; >>> print "script_dir=$script_dir\n"; >>> print "perldoc=$perldoc\n"; >>> print "noperldoc=$noperldoc\n";

>>

Sun Mar 22 12:27:50 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 18:27:39 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Download (untitled)
application/pgp-encrypted 11b

Message body not shown because it is not plain text.

Download encrypted.asc
application/octet-stream 2.2k

Message body not shown because it is not plain text.

Sun Mar 22 12:30:16 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 18:30:05 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. I found the problem: perldoc itself has #!perl, instead of #!/usr/bin/env perl. So, even if perldoc from perl-5.20 running - old perl interpreter started. On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Sure, I will certainly look into adding an option to use base64 > encoded files in the future. Thanks for the suggestion. > > On 22 March 2015 at 15:43, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Maybe add additional option to crypt_file function? >> >> Or even add new function crypt_file_b64? >> >> This is not so important, but will be useful. >> >> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>> I hadn't considered using base64, no :-/ There are doubtless other >>> encodings that are even smaller too, but the trouble with changing it >>> now is that users will have to re-encrypt any previously encrypted >>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>> now unless there is a significant speed-up to be had. I will certainly >>> consider it for the future, though, and definitely if there is ever a >>> need to make some other incompatible change that would require files >>> to be re-encrypted anyway.

>>

>

Sun Mar 22 13:19:55 2015 SREZIC [...] cpan.org - Correspondence added

On 2015-03-11 15:07:35, dzagashev@gmail.com wrote: Show quoted text

> I have two interpreters on this server. > > Standard centos perl 5.14. and working 5.20. > > Working interpreter is in path before /usr/bin/. > > So, shebang #!perl should work correctly, my mistake. > > No, it not works at all. I run small script: > > #!perl > print $^X; > > and get error - bash: ./1.pl: perl: bad interpreter: No such file or > directory

You just put #!perl into a script which is expected to be postprocessed by ExtUtils::MakeMaker. In the postprocessing the "perl" string would be expanded to the full path of the current perl (and additionally there will be some stuff added for strange operating systems, and special handling for Windows will occur). Show quoted text

> > On 11.03.2015 20:52, Slaven_Rezic via RT wrote:

> > <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > > > On 2015-03-11 14:38:19, dzagashev@gmail.com wrote:

> >> Seems I understand what happens. > >> > >> You have shebang #!perl, but more wide used practice (and more > >> correct) > >> is to use #!/usr/bin/env perl.

> > Sorry, I have to disagree. Currently it's incorrect to specify > > #!/usr/bin/env perl for scripts which are installed normally via > > blib/scripts. The reason is that ExtUtils::MakeMaker's > > _fixin_replace_shebang does not work correctly otherwise --- it does > > not replace the shebang. Correct usage is #!perl or #!/usr/bin/perl > > or so. > >

> >> In my case invalid perl interpreter used. > >> > >> On 11.03.2015 20:12, steve.m.hay@googlemail.com via RT wrote:

> >>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > >>> > >>> On 11 March 2015 at 14:36, dzagashev@gmail.com via RT > >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

> >>>> Queue: Filter-Crypto > >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > >>>> > >>>> This error only appears during make test. > >>>> > >>>> And prevent to install package. > >>>>

> >>> Please could you run the attached test.t from the root Filter- > >>> Crypto > >>> folder and send the output? > >>> > >>> I'm just trying to figure out why "$^X script/crypt_file -m" fails > >>> (in > >>> make test), but a simple "perldoc script/crypt_file" works... > >>>

> >>

> > > >

Sun Mar 22 13:31:25 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 17:31:15 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

On 22 March 2015 at 16:23, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > > On 22.03.2015 18:08, steve.m.hay@googlemail.com via RT wrote:

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Ok. It still seems weird to me that scripts with the #!perl shebang >> line are not running on your system, especially given that all the >> t/*.t scripts have that in them! My system's perldoc script also has >> #!perl in it. Does yours too? >> >> How did you run that test script (1.pl)? Did you run "perl 1.pl" or

> just "1.pl"? > > > I run ./1.pl, without perl.

