Add to_string, prints a DDouble in the %e style.#7
Conversation
mwallerb
left a comment
mwallerb
left a comment
There was a problem hiding this comment.
Hi Ben,
Thank you for the PR.
However, this need quite a bit of work. Unit tests are missing completely. The formatting is inconsistent with the rest, and I think you should be getting quite an error buildup due to the repeated multiplications with 10.
Please see "Printing Floating-Point Numbers Quickly and Accurately with
Integers" by Florian Loitsch or the "Dragon4" algorithm. I think this is the way to go.
|
|
||
| XPREC_API_EXPORT | ||
| std::string to_string(DDouble d, size_t nDigits){ | ||
| using namespace std; |
There was a problem hiding this comment.
Please do not import the complete std namespace.
There was a problem hiding this comment.
Why? It's confined, there is no namespace pollution.
| bool isnan(DDouble x); | ||
| bool isnormal(DDouble x); | ||
| bool iszero(DDouble x); | ||
| std::string to_string(DDouble d, size_t nDigits = 34); |
There was a problem hiding this comment.
This is extending the to_string interface. I am not sure that is what we want to do. I think it is better to respect to use << (which can then be used together with a stringstring).
There was a problem hiding this comment.
You're right.
| //Alright. Get a string started. Handle sign. | ||
| string s; | ||
|
|
||
| if(d.hi() < 0.0) s += "-"; |
There was a problem hiding this comment.
Please do not do this.
There is a clang format file in the repo. You can use clang-format -i src/io.cxx to automatically format the file according to the spect.
There was a problem hiding this comment.
Don't do what?
There was a problem hiding this comment.
one-line if's, please split them across lines.
|
|
||
| //Current digit to print is just the leading one. | ||
| auto m = (int)floor(d.hi()); | ||
| assert(m >= 0 && m < 10); |
There was a problem hiding this comment.
This assertion will occasionally fail due to rounding.
| //Loop over digits, print them. | ||
| nDigits = min(max(nDigits, (size_t)3), (size_t)34); |
|
Thanks for the feedback and for pointing me to the other stuff. I'll let you know if I make time to work, it looks like it'd be a rewrite. |
2 participants
You may or may not want to move this around and/or make it match style better.
Sample use:
Output:
1.414213562373095048801688724209708
Actual value of sqrt(2) to same number of digits:
1.414213562373095048801688724209698