"Some people, when confronted with a problem, think 'I know, I'll use regular expressions.' Now they have two problems."

Perl’s documentation is far more clear about the consequences:

(https://perldoc.perl.org/perlop#Regexp-Quote-Like-Operators)

   o   Compile pattern only once.

  […]

  PATTERN may contain variables, which will be
  interpolated every time the pattern search is
  evaluated, except for when the delimiter is a
  single quote. […] Perl will not recompile the
  pattern unless an interpolated variable that
  it contains changes. You can force Perl to skip
  the test and never recompile by adding a /o
  (which stands for "once") after the trailing
  delimiter. Once upon a time, Perl would recompile
  regular expressions unnecessarily, and this
  modifier was useful to tell it not to do so,
  in the interests of speed. But now, the only
  reasons to use /o are one of:

  [reasons]

  The bottom line is that using /o is almost
  never a good idea.

  However, mentioning /o constitutes a promise
  that you won't change the variables in the
  pattern. If you change them, Perl won't even
  notice.

(https://github.com/Perl/perl5/blob/perl-4.0.00/perl.man#L272...)

  PATTERN may contain references to scalar
  variables, which will be interpolated
  (and the pattern recompiled) every time the
  pattern search is evaluated. […] If you want
  such a pattern to be compiled only once, add
  an “o” after the trailing delimiter. This
  avoids expensive run-time recompilations, and
  is useful when the value you are interpolating
  won't change over the life of the script.

> With nothing else to investigate, I finally looked up the docs for what the /o regex modifier does.

I'll probably never understand this mode of thinkning. But then again, Ruby programmers are, after all, people who chose to write Ruby.

> /o is referred to as “Interpolation mode”, which sounded pretty harmless.

Really? Those words sound quite alarming to me, due to personal reminiscences of eval.

Also, this whole "/o" feaure seems insane. If I have an interpolation in my regex, obviously I have to re-interpolate it every time a new value is submitted, or I'd hit this very bug. And if the value is expected to the same every time, then I can just compile it once and save the result myself, right? In which case, I probably could even do without interpolation in the first place.

That is crystal clear to me. It means that on the next execution, the new values of the interpolation will be ignored; the regexp is now "baked" with the first ones.

Like this in C++:

  void fun(int arg)
  {
     static int once = arg;
  }

There is a function in POSIX for once-only initializations: pthread_once. C++ compilers for multithreaded environments emit thread-safe code to do something similar to pthread_once to ensure that even if there are several concurrent first invocations of the function, the initialization happens once.

The opposite behaviour (we have a constant regular expression, we re-use it often but the tooling doesn't realise and so it's created each time we mention it) is not a footgun, it results in poor performance, and so you might want (especially in some managed languages) to just magically optimise this case, but if not you won't cause mysterious bugs. An expert, asked "Why is this slow?" can just fix it - you have to supply basic tools for that, but this flag is not a sensible tool.