I guessed as much. As Slaven has just noted, that's not expected to work in a script that just contains #!perl, rather than what EU::MM would have changed the shebang line to. Presumably if you run "perl 1.pl" instead then that would run?... Show quoted text

> >

>> Please could you also try this for me?: Replace the following line: >> >> chomp($data = qx{$perl $crypt_file -m}); >> >> near the end of t/03_script.t with this: >> >> chomp($data = qx{/usr/local/perl-5.20.2/bin/perldoc

> $crypt_file});

... and likewise if you put "qx{$^X /usr/local/perl-5.20.2/bin/perldoc $crypt_file}" instead then the test would now pass? Show quoted text

>> >> and then try running "make test" again.

> > > > t/03_script.t ........ 94/105 Can't find any loadable formatter class in > Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText > Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text > Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text > Pod::TEXT Pod::Perldoc::Toterm Pod::Perldoc::Toterm Pod::Perldoc::ToTerm > Pod::Perldoc::ToTERM Pod::Simple::term Pod::Simple::term > Pod::Simple::Term Pod::Simple::TERM Pod::term Pod::term Pod::Term > Pod::TERM Pod::Perldoc::Totext Pod::Perldoc::Totext Pod::Perldoc::ToText > Pod::Perldoc::ToTEXT Pod::Simple::text Pod::Simple::text > Pod::Simple::Text Pod::Simple::TEXT Pod::text Pod::text Pod::Text > Pod::TEXT Pod::Perldoc::ToPod?! > Aborting > at /usr/local/perl-5.20.2/bin/perldoc line 6. > > # Failed test '-m option works' > # at t/03_script.t line 673. > # '' > # doesn't match '(?^msx:^ NAME .*? > # ^ SYNOPSIS .*? > # ^ ARGUMENTS .*? > # ^ OPTIONS .*? > # ^ EXIT\ STATUS .*? > # ^ DIAGNOSTICS .*? > # ^ EXAMPLES .*? > # ^ ENVIRONMENT .*? > # ^ SEE\ ALSO .*? > # ^ AUTHOR .*? > # ^ COPYRIGHT .*? > # ^ LICENCE .*? > # ^ VERSION .*? > # ^ DATE .*? > # ^ HISTORY )' > # Looks like you failed 1 test of 105. > t/03_script.t ........ Dubious, test returned 1 (wstat 256, 0x100) > Failed 1/105 subtests > > > >

>> On 22 March 2015 at 15:35, dzagashev@gmail.com via RT >> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>> Queue: Filter-Crypto >>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> Hi. >>> >>> with !#perl shebang: >>> >>> bash: /var/local/1.pl: perl: bad interpreter: No such file or directory >>> >>> with #!/usr/bin/env perl shebang: >>> >>> script_dir=/usr/local/perl-5.20.2/bin >>> perldoc=/usr/local/perl-5.20.2/bin/perldoc >>> noperldoc=0 >>> >>> >>> On 22.03.2015 15:41, steve.m.hay@googlemail.com via RT wrote:

>>>> use strict; >>>> use warnings; >>>> use Config; >>>> use ExtUtils::MakeMaker; >>>> use File::Spec::Functions qw(catfile); >>>> my $script_dir = $Config{scriptdirexp} || $Config{scriptdir}; >>>> my $perldoc = catfile($script_dir, 'perldoc'); >>>> my $noperldoc = MM->maybe_command($perldoc) ? 0 : 1; >>>> print "script_dir=$script_dir\n"; >>>> print "perldoc=$perldoc\n"; >>>> print "noperldoc=$noperldoc\n";

>>>

> >

Sun Mar 22 13:36:13 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 17:36:04 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

