TheDoctor [they/them]

“Oh, they print in that order? That’s weird.”

I did both of these at once last week.

Added a breakpoint. Debugger didn’t break.

Added an echo "here";. Debugger didn’t print.

Added a throw new Exception('fuck');. Debugger didn’t throw.

Stepped through. Debugger wouldn’t let me step in.

It took me almost an hour to realize it wasn’t the debugger’s fault and that a variable I thought was guaranteed to be truthy at that point was actually falsey due to upstream changes in a spreadsheet parser. I felt kind of stupid for not trusting the debugger at that point.

Thank god

In the way that’s common in languages like Java where you’re making a property read-only, yes. But there’s a whole protocol in Python called descriptors where you can override the . on a field. The most common form of these is class methods annotated with the @property annotation, which makes it so the method can be accessed as if it were a property.

Yeah. I can understand the use case when it’s something relating to keeping simple state in sync by replacing it with derived state. But this particular case was flushing a cache after each get, which made each get of the property non-deterministic based on the class’s state.

I helped a friend debug a script last week that was working inconsistently in really weird ways. I looked at the script and it was all event hooks littered with sleep calls. I told him he was basically fuzz testing his own script and then getting surprised when he found race conditions. Shit was wild. Also, sometimes getters in Python are a mistake.

dozen = 12 + 1; // one extra for the baker!

I got mad at this when I first saw it but then I remembered there’s some code at work that defines an hour as 50 minutes

Statistically, this makes your code better

Yeah…. I’ve definitely been the next guy on a couple bad regexes that I wrote

When versioning and feature flags are too hard: just use git and hope for the best

My old senior used to do this before he got laid off and now I’m charge of code that’s littered with old commented out code and no way to know why it was commented out.

Then it breaks years after you’ve left and someone has no choice but to touch it

I often use comments as ways to say, “I know this is cursed, but here’s why the obvious solution won’t work.” Like so:

/**
 * The column on this table is badly named, but
 * renaming it is going to require an audit of our
 * db instances because we used to create them
 * by hand and there are some inconsistencies
 * that referential integrity breaks. This method
 * just does some basic checks and translates the
 * model’s property to be more understandable.
 * See [#27267] for more info.
 */

Writing plain old JavaScript without a library or framework is nice while you’re learning. Too many people will learn a single framework and not have any idea what the underlying APIs are, so the transferable skills are minimal.

I can’t imagine just opening a giant PR without having extensive contact and coordination with the maintainer. Almost any amount of incremental safe steps would be preferable to a giant PR, even with extensive communication. I once introduced fully strict typescript into a vanilla js codebase and it took dozens of small PRs to do so. It was made more complicated by the fact that it was a library, but still. The communication made the entire process smooth and let everyone be confident the changes were correct along the way. If I’d done it all at once without any coordination, it would have been faster for me, but at the cost of the maintainer’s sanity and time.