Ah, I've found where the bug is.
If it's written there as r = r * x, it's compiled into r = (NI)(r * x); and
it's ok.
But if it's written as r *= x, it's compiled into stareq__blah-blah((&r), x);
which obviously ignores that pragma. So it must have been like r *= x; in C
code instead of stareq
Yeah, the manual is absolutely helpless here. Now the question stays -- how to
make these checks off in the example above.
>From the [compiler
>manual](https://nim-lang.org/docs/nimc.html#compiler-usage-command-line-switches):
\--overflowChecks:on|off \- turn int over-/underflow checks on|off
When I run this program in a usual way (nim c -r), it stops with overflow
error. How can I switch this check inside? I tried what I could find in
doc/sources, but it does not seem to work.
{.push overflowChecks: off.}
proc `^`(x,y:int):int = (var r=1;for i in 1..y:r*=x;r)
