Let std::format do the work #721
Conversation
Ubuntu 22.04 LTS does not support the C++20 features, but Ubunntu 24.04 LTS does.
C++20 features are supported starting from MacOS 13.3.
tobiolo
force-pushed
the
double-to-string.formatting
branch
from
a2bc02b
to
4174399
Compare
|
|
||
| t = s; | ||
| } | ||
| void SetNum(double d) { t = std::format("{:g}", d); } |
There was a problem hiding this comment.
This is unlikely to produce the same formatting as the code below. The existing code does a lot of specific things to achieve a precise formatting.
There was a problem hiding this comment.
But is this stuff really necessary?
There was a problem hiding this comment.
Most code that is there, especially if it does something which appears over complicated or does extra things, does so for a reason, yes.
Simplifying it is good, if you can indeed verify that the simplified version does the same thing, and if it doesn't, that should only be done if its an improvement.
For example, from the comments you can see that this implements a variable length float format, that avoids low precisions last digits and/or scientific notation.. basically a friendly human-like representation. There's no way to get that formatting with the standard functions that I know of.
There was a problem hiding this comment.
Thanks for your explanation! Thus I close this PR.
No description provided.