Some folks have commented on the recent removal of
Mojo::Server::FastCGI from the Mojolicious core; one google groups
entry was even removed. (not by the Mojolicious people!)
Among the complaints:
- ACK! I need that!
- OH NOES! I use that!
- Other variations of the above
I understand. All of these responses are valid and need to be
addressed.
First, Mojolicious does move quickly. It's a of the hallmark of the
project. It will continue to move quickly, as it's woven into the
code, into the very bits of each character. I hear it's a new unicode
standard.
Second, Mojolicious never leaves people out to dry - there's always a
path for you to do what you need to do.
In the case of FastCGI, you have several options:
1) You can use Plack:
http://mojolicio.us/perldoc/Mojolicious/Guides/Cookbook#PSGI2FPlack
2) You can use the most recent Mojolicious distribution that supports
it directly:
Tar:
https://github.com/kraih/mojo/tarball/v1.97
Zip:
https://github.com/kraih/mojo/zipball/v1.97
3) Wait for
https://github.com/arpadszasz to fulfill his commitment to
the project, and nudge him accordingly. *nudge*
4) Port the 1.97 Mojo::Server::FastCGI module to your own custom fork
for Mojolicious
Which one you choose is dependent on your needs, and there's a good
solution for everyone. For a production system, we presume everyone
is putting new releases of any software package, including the
Mojolicious framework, through a rigorous testing process to make sure
each release of your final product works as expected. You do have a
rigorous test suite...right? Of course you do.
This is the first time a a feature has been split off from the
Mojolicious framework, and it's clear that rules need to be set in
place for how it's done, so there's no misunderstanding and everyone
knows what's going on. Once the rules are in place, they will be
posted along side the deprecation policy.
In the end, Mojolicious will keep getting new features. At times,
cumbersome pieces will be removed. It will always be awesome.