I wonder how that has happened? My perldoc has only #!perl in it, but I'm running on Windows, which doesn't use the shebang mechanism at all. Instead, my scripts are turned into "batch files", with various bits of black magic added to them to make them run nicely. In your case, EU::MM should have changed that #!perl into something more useful :-/ Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is shipped with the core perl distribution, but maybe the way in which it gets installed by that is not quite right and bypasses the usual shebang line processing?) On 22 March 2015 at 16:30, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. > > I found the problem: > > perldoc itself has #!perl, instead of #!/usr/bin/env perl. > > So, even if perldoc from perl-5.20 running - old perl interpreter started. > > On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Sure, I will certainly look into adding an option to use base64 >> encoded files in the future. Thanks for the suggestion. >> >> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>> Queue: Filter-Crypto >>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> Maybe add additional option to crypt_file function? >>> >>> Or even add new function crypt_file_b64? >>> >>> This is not so important, but will be useful. >>> >>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>> encodings that are even smaller too, but the trouble with changing it >>>> now is that users will have to re-encrypt any previously encrypted >>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>> now unless there is a significant speed-up to be had. I will certainly >>>> consider it for the future, though, and definitely if there is ever a >>>> need to make some other incompatible change that would require files >>>> to be re-encrypted anyway.

>>>

>>

>

Sun Mar 22 15:11:30 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 21:11:13 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

I found that system('/usr/local/perl-5.20.2/bin/perldoc', '-F', '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); generate error, when executed from make test, and works normally when I run it directly from bash. Maybe MM redefine system? On 22.03.2015 19:36, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I wonder how that has happened? My perldoc has only #!perl in it, but > I'm running on Windows, which doesn't use the shebang mechanism at > all. Instead, my scripts are turned into "batch files", with various > bits of black magic added to them to make them run nicely. In your > case, EU::MM should have changed that #!perl into something more > useful :-/ > > Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is > shipped with the core perl distribution, but maybe the way in which it > gets installed by that is not quite right and bypasses the usual > shebang line processing?) > > > On 22 March 2015 at 16:30, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. >> >> I found the problem: >> >> perldoc itself has #!perl, instead of #!/usr/bin/env perl. >> >> So, even if perldoc from perl-5.20 running - old perl interpreter started. >> >> On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> Sure, I will certainly look into adding an option to use base64 >>> encoded files in the future. Thanks for the suggestion. >>> >>> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> Maybe add additional option to crypt_file function? >>>> >>>> Or even add new function crypt_file_b64? >>>> >>>> This is not so important, but will be useful. >>>> >>>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>>> encodings that are even smaller too, but the trouble with changing it >>>>> now is that users will have to re-encrypt any previously encrypted >>>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>>> now unless there is a significant speed-up to be had. I will certainly >>>>> consider it for the future, though, and definitely if there is ever a >>>>> need to make some other incompatible change that would require files >>>>> to be re-encrypted anyway.

Sun Mar 22 15:21:28 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 21:21:17 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Download (untitled)
application/pgp-encrypted 11b

Message body not shown because it is not plain text.

Download encrypted.asc
application/octet-stream 2.5k

Message body not shown because it is not plain text.

Sun Mar 22 15:38:13 2015 steve.m.hay [...] googlemail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 19:38:04 +0000
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	Steve Hay <steve.m.hay [...] googlemail.com>

I don't think it redefines system(), but the PATH environment variable may well be different when running test scripts via "make test" compared to your standard bash shell. It might be worth printing out $ENV{PATH} before the system() call to see what's going on? I still think reinstalling Pod::Perldoc is worth a try too - it doesn't sound correct to me that perldoc contains an unmodified #!perl. On 22 March 2015 at 19:11, dzagashev@gmail.com via RT <bug-Filter-Crypto@rt.cpan.org> wrote: Show quoted text

