Chapter 13. Writing a Distribution
In Chapter 12,
you created a fictional
module, and built the right support for
so that you could include
use Island::Plotting::Maps in a program.
While the resulting
.pm file is useful,
it’s not very practical. There’s a
lot more to building a real module than just creating the
.pm file. You’ll also need to
consider and implement the following questions:
- Installation location
How and where is the
.pmfile installed so a program can find the module in its
Where is the documentation for the module? How is the documentation installed so the user can read it?
- Archive completeness
If there are any accompanying files, where are they? How can the end user know if any missing files are missing?
What test harnesses can the developer run to verify correct operation, including testing against older known bugs? Can these same tests be run by the installer to ensure proper operation in the target environment?
- C-language interfaces
If the module contains C or C++ code (not covered here), how can the developer describe how to compile and link the code in the developer’s environment, or the end user environment?
As Roy Scheider uttered in the movie Jaws: “You’re gonna need a bigger boat.” That “bigger boat” is the difference between a module and a distribution.
Starting with h2xs
A distribution contains the module (or collection of related modules), plus all the support files required to document, test, ...