Like <DATA>, Only Less Stupid
First, An Alternative
Data::Section was actually written when I realized I couldn't use Inline::Files. I'd wanted to use that ages ago, but it had a number of little problems. The biggest problem, unfortunately, was that it just tried to be way, way too powerful.
I like overpowered modules as much as the next guy, but here's about where I draw the line: Inline::Files is a form of source filter, and includes this warning:
It is possible that this module may overwrite the source code in files that use it. To protect yourself against this possibility, you are strongly advised to use the -backup option described in "Safety first".
The DATA Section
Data::Section (and Inline::Files) are meant to make the
__DATA__ section both more powerful and more reliable. For those of you who haven't encountered it, Perl lets you access a virtual filehandle that reads non-code content at the end of your program file. In other words, if you write this program:
...then your program would output:
Got number: 2 Got number: 4 Got number: 6 Got number: 8 Total: 20
Why Data::Section Beats DATA
No Need to seek
There are a few problems with
DATA, though. The most annoying is that you can't easily seek on it. Its read position is actually relative to the whole source file, so you'd need to try to
tell it first, then save that, and things get messy.
Data::Section caches your data for you, so you can reread it any time you want.
Multiple "Files" Per Module
DATA filehandle is one big entity, but Data::Section lets you provide multiple virtual files in your data:
You'll get two "filenames" and each one will have five lines in it.
(By the way, that
use Data::Section line? Yeah, that's all you have to do. Obviously, Data::Section uses Sub::Exporter, so you can rename its methods, and there are a few parameters you can use to customize them.)
If you're wondering what's "merged" in the
merged_section_data method, it's inherited files. In other words, if you wrote a subclass of the Your::Package code above, you could add new files in its
DATA section, but the old files would still be accessible -- unless you added a so-far.txt definition to your subclass. It behaves very much like an overridden method, in that way.
When is this useful? It's great to use instead of here-docs for a lot of cases. It's also a nice way to avoid having to install and manage text files. A package can store its own bundle of configuration or template files, and it will have methods to access them. Then if you want, you can specialize your pack of template files by subclassing the original one.
Even without all its new features, the simple ability to never re-seek
DATA is a big one. (Please note that Data::Section caches the contents, it doesn't magically take care of re-seeking. That ends up being a real problem.)
DATA is a global variable, and if code ever tries to read from it more than once, the second time will fail, and that is a really annoying bug to track down -- trust me, I've had to do so many times. So, forget about plain old
DATA sections and start using Data::Section. Your future self will thank you.