Handle zero-like imaginary part as zero in to_r #9581
Conversation
| @@ -1851,8 +1851,11 @@ nucomp_to_r(VALUE self) | |||
| get_dat1(self); | |||
|
|
|||
| if (!k_exact_zero_p(dat->imag)) { | |||
| rb_raise(rb_eRangeError, "can't convert %"PRIsVALUE" into Rational", | |||
| self); | |||
| VALUE imag = rb_check_convert_type_with_id(dat->imag, T_RATIONAL, "Rational", idTo_r); | |||
There was a problem hiding this comment.
why convert imag to rational here just to pass it into k_exact_zero_p()? Wouldn't it be easier to just change the check k_exact_zero_p(dat->imag) to f_zero_p(dat->imag) instead?
There was a problem hiding this comment.
@KJTsanaktsidis This approach doesn’t impose the responsibility of deciding what exactly constitutes zero on the Complex class. Instead, it assigns the responsibility to each individual class.
There was a problem hiding this comment.
And this implementation is for proof of concept. Before merging this to the master branch, I’ll add the optimization for Float class later.
There was a problem hiding this comment.
Ah that makes sense - thanks for the explanation!
mrkn
force-pushed
the
complex_to_r
branch
2 times, most recently
from
b019354
to
ab737e3
Compare
|
Waiting for ruby/prism#2468. |
|
Sorry about the delay @mrkn, you should be able to rebase and it should pass. Please let me know if you run into any more trouble! |
Use imag.to_r when imag is not 0 or 0r. Fixes [Bug ruby#5179]
Use imag.to_r when imag is not 0 or 0r.
Fixes [Bug #5179]