Bug #101615 for DBIx-Class: t/54taint.t fails on Windows 7 "perl in space"

Mon Jan 19 06:21:18 2015 ETJ [...] cpan.org - Ticket created

Subject:

t/54taint.t fails on Windows 7 "perl in space"

When perl is installed as D:\straw(berry) perl\perl\bin\perl.exe, t/54taint.t says: Can't open perl script "perl\perl\bin\perl.exe": No such file or directory I had a look for why (it's the "exec" line), and I tried making a list of @execargs and using the form "exec { $execargs[0] } @execargs", per perldoc -f exec. Couldn't figure a way to force not to use shell.

Mon Jan 19 10:05:45 2015 ribasushi [...] leporine.io - Correspondence added

On Mon Jan 19 12:21:18 2015, ETJ wrote: Show quoted text

> When perl is installed as D:\straw(berry) perl\perl\bin\perl.exe, > t/54taint.t says: > > Can't open perl script "perl\perl\bin\perl.exe": No such file or > directory > > I had a look for why (it's the "exec" line), and I tried making a list > of @execargs and using the form "exec { $execargs[0] } @execargs", per > perldoc -f exec. Couldn't figure a way to force not to use shell.

Um... I am confused... Are you saying that contrary to `perldoc -f exec` you simply can't avoid the shell? Will "check for space in $^X" escape at the start of the BEGIN suffice to avoid the entire brouhaha?

Mon Jan 19 10:05:45 2015 The RT System itself - Status changed from 'new' to 'open'

Mon Jan 19 10:30:32 2015 ETJ [...] cpan.org - Correspondence added

On Mon Jan 19 10:05:45 2015, RIBASUSHI wrote: Show quoted text

> Um... I am confused... Are you saying that contrary to `perldoc -f > exec` you simply can't avoid the shell? > > Will "check for space in $^X" escape at the start of the BEGIN suffice > to avoid the entire brouhaha?

I'm saying I couldn't figure how to. Windows 7, Strawberry Perl 5.20.0. An effective workaround would be to switch that test to using string-arg system and exec, and quote $^X et al. Someone cleverer than me might be able to figure out why this is happening, and that might be worthwhile.

Mon Jan 19 10:36:58 2015 ribasushi [...] leporine.io - Correspondence added

Subject:	Re: [rt.cpan.org #101615] t/54taint.t fails on Windows 7 "perl in space"
Date:	Mon, 19 Jan 2015 16:36:39 +0100
To:	bug-DBIx-Class [...] rt.cpan.org
From:	Peter Rabbitson <ribasushi [...] cpan.org>

On 01/19/2015 04:30 PM, Ed J via RT wrote: Show quoted text

> Queue: DBIx-Class > Ticket <URL: https://rt.cpan.org/Ticket/Display.html?id=101615 > > > On Mon Jan 19 10:05:45 2015, RIBASUSHI wrote:

>> Um... I am confused... Are you saying that contrary to `perldoc -f >> exec` you simply can't avoid the shell? >> >> Will "check for space in $^X" escape at the start of the BEGIN suffice >> to avoid the entire brouhaha?

> I'm saying I couldn't figure how to. Windows 7, Strawberry Perl 5.20.0.

Am I understanding correctly that the system() a little higher up works, yet the exec() doesn't? Show quoted text

> An effective workaround would be to switch that test to using string-arg system and exec, and quote $^X et al.

As far as I understand there is no universal quoting (working both on linux AND windows). If this is indeed the case the workaround is hardly "effective" :) The reason I proposed preempting the entire BEGIN is because this is a "canary test", which is not relevant to the functionality on an end-user's machine: https://github.com/dbsrgits/dbix-class/commit/ade96c1f2c6. Therefore, in the interest of perl-in-space, I am perfectly fine with skipping the test on $^X =~ /\s/, and be done with it. Especially if this is the *only* thing that failed (which is extremely surprising to me).

Mon Jan 19 16:54:50 2015 ETJ [...] cpan.org - Correspondence added

On Mon Jan 19 10:36:58 2015, RIBASUSHI wrote: Show quoted text

> As far as I understand there is no universal quoting (working both on > linux AND windows). If this is indeed the case the workaround is > hardly

I have some modest experience with making command lines working "in space", and applying this concept seems to work great everywhere (at least on Unix, Windows, VMS): "command" "arg1" "arg2" I claim support for this, from the general success of the "perl in space" changes to EUMM. Show quoted text

> Therefore, in the interest of perl-in-space, I am perfectly fine with > skipping the test on $^X =~ /\s/, and be done with it. Especially if > this is the *only* thing that failed (which is extremely surprising to > me).

DBIC uses good practice with list-args for system/exec (which is what it's trying for here as well), so it's not very surprising to me.

Sun Feb 15 05:49:58 2015 ribasushi [...] leporine.io - Correspondence added

On Mon Jan 19 16:36:58 2015, RIBASUSHI wrote: Show quoted text

> > Am I understanding correctly that the system() a little higher up > works, > yet the exec() doesn't?

You never answered this part of my question. Is it the case that exec LIST does not behave the same as system LIST when it comes to a space-containing $^X? If this is the case is there a bug about it in the perl RT system? Thanks

Mon Feb 23 05:03:13 2015 MITHALDU [...] cpan.org - Correspondence added

This appears to be a core bug: https://rt.perl.org/Ticket/Display.html?id=123907

Thu Mar 05 14:28:49 2015 ribasushi [...] leporine.io - Correspondence added

Given it's a core bug, the problematic test was shunted in[1]. Please test the newly uploaded DBIC 0.082899_03[2] and let me know if the problem is resolved. Cheers [1] https://github.com/dbsrgits/dbix-class/commit/a5741f5c#diff-5084a4f11dec04497be5c5a23996f4d1 [2] https://metacpan.org/release/RIBASUSHI/DBIx-Class-0.082899_03

Thu Mar 05 14:28:51 2015 ribasushi [...] leporine.io - Status changed from 'open' to 'patched'

Mon Mar 09 06:34:17 2015 ETJ [...] cpan.org - Correspondence added 10 min

On Thu Mar 05 14:28:49 2015, RIBASUSHI wrote: Show quoted text

> [...] Please > test the newly uploaded DBIC 0.082899_03[2] and let me know if the > problem is resolved.

That version works on my system now (all tests pass and it installs), nice work.

Mon Mar 09 06:34:20 2015 ETJ [...] cpan.org - Fixed in 0.082899_03 added

Mon Mar 23 09:34:09 2015 ribasushi [...] leporine.io - Fixed in 0.082820 added

Mon Mar 23 09:34:11 2015 ribasushi [...] leporine.io - Fixed in 0.082899_03 deleted

Mon Mar 23 09:34:12 2015 ribasushi [...] leporine.io - Status changed from 'patched' to 'resolved'