Seemingly permanent performance degradation after fiddling with command-line options

bug

#1

Yesterday evening I was working on trying to speed up some functions. Most I was tweaking the regex which provided a 13% speed boost. In an attempt to get more out of it I tried running Julia with different options. Namely --compile, largely out of curiosity. Everything I tried was slower, even when I returned it to “normal” meaning no options given.

I then gave up. Now I’m back at it and noticing that it is indeed slower, even after a reboot, and the memory usage is outrageous. What was using less than 700kb of memory is now using nearly 5.5MB (both after compiling)! Wondering if I might have introduced some kind of regression, I rolled back to various tagged points in my repo with similar results.

How I’m timing this is by loading the REPL with “julia” in terminal and include("path/to/my/script.jl") and running @time function("path/to/same/file/as/yesterday.tex") and reading off the time. I’m using the julia package from the official Arch repo. As far as I can tell between yesterday and today nothing was updated that should have caused performance degradation like this. The following were updated last night. (What I put up in my original opening post ISN’T what I updated last night. I’m not sure was since /var/log/packman.log is only showing that only yay was upgraded, which I don’t think is right.)

The only think of is that somehow, something I did last night stuck and isn’t changing back. It shouldn’t. I didn’t mess with any config files, just command line options.

I tried moving ~/.julia to a new location but this didn’t help.