Pedro J. Aphalo

Here is an example with a CI (I removed the smooth line given by the fitted spline, to avoid confusions, as the _R_ value is for a linear correlation, not...

- [x] support `p.digits = Inf` to force use of scientific notation. - [x] test `p.digits = Inf` with `stat_correlation()`. - [x] test `p.digits = Inf` with the other stats....

@wbvguo A preliminary implementation of your request is now done. In the case of labels as R expressions, the e-notation is replaced by actual powers of 10. With labels as...

@wbvguo Many thanks for your suggestions! The idea with `p.digits` is to avoid showing too many digits. Although smart, the idea of using the fractional part of numbers may be...

I think that the best approach would be to add formal parameters named `p.critical` and `p.highlight.action`. This would make it possible to show values like _P_ < 0.050 for example...

@wbvguo Hi Wenbin, I did some additional changes to the logic of how nurmber formatting is applied to _P_-values. These changes, I hope address a bit better your use case....

Support for most of these features is now implemented and documented, unit tests written, and testing completed.

Not easy to implement because we need to pass all the variables to the geometry. As different geometries could be used this is not straightfoward, specially for lines.

At least all variables mapped by default would need to be set to NA or "", but to do this correctly every value returned would need to be set also...

One could possibly simply return an _empty_ data frame with `data.frame()`, I think. It needs to be tested. A warning would also need to be issued.