Skip Menu |
 

This queue is for tickets about the Module-Starter CPAN distribution.

Report information
The Basics
Id: 93023
Status: open
Priority: 0/
Queue: Module-Starter

People
Owner: Nobody in particular
Requestors: ribasushi [...] cpan.org
Cc:
AdminCc:

Bug Information
Severity: Important
Broken in: (no value)
Fixed in: (no value)



Subject: Non-optional fatalization of strictures
Module injects fatal warnings all over the place, without the ability to turn it off. At least one of the following needs to take place (in order of least work => best outcome) *) M::S needs to unamibuously document what it is emitting *) Needs to provide an option to disable the fatals *) Needs to provide the option and default it to "no fatals *) Needs to stop copying the questionable ideas of strictures altogether More discussion: http://blogs.perl.org/users/peter_rabbitson/2014/01/fatal-warnings-are-a-ticking-time-bomb-via-chromatic.html https://rt.cpan.org/Ticket/Display.html?id=93004 Affected pieces of code: https://metacpan.org/source/XSAWYERX/Module-Starter-1.62/lib/Module/Starter/Simple.pm#L759 https://metacpan.org/source/XSAWYERX/Module-Starter-1.62/lib/Module/Starter/Simple.pm#L808 https://metacpan.org/source/XSAWYERX/Module-Starter-1.62/lib/Module/Starter/Simple.pm#L896 https://metacpan.org/source/XSAWYERX/Module-Starter-1.62/lib/Module/Starter/Simple.pm#L1103 https://metacpan.org/source/XSAWYERX/Module-Starter-1.62/lib/Module/Starter/Simple.pm#L1620
Download (untitled) / with headers
text/plain 994b
On Fri Feb 14 03:40:11 2014, RIBASUSHI wrote: Show quoted text
> Module injects fatal warnings all over the place, without the ability > to turn it off. At least one of the following needs to take place (in > order of least work => best outcome)
I've come to the conclusion that at least for library code, some warnings can't be fatalised for the forwards compatibility reasons chromatic outlines (which is the one actual problem in amongst the doom mongering :). I'd suggest backing back out to just 'use warnings;' and revisiting this once we've got the research done for strictures 2.0 - that should give a clearer idea what's best on/off. You're probably going to end up needing an option anyway, though, while a decent percentage of people seem to find fatal warnings addictive (myself included), it also completely throws off the development patterns of another fair chunk. However I think keeping the default as strict as can be done without forwards compatibility concerns is still a reasonable idea.
Download (untitled) / with headers
text/plain 698b
On Fri Feb 14 13:37:19 2014, MSTROUT wrote: Show quoted text
> On Fri Feb 14 03:40:11 2014, RIBASUSHI wrote:
> > Module injects fatal warnings all over the place, without the ability > > to turn it off. At least one of the following needs to take place (in > > order of least work => best outcome)
> > I've come to the conclusion that at least for library code, some > warnings can't be fatalised for the forwards compatibility reasons > chromatic outlines (which is the one actual problem in amongst the > doom mongering :).
Thank you. I came here just to file this bug report. Maybe fatals warnings can be turned on specifically with a "--i-like-my-users-code-to-break-in-unexpected-ways" switch. Cheers, Ovid


This service is sponsored and maintained by Best Practical Solutions and runs on Perl.org infrastructure.

Please report any issues with rt.cpan.org to rt-cpan-admin@bestpractical.com.