> Queue: Filter-Crypto > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I found that > > system('/usr/local/perl-5.20.2/bin/perldoc', '-F', > '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); > > generate error, when executed from make test, and works normally when I > run it directly from bash. > > Maybe MM redefine system? > > > On 22.03.2015 19:36, steve.m.hay@googlemail.com via RT wrote:

>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> I wonder how that has happened? My perldoc has only #!perl in it, but >> I'm running on Windows, which doesn't use the shebang mechanism at >> all. Instead, my scripts are turned into "batch files", with various >> bits of black magic added to them to make them run nicely. In your >> case, EU::MM should have changed that #!perl into something more >> useful :-/ >> >> Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is >> shipped with the core perl distribution, but maybe the way in which it >> gets installed by that is not quite right and bypasses the usual >> shebang line processing?) >> >> >> On 22 March 2015 at 16:30, dzagashev@gmail.com via RT >> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>> Queue: Filter-Crypto >>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. >>> >>> I found the problem: >>> >>> perldoc itself has #!perl, instead of #!/usr/bin/env perl. >>> >>> So, even if perldoc from perl-5.20 running - old perl interpreter started. >>> >>> On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> Sure, I will certainly look into adding an option to use base64 >>>> encoded files in the future. Thanks for the suggestion. >>>> >>>> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>> Queue: Filter-Crypto >>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> Maybe add additional option to crypt_file function? >>>>> >>>>> Or even add new function crypt_file_b64? >>>>> >>>>> This is not so important, but will be useful. >>>>> >>>>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>>>> encodings that are even smaller too, but the trouble with changing it >>>>>> now is that users will have to re-encrypt any previously encrypted >>>>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>>>> now unless there is a significant speed-up to be had. I will certainly >>>>>> consider it for the future, though, and definitely if there is ever a >>>>>> need to make some other incompatible change that would require files >>>>>> to be re-encrypted anyway.

> >

Sun Mar 22 15:48:49 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 21:48:36 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

I reinstall perl completely - this not help. On 22.03.2015 21:38, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I don't think it redefines system(), but the PATH environment variable > may well be different when running test scripts via "make test" > compared to your standard bash shell. It might be worth printing out > $ENV{PATH} before the system() call to see what's going on? > > I still think reinstalling Pod::Perldoc is worth a try too - it > doesn't sound correct to me that perldoc contains an unmodified > #!perl. > > > On 22 March 2015 at 19:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> I found that >> >> system('/usr/local/perl-5.20.2/bin/perldoc', '-F', >> '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); >> >> generate error, when executed from make test, and works normally when I >> run it directly from bash. >> >> Maybe MM redefine system? >> >> >> On 22.03.2015 19:36, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> I wonder how that has happened? My perldoc has only #!perl in it, but >>> I'm running on Windows, which doesn't use the shebang mechanism at >>> all. Instead, my scripts are turned into "batch files", with various >>> bits of black magic added to them to make them run nicely. In your >>> case, EU::MM should have changed that #!perl into something more >>> useful :-/ >>> >>> Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is >>> shipped with the core perl distribution, but maybe the way in which it >>> gets installed by that is not quite right and bypasses the usual >>> shebang line processing?) >>> >>> >>> On 22 March 2015 at 16:30, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. >>>> >>>> I found the problem: >>>> >>>> perldoc itself has #!perl, instead of #!/usr/bin/env perl. >>>> >>>> So, even if perldoc from perl-5.20 running - old perl interpreter started. >>>> >>>> On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> Sure, I will certainly look into adding an option to use base64 >>>>> encoded files in the future. Thanks for the suggestion. >>>>> >>>>> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>> Queue: Filter-Crypto >>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> Maybe add additional option to crypt_file function? >>>>>> >>>>>> Or even add new function crypt_file_b64? >>>>>> >>>>>> This is not so important, but will be useful. >>>>>> >>>>>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>>>>> encodings that are even smaller too, but the trouble with changing it >>>>>>> now is that users will have to re-encrypt any previously encrypted >>>>>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>>>>> now unless there is a significant speed-up to be had. I will certainly >>>>>>> consider it for the future, though, and definitely if there is ever a >>>>>>> need to make some other incompatible change that would require files >>>>>>> to be re-encrypted anyway.

