Adjust claims about the power of parse(Complex...)

[traktofon]

The statement

parse(Complex{T}, string) can parse complex numbers in common formats

is a bit exaggerating. Arguably one of the more common complex number output formats is that used by Fortran, which is of the form "(real, imag)". The current implementation of parse fails to parse this:

julia> parse(Complex{Float64}, "(1.0, 2.0)")
ERROR: ArgumentError: cannot parse "(1.0, 2.0)" as Float64

On version 0.7.0-DEV.3712.

[stevengj]

The NEWS file doesn't generally list all of the features the standard library doesn't have...

It doesn't say all common formats. I would just change it to say "several common formats" if you insist on qualifying this.

[traktofon]

How about "some common formats"? For me that would be a more accurate statement of the current abilities.

[StefanKarpinski]

We could also just implement parse(Complex{Float64}, "(1.0, 2.0)").

[stevengj]

@StefanKarpinski, in #22250 it was decided not to allow Fortran format. I didn't think it makes sense to parse Fortran-style complex numbers if we can't even parse Fortran-style real numbers (e.g. 1.0D+00).

[traktofon]

Thanks @stevengj for pointing out #22250, I hadn't seen that one. I agree that without support for parsing Fortran doubles, adding support for Fortran complex numbers doesn't make sense. Would it make sense to reference #22250 from the NEWS.md directly? Or is it policy to only reference PRs?

[stevengj]

Usually we reference the PR that closes the issue, especially (as it does in this case) when the PR links to the issue that it addresses.

[traktofon]

Thanks!