The Rakudo Codebase: Visualized! (Partially)

Yesterday I stumbled upon this old perl6-compiler mailing list message which inspired me to actually try to split up the compilation of CORE.setting into smaller, saner pieces. I’ve started work on this already (so far having just modified the Makefile).

The one and only response to that post suggested using the stub syntax to resolve missing class issues. Sadly, that does not work, because perl6 needs to see the stubbed class in full sometime later in the file.

So, the best option I’ve seen is to include the files containing needed classes and roles into files that need them (through either use or require). This requires knowing what to include (as including everything everywhere ruins the point), and I’ve started to try to grasp what’s going on in the rakudo codebase, specifically src/core, in terms of dependencies.

Here’s what I have so far. It’s right now just one, single script. This single script graphs the inheritance chain of all the classes and roles in src/core (what classes and roles “is” and “does”). You need the modules Term::ANSIColor, Term::ProgressBar, and IO::Capture::Simple (at this time pulled in, but apparently unused, by Term::ProgressBar). You also need ack and graphviz (which is not used in the script, but needed to process the resulting file).

Wanna see the graph for nom commit bf472b0, the latest as of this writing? Here you go (click to embiggen, clearly):

The arrows point towards what a particular box is inheriting, and the one or two white boxes mean that the class or role doesn’t exist within the src/core files. Yellow denotes classes, red denotes roles. And each box contains the name of the class/role, and the file it was found in and the line number of its definition.

This, however, doesn’t even begin to deal with all the dependencies abound in the src/core files. See all those abandoned colored blocks? Those aren’t inherited by anything in src/core, but they most certainly are utilized (e.g. the Pod::Config block, though not inherited, could easily be utilized as a variable type in the Pod code). I have a lot more work to do before I have a clear picture on how to separate CORE.setting into smaller pieces. (I might even need to utilize STD.pm6 or similar to do all the parsing I’ll eventually be doing!)

Just to finish off, Here are a couple more graphs of the same data. The first graph above was generated with GraphViz’s “dot”. This one (generated by their “twopi” program) I think looks hilarious:

And this picture was generated by GraphViz’s “fdp”. I think this one is the easiest to follow.

Happy viewing!

This entry was posted in Research Department, Progress Happened and tagged , , , . Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s