>>

Sun Mar 22 15:54:51 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 21:54:39 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

Also, I debug all calls in Pod::Usage. All variables defined correctly: perl path, perldoc path, crypt_file path. But system() call with absolutely correct absolute path to perldoc dies with error. system('/usr/local/perl-5.20.2/bin/perldoc', '-F', '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); Why this trivial call could not work? And it not work only from make test. On 22.03.2015 21:38, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I don't think it redefines system(), but the PATH environment variable > may well be different when running test scripts via "make test" > compared to your standard bash shell. It might be worth printing out > $ENV{PATH} before the system() call to see what's going on? > > I still think reinstalling Pod::Perldoc is worth a try too - it > doesn't sound correct to me that perldoc contains an unmodified > #!perl. > > > On 22 March 2015 at 19:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> I found that >> >> system('/usr/local/perl-5.20.2/bin/perldoc', '-F', >> '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); >> >> generate error, when executed from make test, and works normally when I >> run it directly from bash. >> >> Maybe MM redefine system? >> >> >> On 22.03.2015 19:36, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> I wonder how that has happened? My perldoc has only #!perl in it, but >>> I'm running on Windows, which doesn't use the shebang mechanism at >>> all. Instead, my scripts are turned into "batch files", with various >>> bits of black magic added to them to make them run nicely. In your >>> case, EU::MM should have changed that #!perl into something more >>> useful :-/ >>> >>> Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is >>> shipped with the core perl distribution, but maybe the way in which it >>> gets installed by that is not quite right and bypasses the usual >>> shebang line processing?) >>> >>> >>> On 22 March 2015 at 16:30, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. >>>> >>>> I found the problem: >>>> >>>> perldoc itself has #!perl, instead of #!/usr/bin/env perl. >>>> >>>> So, even if perldoc from perl-5.20 running - old perl interpreter started. >>>> >>>> On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> Sure, I will certainly look into adding an option to use base64 >>>>> encoded files in the future. Thanks for the suggestion. >>>>> >>>>> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>> Queue: Filter-Crypto >>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> Maybe add additional option to crypt_file function? >>>>>> >>>>>> Or even add new function crypt_file_b64? >>>>>> >>>>>> This is not so important, but will be useful. >>>>>> >>>>>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>>>>> encodings that are even smaller too, but the trouble with changing it >>>>>>> now is that users will have to re-encrypt any previously encrypted >>>>>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>>>>> now unless there is a significant speed-up to be had. I will certainly >>>>>>> consider it for the future, though, and definitely if there is ever a >>>>>>> need to make some other incompatible change that would require files >>>>>>> to be re-encrypted anyway.

>>

Sun Mar 22 16:03:18 2015 dzagashev [...] gmail.com - Correspondence added

Subject:	Re: [rt.cpan.org #102525] Tests not passing
Date:	Sun, 22 Mar 2015 22:03:05 +0200
To:	bug-Filter-Crypto [...] rt.cpan.org
From:	"dzagashev [...] gmail.com" <dzagashev [...] gmail.com>

/usr/local/perl-5.20.2/bin/perldoc has hardcoded shebang '#!/usr/local/perl-5.20.2/bin/perl', so this is impossible, that some other perl can be executed. On 22.03.2015 21:38, steve.m.hay@googlemail.com via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > > > I don't think it redefines system(), but the PATH environment variable > may well be different when running test scripts via "make test" > compared to your standard bash shell. It might be worth printing out > $ENV{PATH} before the system() call to see what's going on? > > I still think reinstalling Pod::Perldoc is worth a try too - it > doesn't sound correct to me that perldoc contains an unmodified > #!perl. > > > On 22 March 2015 at 19:11, dzagashev@gmail.com via RT > <bug-Filter-Crypto@rt.cpan.org> wrote:

