Why I always keep doing it...

xkcd 242 obviously

But sometimes it works, or throws a different error ...

You make a change. It doesn't fix it.

You change it back. The code now works.

The usual for me is that I flip back over to my editor and hit ctrl+save, cause heaven forbid I ever remember to do that before running.

Trying to debug race conditions be like

The first one is to warm up the engine. Like getting your car ignition to kick over in the winter

Code doesn't work; don't know why.

Code works; don't know why.

The absolute worst thing that can happen is if it suddenly starts working without doing anything

The error message goes stale when it's been sitting for a while. I need to see a fresh one.

You know, youve gotta give your computer some warmup.

Not sure which is worse. When you know you changed nothing and it inexplicably starts|stops working compared to yesterday

Running the code again is fast and requires no thinking. Finding the problem is slow and requires a lot of thinking.

It's worth looking under the light-post in case your keys somehow rolled there. Just not for long.

You jest but “wait and retry” is such a powerful tool in my DevOps toolbox. First thing I tell junior engineers when they run across anything weird

Ah, come on this is valid investigation.

If you get the same error every time, you know you can find it and debug it, somewhat with ease.

If you don’t, you might have a thornier issue at hand.

Computer needs practice to get program right.

I started coding professionally using Visual Basic (3!). Everybody made fun of VB's On Error Resume Next "solution" to error handling, which basically said if something goes wrong just move on to the next line of code. But apparently nobody knew about On Error Resume, which basically said if something goes wrong just execute the offending line again. This would of course manifest itself as a locked app and usually a rapidly-expanding memory footprint until the computer crashed. Basically the automated version of this meme.

BTW just to defend VB a little bit, you didn't actually have to use On Error Resume Next, you could do On Error Goto errorHandler and then put the errorHandler label at the bottom of your routine (after an Exit Sub) and do actual structured error handling. Not that anybody in the VB world ever actually did this.

gotta rule out cosimc rays flipping a bit or two

This would be more mockable if it didn't often WORK.

Just making sure that the write buffer was flushed or something.

When your Makefile is so fucked up that you have to run it multiple times to get everything to build and link properly.

sometimes it needs to warm up.. or cool down

And run it with the debugger.

sometimes you don't compile it enthusiastically enough

demons ahem. data-races.

Because you're Good developer

This is just how you use Visual Studio

Or the code you are working on is calling a system that is currently unreliable which you cannot be responsible for.

Fuck test automation, it's a fucking trap get out of it as soon as you can

... try it now.

does it count if i run it again, but with a debugger attached?

This genuinely happens regularly in our testing environment 🥲

Most applications aren't written to compile deterministically so there is always a chance.