[tpm] mod_perl ... still suitable choice?

[Olaf Alders]

On Jun 17, 2014, at 10:58 AM, Dominic Thoreau <dominic at thoreau-online.net> wrote:

> 
> We're [1] running Apache/mod_perl. We have, by quick grep, 1.1M lines of perl code. Continuing to run mod_perl is pretty much the only choice. Selling management on "we need to re-write this, it'll take months, and the net effect, hopefully, will be no new functionality, and probably new bugs" is a non-starter.

I’m in a similar situation at $work.  Lots of critical code still running under mod_perl.  The biggest problem is that it’s sometimes hard to test the mod_perl code.  We do have some WWW::Mechanize tests that fire up their own Apache on a different port in order to run the tests, but it’s still kind of slow to do that for individual unit tests.  It’s not ideal for that kind of stuff, but it’s a huge pain to switch over and it’s not broken so there’s no hurry to fix it.

I did have some of my own code that was built in CGI::Application that I was running under ModPerl::Registry.  That actually was mostly trivial to switch over to Plack because it didn’t rely directly on any Apache::* modules.

Olaf