|Perl: the Markov chain saw|
Stop suggesting to upgrade perlby vsespb (Hermit)
|on Aug 31, 2013 at 17:25 UTC||Need Help??|
Often on PerlMonks I see posts, asking for some help, with a requirement for perl minimum version (i.e. "perl 5.8.8+" or "perl 5.10+", or "any perl above 5.8" etc).
What surprise me is that people still trying to advice to "upgrade perl" again and again. Why??
Advice to upgrade perl comes from wrong assumption that:
There is only one instance where this code supposed to work.
That's often not true.
Cases where there is only one instance where code should work:
1) You are the only user of your own code. You hack it for yourself.
2) It's a webdevelopment and there is just single instance of web application (I am not talking about single server vs cluster, I am talking about case when you develop one site somesite.com, but not a web framework, not a software-as-service to be installed on your clients machines)
Cases when there are many instances of your code (i.e. when your code have users):
1) You are developing CPAN distribution
2) You are developing standalone application
And it's not an option to ask your users to upgrade perl.
Imaging that all the software that you use (say, under Linux) suddenly change install instruction and and ask you to install "rvm", "virtualenv", "perlbrew", upgrade Glib and GCC, and install certain versions of Python, Ruby and Perl into virtual environments.
If no one would be allowed to run system, vendor, perl, perl would be already dead.
And yes, about "unsupported" perl versions. Most verndor perl versions are unsupported (well, they supposed to be supported by vendors)
Conclusion: Please, don't suggest to upgrade perl if you see that there are clear requirements about *range* of perl versions in the post.