Sometimes I enjoy reading the source code to projects I like, use, or am about to install for the first time. This was something I used to do on a very regular basis, looking for security issues to report. Nowadays I don't have so much free time, but I still like to inspect the source code to new applications I install, and every now and again I'll find the time to look at the source to random projects.
Reading code is good. Reading code is educational.
One application I've looked at multiple times is redis, which is a great example of clean and well-written code. That said when reading the redis codebase I couldn't help noticing that there were a reasonably large number of typos/spelling mistakes in the comments, so I submitted a pull-request:
Sadly that particular pull-request didn't receive too much attention, although a previous one updating the configuration file was accepted. I was recently reminded of these pull-requests when I was when I was doing some other work. So I figured I'd have a quick scan of a couple of other utilities.
In the past I'd just note spelling mistakes when I came across them, usually I'd be opening each file in a project one by one and reading them from top to bottom. (Sometimes I'd just open files in emacs and run "M-x ispell-comments-and-strings", but more often I'd just notice them with my eyes). It did strike me that if I were to do this in a more serious fashion it would be good to automate it.
So this time round I hacked up a simple "dump comments" utility, which would scan named files and output the contents of any comments (be they single-line, or multi-line). Once I'd done that I could spell-check easily:
$ go run dump-comments.go *.c > comments
$ aspell -c comments
Anyway the upshot of that was a pull-request against git:
We'll see if that makes its way live sometime. In case I get interested in doing this again I've updated my sysbox-utility collection to have a comments sub-command. That's a little more robust and reliable than my previous hack:
$ sysbox comments -pretty=true $(find . -name '*.c')
..
..
The comments sub-command has support for:
- Single-line comments, for C, as prefixed with
//. - Multi-line comments, for C++, as between
/*and*/. - Single-line comments, for shell, as prefixed with
#. - Lua comments, both single-line (prefixed with
--) and multiline between--[[and--]].
Adding new support would be trivial, I just need a start and end pattern to search against. Pull-requests welcome:
https://steve.fi/
My patch was accepted!