>> Queue: Filter-Crypto >> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >> >> I found that >> >> system('/usr/local/perl-5.20.2/bin/perldoc', '-F', >> '/root/.cpanm/work/1427047246.28711/Filter-Crypto-2.07/blib/script/crypt_file'); >> >> generate error, when executed from make test, and works normally when I >> run it directly from bash. >> >> Maybe MM redefine system? >> >> >> On 22.03.2015 19:36, steve.m.hay@googlemail.com via RT wrote:

>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>> >>> I wonder how that has happened? My perldoc has only #!perl in it, but >>> I'm running on Windows, which doesn't use the shebang mechanism at >>> all. Instead, my scripts are turned into "batch files", with various >>> bits of black magic added to them to make them run nicely. In your >>> case, EU::MM should have changed that #!perl into something more >>> useful :-/ >>> >>> Does it help if you re-install Pod::Perldoc from CPAN? (perldoc is >>> shipped with the core perl distribution, but maybe the way in which it >>> gets installed by that is not quite right and bypasses the usual >>> shebang line processing?) >>> >>> >>> On 22 March 2015 at 16:30, dzagashev@gmail.com via RT >>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>> Queue: Filter-Crypto >>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>> >>>> Sorry, Thunderbird start encrypt all messages with PGP unexpectedly. >>>> >>>> I found the problem: >>>> >>>> perldoc itself has #!perl, instead of #!/usr/bin/env perl. >>>> >>>> So, even if perldoc from perl-5.20 running - old perl interpreter started. >>>> >>>> On 22.03.2015 18:09, steve.m.hay@googlemail.com via RT wrote:

>>>>> <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>> >>>>> Sure, I will certainly look into adding an option to use base64 >>>>> encoded files in the future. Thanks for the suggestion. >>>>> >>>>> On 22 March 2015 at 15:43, dzagashev@gmail.com via RT >>>>> <bug-Filter-Crypto@rt.cpan.org> wrote:

>>>>>> Queue: Filter-Crypto >>>>>> Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=102525 > >>>>>> >>>>>> Maybe add additional option to crypt_file function? >>>>>> >>>>>> Or even add new function crypt_file_b64? >>>>>> >>>>>> This is not so important, but will be useful. >>>>>> >>>>>> On 22.03.2015 15:46, steve.m.hay@googlemail.com via RT wrote:

>>>>>>> I hadn't considered using base64, no :-/ There are doubtless other >>>>>>> encodings that are even smaller too, but the trouble with changing it >>>>>>> now is that users will have to re-encrypt any previously encrypted >>>>>>> files if they upgrade their Filter-Crypto. So I'm not keen on doing it >>>>>>> now unless there is a significant speed-up to be had. I will certainly >>>>>>> consider it for the future, though, and definitely if there is ever a >>>>>>> need to make some other incompatible change that would require files >>>>>>> to be re-encrypted anyway.

>>

Wed May 27 17:44:13 2015 ZDM [...] cpan.org - Correspondence added

Hi. Seems, that this issue doesn't appears anymore. I don't know the reason, maybe some dependencies was updated during this two months. Ticked can be closed. Thanks a lot.

Thu Aug 25 08:35:37 2016 shay [...] cpan.org - Correspondence added

On Wed May 27 17:44:13 2015, ZDM wrote: Show quoted text

> Hi. > > Seems, that this issue doesn't appears anymore. > > I don't know the reason, maybe some dependencies was updated during > this two months. > > Ticked can be closed. > > Thanks a lot.

Thanks, closing ticket as requested.

Thu Aug 25 08:35:42 2016 shay [...] cpan.org - Status changed from 'open' to 'resolved'